Filename
stringlengths 17
17
| Abstract
stringlengths 76
3.38k
| Highlights
stringlengths 80
1.73k
|
|---|---|---|
S0968090X20305891 | Short term traffic forecasting based on deep learning methods especially recurrent neural networks has received much attention in recent years . However the potential of RNN based models in traffic forecasting has not yet been fully exploited in terms of the predictive power of spatialtemporal data and the capability of handling missing data . In this paper we focus on RNN based models and attempt to reformulate the way to incorporate RNN and its variants into traffic prediction models . A stacked bidirectional and unidirectional LSTM network architecture is proposed to assist the design of neural network structures for traffic state forecasting . As a key component of the architecture the bidirectional LSTM is exploited to capture the forward and backward temporal dependencies in spatiotemporal data . To deal with missing values in spatialtemporal data we also propose a data imputation mechanism in the LSTM structure by designing an imputation unit to infer missing values and assist traffic prediction . The bidirectional version of LSTM I is incorporated in the SBU LSTM architecture . Two real world network wide traffic state datasets are used to conduct experiments and published to facilitate further traffic prediction research . The prediction performance of multiple types of multi layer LSTM or BDLSTM models is evaluated . Experimental results indicate that the proposed SBU LSTM architecture especially the two layer BDLSTM network can achieve superior performance for the network wide traffic prediction in both accuracy and robustness . Further comprehensive comparison results show that the proposed data imputation mechanism in the RNN based models can achieve outstanding prediction performance when the models input data contains different patterns of missing values . | A stacked bidirectional and unidirectional LSTM architecture for traffic forecasting. An LSTM structure with an imputation unit to infer missing values is proposed. The trade off between model capacity and complexity is evaluated. |
S0968090X20305908 | Ride sharing services that have been growing in recent years with the start of network service companies will be further enhanced by the recently emerging trend of applications for autonomous vehicles for future traveler mobility . One fundamental question that transportation managers should address is how to capture the endogenous traffic patterns involving the new and uncertain elements facing future transportation planning and management . By concentrating on one ideal system optimal scenario in which all vehicles are autonomous or can be centrally guided and all passengers pickup drop off trip requests can be given at the beginning this paper aims to integrate travel demand vehicle supply and limited infrastructure . Available ride shared and autonomous vehicles from different depots can be optimally assigned to satisfy passengers trip requests while considering the endogenous congestion in capacitated networks . A number of decomposition approaches are adopted in this research . Focusing on this primal problem we propose an arc based vehicle based integer linear programming model in space time state networks which is solved by Dantzig Wolfe decomposition . From the perspective of dynamic traffic assignment a space time state path based flow based linear programming model is also provided as an approximation according to the mapping information between vehicle and passenger and between a vehicle and the space time arc in each STS path in our priori generated column pool . Finally numerical experiments are performed to demonstrate our decomposition approaches and their computation efficiency . From our preliminary experiments we have a few interesting observations without considering road congestion the network performance efficiency could be overestimated passengers required pickup and drop off time windows could be a buffer to mitigate road congestion without impacting system performance the ride sharing service could reduce the total transportation system cost under centralized control . | Integrate trip requests vehicle supply and infrastructure with endogenous congestions. Develop space time state network flow models with ridesharing and road capacity. Apply Dantzig Wolfe decomposition to decompose the primal problem. A column pool based linear programming model proposed as an approximation solution approach. |
S0968090X2030591X | This study models and manages the parking sharing problem in urban cities where private parking owners can share their vacant spaces to parking users via a parking sharing platform . The proposed model takes into account the spatial dimension of parking where clusters of curbside spaces and private shareable ones are distributed over different locations . On the supply side private parking owners can sell the right of use of their spaces to the platform based on the rent they can receive and the inconvenience they would experience due to sharing . On the demand side travelers make their parking choices of space type and location under given parking capacities and prices . The resulting parking choice equilibrium is formulated as a minimization problem and several properties of the equilibrium are identified and discussed . The platform operators pricing strategy i.e . rent paid to space owners and price charged on space users can significantly affect the private parking owners sharing decisions and the choice equilibrium of parking users . In this context we examine the platform operators optimal pricing strategies for revenue maximization or social cost minimization . Numerical examples are also presented to illustrate the models and results and to provide further insights . | This paper models the parking sharing problem with spatially distributed parking supplies. This paper studies the parking choice equilibrium of travelers. This paper examines the pricing strategies of a revenue maximizing operator. This paper examines the pricing strategies of a social cost minimizing operator. |
S0968090X20305921 | This study examines the effects of a time varying congestion toll and a location dependent parking fee on the behavior of heterogeneous commuters and their commuting costs . To this end we develop a model of departure time and parking location choices by heterogeneous commuters and characterize its equilibrium . By comparing the equilibrium with and without pricing policies we obtain the following results imposing a parking fee and expanding parking capacity may concentrate the temporal distribution of traffic demand thereby exacerbating traffic congestion the expansion of parking capacity does not necessarily lead to a Pareto improvement when a parking fee is not imposed the social optimum is achieved by combining a parking fee with a congestion toll and the revenue obtained from pricing of roads and parking exactly equals the costs for optimal bottleneck and parking capacities respectively that is the self financing principle holds separately for bottleneck capacity and parking capacity . | We develop a model of departure time and parking location choices by heterogeneous commuters. We examine the distributional effects of imposing a congestion toll and or a parking fee. The expansion of parking capacity does not necessarily lead to a Pareto improvement when a parking fee is not imposed. The self financing principle holds separately for bottleneck capacity and parking capacity. |
S0968090X20305933 | One major challenge for on demand mobility service providers is to seamlessly match empty vehicles with trip requests so that the total vacant mileage is minimized . In this study we develop an innovative data driven approach for devising efficient vehicle relocation policy for OMS that proactively relocates vehicles before the demand is observed and reduces the inequality among drivers income so that the proactive relocation policy is fair and is likely to be followed by drivers . Our approach represents the fusion of optimization and machine learning methods which comprises three steps First we formulate the optimal proactive relocation as an optimal stable matching problems and solve for global optimal solutions based on historical data . Second the optimal solutions are then grouped and fed to train the deep learning models which consist of fully connected layers and long short term memory networks . Low rank approximation is introduced to reduce the model complexity and improve the training performances . Finally we use the trained model to predict the relocation policy which can be implemented in real time . We conduct comprehensive numerical experiments and sensitivity analyses to demonstrate the performances of the proposed method using New York City taxi data . The results suggest that our method will reduce empty mileage per trip by 5470 under the optimal matching strategy and a 2532 reduction can also be achieved by following the stable matching strategy . We also validate that the predicted relocation policies are robust in the presence of uncertain passenger demand level and passenger trip requesting behavior . | Developing the proactive vehicle relocation method for the managment of city wide OMS vehicles. Optimal vehicle relocation by fusing optimization and machine learning methods. Proposing two relocation schemes suitable for both human drivers and future autonomous vehicles of OMS platform. Using a low rank approximation method to reduce complexity and improve prediction performances. Extensive numerical experiments and sensitivity analyses to demonstrate the effectiveness of the proposed method. |
S0968090X20305945 | Operating speed profiles represent drivers responses to roadway geometry and are widely used to evaluate safety performance of roadway design . To predict operating speed profile the majority of early research followed a two step modeling procedure estimate speeds at start middle and end points of road segments and fill the profile between the points with assumed driver behavior . This sparse spot based modeling strategy has been shown to be inadequate for capturing the complex speed changes resulting from the overlapping horizontal and vertical curves on mountainous roads . This paper proposes a high resolution modeling approach for operating speeds measured in a dense series of equidistant spots along a road . This type of model is more conducive to analysis of mountainous freeway alignments as operating speeds are predicted along the entire roadway . The high resolution data were obtained using the Tongji University Driving Simulator from a simulated section of mountainous freeway . The estimated linear mixed model includes geometric variables representing the road upstream and downstream of each data collection spot . To determine the suitable lengths of the upstream and downstream segments the data were extracted from several alternative segment lengths including fixed lengths and varying downstream length accordingly to sight distances . The model with a spherical structure of error covariance using geometric data extracted from 300 meter upstream and downstream segments performed the best . An out of sample evaluation of the model has the mean absolute error of 3.2km h and the root mean square error of 4.2km h which indicates a promising prediction ability of the proposed model . | A high resolution speed modeling strategy for mountainous freeways was proposed. Geometric parameters from adjacent segments were used to model speed changes. The linear mixed model was applied to handle the spatial autocorrelation problem. An out of sample test showed the promising prediction ability of the model. |
S0968090X20305957 | In this paper we study the problem of computing train trajectories in an uncertain environment in which the values of some system parameters are difficult to determine . Specifically we consider uncertainty in traction force and train resistance and their impact on travel time and energy consumption . Our ultimate goal is to be able to control trains such that they will arrive on time i.e . within the planned running time regardless of uncertain factors affecting their dynamic or kinematic performance . We formulate the problem as a Markov decision process and solve it using a novel numerical approach which combines an off line approximate dynamic programming method to learn the energy and time costs over iterations and an on line search process to determine energy efficient driving strategies that respect the real time time windows more in general expressed as train path envelope constraints . To evaluate the performance of our approach we conducted a numerical study using real life railway infrastructure and train data . Compared to a set of benchmark driving strategies the trajectories from our ADP based method reduce the probability of delayed arrival and at the same time are able to better use the available running time for energy saving . Our results show that accounting for uncertainty is relevant when computing train trajectories and that our ADP based method can handle this uncertainty effectively . | The uncertainty in traction effort and train resistance is considered. An approximate dynamic programming approach for off line value function learning. A comprehensive comparison of the proposed method and other popular approachess. |
S0968090X20305969 | Train stop plans and timetables play key roles in railway operation . Previous research has demonstrated that their integration can significantly improve the quality of a train timetable especially for commuter railways with flexible service frequencies and multiple stop plans . However solving the dilemma of the mathematical tractability and practicality of the model is still an open challenge . To obtain a high quality timetable and simultaneously consider more realistic conditions an integrated combination optimization model of both train stop plans and timetables under time dependent passenger demand is proposed in this article . More realistic conditions such as no predefined schedule a variable total number of trains and oversaturation are taken into account . The problem is modeled as a mixed integer nonlinear programming problem to optimize passenger travel efficiency and mainly consists of the total waiting time at stations the delay time for trains due to a train stop and the minimization of the total train running time . An extended adaptive large scale neighborhood search algorithm is developed to solve the problem . A numerical experiment is designed to test the validity of the model and the algorithm . Then the integrated approach is applied in a real world case . The results show that the proposed approach can simultaneously reduce the passenger total waiting time and delay time as well as the train running time within a short computation time and demonstrate the effectiveness of the model and the approach . | Develop integrated model to optimize the train stop plan and timetable under time dependent passenger demand. Improves passenger travel efficiency and reduces train running costs. The number of trains and stops can be changed freely oversaturation is permitted. Improved extended adaptive large neighborhood search approach for solving. |
S0968090X20305970 | The theoretical analysis of traffic flow with empirical vehicle trajectory data contained within this study allows for the explanation reconstruction and prediction of spatiotemporal transition characteristics of traffic conditions . Using an unmanned aerial vehicle during a morning rush hour on a working day observations of congestion evolution near an on ramp bottleneck of an expressway was captured . The empirical high fidelity trajectory data of 621 vehicles were extracted . The major findings include | The macroscopic and microscopic characteristics of traffic flow in congestion evolution near expressway bottleneck were captured. The impact of lane changing behavior on the formation and propagation of oscillation was quantified. Inducement for triggering traffic phase transition was investigated. Investigation on the critical headway of engendering perturbation. Fluctuation magnitude of speed and the theoretical lag distance. |
S0968090X20305982 | This study proposes a novel approach to predict real time crash risk at signalized intersections at the signal cycle level . The approach uses traffic conflicts extracted from informative vehicle trajectories as an intermediate for crash prediction and develops generalized extreme value models based on conflict extremes . Moreover a Bayesian hierarchical structure is developed for the GEV model to combine conflict extremes of different intersections and the aim is to further improve safety estimates through borrowing strength from different intersections and accounting for non stationarity and unobserved heterogeneity in conflict extremes . The proposed approach was applied to four signalized intersections in City of Surrey British Columbia . Traffic conflicts measured by modified time to collision and three cycle level traffic parameters were extracted from collected video data using computer vision techniques and a best fitted model was then developed . Two safety indices risk of crash and return level of a cycle were derived from the GEV model to quantitatively measure the safety cycle by cycle . The results show that the non negative RC can directly point out cycles with crash prone traffic conditions with RC 0 and RLC is a more flexible safety index which can differentiate between safety levels even for safe cycles with RC 0 . The real time crash prediction results are validated at an aggregate level by comparing to observed crashes . | A Bayesian hierarchical extreme value model is developed for intersection safety prediction at the signal cycle level. Traffic conflicts automated extracted from informative vehicle trajectories are used for model development. Non stationarity and unobserved heterogeneity in conflict extremes are accounted for to improve safety estimates. Risk of crash RoC and return level of a cycle RLC are developed as real time safety indicators. RoC and RLC quantitatively show how risky a crash prone traffic condition is. |
S0968090X20305994 | This paper presents the development of a multilevel optimization framework for the design and selection of departure routes and the distribution of aircraft movements among these routes while taking the sequence and separation requirements for aircraft on runways and along selected routes into account . The main aim of the framework is to minimize aircraft noise impact on communities around an airport and the associated fuel consumption . The proposed framework features two consecutive steps . In the first step for each given Standard Instrument Departure multi objective trajectory optimization is utilized to generate a comprehensive set of possible alternative routes . The obtained set is subsequently used as input for the optimization problem in the second step . In this step the selection of routes for each SID and the distribution of aircraft movements among these routes are optimized simultaneously . To ensure the feasibility of optimized solutions for an entire operational day the sequence and separation requirements for aircraft on runways and along selected routes are included in this second phase . In order to address these issues three novel techniques are developed and added to a previously developed multilevel optimization framework | The route design and flight allocation problems are considered in a linked manner. Three novel techniques are developed and integrated into the proposed framework. The reliability and efficiency of the proposed approach are evaluated. The framework can offer conflict free solutions outperforming the reference case. |
S0968090X20306008 | Real time network control strategies such as congestion pricing have been used in a number of metropolitan areas around the world for traffic congestion mitigation . Recent advances in Global Navigation Satellite System technology have led to increasing interest in distance or usage based road pricing as an effective alternative to traditional facility cordon and area based pricing that typically rely on fixed infrastructure . In this paper we propose the use of feature variant clustering methods OPTICS and HDBSCAN as a systematic approach for tolling zone definition to operationalize distance based tolling schemes . Subsequently we develop a framework for predictive distance based toll optimization to evaluate network performance for the various tolling zone definitions derived from the aforementioned feature variant clustering methods . In this framework for a specific tolling zone definition tolling function parameters are optimized using a simulation based Dynamic Traffic Assignment model operating within a rolling horizon scheme . Predictive optimization is integrated with the guidance information generation . Behavioral models capture drivers responses to the tolls in terms of trip cancellation and choices of mode route and departure time . Experiments on the real world Expressway and Major Arterials network of Singapore demonstrate improved effectiveness of distance based toll optimization given tolling zone definitions derived from feature variant clustering compared to fixed cordon based pricing adaptive cordon based pricing as well as distance based pricing with ad hoc tolling zone definitions . Further the results indicate that the use of the marginal link cost tolls as a clustering feature produces the most robust tolling zone definitions and yields significant improvements in social welfare over ad hoc zone definitions and cordon pricing . Finally experiments on the Boston CBD network also demonstrate the effectiveness of distance based toll optimization schemes on urban traffic networks . | OPTICS HDBSCAN used to define tolling zones for distance based pricing schemes. Clustering features link speeds marginal cost tolls for tolling zone definition. Experiments on Singapore Expressway Boston CBD networks for various pricing schemes. Results show that distance based schemes outperform cordon based schemes. OPTICS HDBSCAN using marginal costs tolls suitable for tolling zone definitions. |
S0968090X2030601X | This paper builds an economic model to investigate the interactions between baggage fee and airport congestion . With checked bag fees we find that the socially optimal numbers of passengers and checked bags will not be achieved if the marginal congestion cost from carry ons is larger than that from checked bags . However it is possible with bundle pricing . Furthermore we use numerical experiments to illustrate that checked bag pricing may underperform bundle pricing in terms of social welfare . | We investigate airline baggage fees impacts on airport congestion. Both checked bags and carry ons could cause airport congestion. Airline competition may not be able to induce social optimum. Shifting from bundle pricing to checked bag pricing may lower social welfare. |
S0968090X20306124 | Real time crash prediction is essential for proactive traffic safety management . However developing an accurate prediction model is challenging as the traffic data of crash and non crash cases are extremely imbalanced . Most of the previous studies undersampled non crash cases to balance the data which may not capture the heterogeneity of the full non crash data . This study aims to use the emerging deep learning method called deep convolutional generative adversarial network model to fully understand the traffic data leading to crashes . With the full understanding of the traffic data of crashes the DCGAN model could generate more synthetic data related to crashes to balance the dataset . All non crash data could be used for developing the prediction models . To capture the correlations between different variables the data are augmented to 2 D matrix as the input for the DCGAN model . The suggested model is evaluated based on data from expressways and compared to two counterparts synthetic minority over sampling technique random undersampling technique . The results suggest that the DCGAN could better understand the crash data characteristics by generating data with better fit of the real data distribution . Four different crash prediction algorithms are developed based on each balanced data and totally twelve models were estimated . The results indicate that the convolutional neural network model based on the DCGAN balanced data could provide the best prediction accuracy validating that the proposed oversampling method could be used for the data balance . Besides compared to other two models only the DCGAN based model could identify the significant effects of speed difference between the upstream and downstream locations which could help guide traffic management strategies . With the prediction model developed based on the balanced data by DCGAN it is expected that more crashes could be predicted and prevented with more appropriate proactive traffic safety management strategies such as Variable Speed Limits and Dynamic Message Signs . | Introduce a GAN model to generate traffic data related to crashes for data balance. Use the CNN model to incorporate the correlations between variables. Develop various models to compare oversampling and undersampling data. |
S0968090X20306136 | This study provides a new method for better incorporating human factors in modeling car following behavior . As the primary decision maker and vehicle operator human driver is the vital component of the driving process . During the driving process an external stimulus may trigger short term psychological changes and these changes are considered as the endogenous cause of many abnormal driving behaviors which often lead to unsafe traffic disturbances and even crashes . In this paper we investigate the intrinsic long term driving characteristics and its short term changes after driver experiences an external stimulus . A long and short term driving model is proposed to incorporate such changes into car following driving behavior modelling . The long term driving characteristics are extracted through a cluster analysis and the changes after an external stimulus are identified and measured as the indicator of the short term driving characteristics . NGSIM data are used to demonstrate the existence of LSTD characteristics and the soundness of the LSTD model . Two classical car following models are integrated with the LSTD model and the integrated models show a promising performance as the errors decrease by 36.7 and 35.7 respectively . | The LSTD model is proposed to describe long and short term driving characteristics. The long term characteristics are classified into three types via cluster analysis. The short term characteristics are identified and divided into two types. The six types of LSTD characteristics are incorporated into IDM and Gipps model. Significant improvement is demonstrated by incorporating LSTD into CF models. |
S0968090X20306148 | Mobility as a Service or MaaS offers potential consumers access to multiple transport modes and services owned and operated by different mobility service providers through an integrated digital platform for planning booking and payment . We surveyed 3985 geographically and demographically representative Australians nationwide to understand consumer demand and willingness to pay for MaaS in Australia . Our analysis reveals that there is definitely a market for MaaS in Australia . Depending on the service offering we find that up to 46 per cent of the Australian population would be willing to adopt MaaS . On average consumers prefer pay as you go schemes to bundled schemes that offer unlimited access to one or more transport modes and services at fixed monthly costs . Local public transport taxis and long distance public transport are the most popular transport services bikeshare is the least popular . Willingness to use MaaS is strongly correlated with age and lifecycle stage young individuals who are employed full time are most likely to use MaaS older adults who have retired from the workforce and whose children have left home are least likely to use MaaS . Our analysis identifies Melbourne Canberra and Sydney as good markets for MaaS trials and early launches . | Up to 46 per cent of Australian population willing to adopt MaaS. Consumers prefer pay as you go MaaS schemes to bundled MaaS schemes. Public transport and taxis are the most popular transport services. Willingness to use MaaS is negatively correlated with age and lifecycle stage. |
S0968090X2030615X | Disruptive events lead to capacity degradation of transportation infrastructure and a good restoration plan could minimize the aftermath impacts during the recovery period . This is considered one aspect of resiliency for transportation systems . Although unmet demand has been proposed as one measure of resilience for freight transportation it has rarely been used for general transportation systems . This study takes unmet demand and total travel time as two measures in modeling the restoration plan problem and proposes a bi objective bi level optimization framework to determine an optimal transportation infrastructure restoration plan . The lower level problem uses Elastic User Equilibrium to model the imbalance between demand and supply and measures the unmet demand for a given transportation network . The upper level problem formulated as bi objective mathematical programming determines optimal resource allocation for roadway restoration . The bi level problems are solved by a modified active set algorithm and a network representation method derived from Network Design Problems . The Weighted Sum Method is adopted to solve the Pareto Frontier of this bi objective optimization problem . The proposed restoration plan optimization method was applied to a typical road network in Sioux Falls to verify the effectiveness of the methodology . For a given failure scenario the Pareto Frontier of this bi objective bi level optimization problem with various budget levels cross referring to the travel efficiency of each solution was illustrated to demonstrate how the proposed method can support decision making for road network restoration . To further study the performance of the proposed method different scenarios were generated with one to five links disrupted and the proposed methodology was applied with different budget levels . The statistical analysis of the optimized solutions for these scenarios demonstrates that a higher budget could help reduce unmet demand in the system by providing more restoration options . | Unmet demand is proposed as one resilience measure for roadway restoration. A bi objective bi level optimization framework is proposed for the RPO problem. Integrated different methods to solve the RPO problem efficiently. Insights obtained from statistical analysis of outcomes from comprehensive numerical experiments. Cross reference of bi objectives and additional metrics helps decision makers choose the best plan. |
S0968090X20306161 | While deep neural networks have been increasingly applied to choice analysis showing high predictive power it is unclear to what extent researchers can interpret economic information from DNNs . This paper demonstrates that DNNs can provide economic information as | Extract economic information from DNN for choice analysis. Introduce both function based and gradient based interpretations. Highlight three challenges associated with the automatic learning capacity of DNN. Compare economic information from DNN with those from discrete choice models. Economic information aggregated either over trainings or population is more reliable. |
S0968090X20306173 | For the transport sector promoting carpooling to private car users could be an effective strategy over reducing vehicle kilometers traveled . Theoretical studies have verified that carpooling is not only beneficial to drivers and passengers but also to the environment . Nevertheless despite carpooling having a huge potential market in car commuters it is not widely used in practice worldwide . In this paper we develop a passenger to driver matching model based on the characteristics of a private car based carpooling service and propose an estimation method for time based costs as well as the psychological costs of carpooling trips taking into account the potential motivations and preferences of potential carpoolers . We test the model using commuting data for the Greater London from the UK Census 2011 and travel time data from Uber . We investigate the service sensitivity to varying carpooling participant rates and fee sharing ratios with the aim of improving matching performance at least cost . Finally to illustrate how our matching model might be used we test some practical carpooling promotion instruments . We found that higher participant role flexibility in the system can improve matching performance significantly . Encouraging commuters to walk helps form more carpooling trips and further reduces carbon emissions . Different fee sharing ratios can influence matching performance hence determination of optimal pricing should be based on the specific matching model and its cost parameters . Disincentives like parking charges and congestion charges seem to have a greater effect on carpooling choice than incentives like preferential parking and subsidies . The proposed model and associated findings provide valuable insights for designing an effective matching system and incentive scheme for carpooling services in practice . | A passenger to driver matching model for commuter carpooling was initially proposed. 38.3 of trips can successfully form carpooling trips and save fuel significantly. The participant flexibilities can help matching performance especially when participant rate in a lower level. The optimal fee sharing ratio is not the traditional half to half. The disincentives seem to have a greater effect on carpooling choice than the incentives. |
S0968090X20306185 | This study proposes an addition to the architecture of the context aware Driver Assistance Systems by introducing a context identification layer to the reasoning subsystem . The proposed layer contains two algorithms that work in sequence the infrastructure detection algorithm and the driver classification algorithm respectively . The infrastructure detection algorithm aims to identify intersection related driving when the driver adjusts his or her behavior due to the presence of an intersection ahead . Then the driver classification algorithm categorizes drivers into cautious normal and aggressive at both locations . Data from 64 drivers in a Naturalistic Driving Study was used to prove the concept of the proposed layer . Several behavioral measures were extracted including following distance relative speed headway acceleration time to collision and jerk . These behavioral measures were then used to train the algorithms in the context identification layer . The results of both algorithms supported the concept of the proposed layer . These results have implications related to driver behaviors including i the intersection related driving behavior can be detected ii the drivers tend to be relatively aggressive at intersections when compared to segments and iii the driver classification which ignores the drivers relative location to intersections were more likely to misclassify drivers as aggressive when they were in high intersection density areas such as downtown cores . The findings of this study emphasized the importance of context aware DAS architecture that acknowledges and integrates the variation in driver behavior due to both a change in the surrounding environment and drivers individual needs . | Adding a context identification layer to context aware Driver Assistance Systems. Integrating the surrounding environment and drivers needs variation into the system. Stressing on the difference between driver behavior at intersections and on segments. Detecting intersection related driving behavior. Classifying drivers at intersections and on segments using six behavioral measures. |
S0968090X20306197 | Smart card data enables the estimation of passenger delays throughout the public transit network . However this delay is measured per passenger trajectory and not per network component . The implication is that it is currently not possible to identify the contribution of individual system components stations and track segments to overall passenger delay and thus prioritize investments and disruption management measures accordingly . To this end we propose a novel method for attributing passenger delays to individual transit network elements from individual passenger trajectories . We decompose the delay along a passenger trajectory into its corresponding track segment delay initial waiting time and transfer delay . Using these delay components we construct a solvable system of equations using which the delays on each network component can be computed . The estimation method is demonstrated on one year of data from the Washington DC metro network . Our approach produces promising results by compressing millions of individual trajectories into 3D networks leading to a dimensionality reduction of 94 . Moreover the mean slack variable value is smaller than five seconds per passenger and has the desired positive sign for almost 90 of all travelers . Applications using the estimation results include revealing network wide recurrent delay patterns modeling delay propagation and detecting disruptions . | Decomposed delay along a passenger trajectory into corresponding network components. Constructed 3D delay networks from trajectories to represent transit delay dynamics. Demonstrated the method on one year smart card data from Washington DC metro network. Achieved a dimensionality reduction of 94 . Computed almost 90 of the passengers delay with a mean error of 5 s per passenger. |
S0968090X20306203 | Bike sharing systems being viewed as providing green transportation modes are growing rapidly in recent years . While BSS operators try to improve the system performance through bike rebalancing and launching more bikes the current BSSs are facing several sustainability challenges . Bike oversupply could bring intensive greenhouse gas emissions due to manufacturing excessive bikes while frequent bike rebalancing could significantly increase fuel consumption of rebalancing vehicles . Existing studies only optimized given BSS from the system operational perspective with predetermined bike fleet size and rebalancing frequencies . However the bike fleet size and rebalancing should also be optimized from the life cycles perspective . This study proposes a framework to obtain the optimal bike fleet size and rebalancing strategy to minimize the systems life cycle GHG emissions integrating a simulation model for fleet size estimation an optimization model for bike rebalancing and a life cycle assessment model to quantify the systems GHG emission rate . The framework is applied to a dock less BSS in Xiamen China as a case study to evaluate the tradeoff between having more bikes and more frequent rebalancing . Our results show that the current BSS in Xiamen is significantly oversupplied with only 15 of current bikes needed to serve the same demands decreasing bike fleet size through more frequent rebalancing will increase the systems life cycle GHG emissions and choosing appropriate rebalancing fleet size loading capacity and setting multiple depots can reduce a BSS rebalancing GHG emissions . | Proposed a modeling framework to optimize bike sharing systems fleet size and rebalancing. Less frequent rebalancing can reduce dock less BSS life cycle GHG emission rate. The current dock less BSS in Xiamen China is significantly oversupplied with bikes. Rebalancing vehicle fleet can be optimized to reduce rebalancing GHG emissions. Having multiple rebalancing depots can help reduce systems GHG emissions. |
S0968090X20306215 | Understanding individual and crowd dynamics in urban environments is critical for numerous applications such as urban planning traffic forecasting and location based services . However researchers have developed travel demand models to accomplish this task with survey data that are expensive and acquired at low frequencies . In contrast emerging data collection methods have enabled researchers to leverage machine learning techniques with a tremendous amount of mobility data for analyzing and forecasting peoples behaviors . In this study we developed a reinforcement learning based approach for modeling and simulation of people mass movement using the global positioning system data . Unlike traditional travel demand modeling approaches our method focuses on the problem of inferring the spatio temporal preferences of individuals from the observed trajectories and is based on inverse reinforcement learning techniques . We applied the model to the data collected from a smartphone application and attempted to replicate a large amount of the populations daily movement by incorporating with agent based multi modal traffic simulation technologies . The simulation results indicate that agents can successfully learn and generate human like travel activities . Furthermore the proposed model performance significantly outperforms the existing methods in synthetic urban dynamics . | A reinforcement learning based agent model is developed for travel demand forecasting. The method is capable of learning and imitating people travel behavior from unlabeled GPS data. A synthetic dataset is generated and evaluated by comparing with real mobility dataset. |
S0968090X20306227 | Studies in several cities indicate that ridesourcing may increase traffic and congestion given the substitution of more sustainable modes and the addition of empty kilometers . On the other hand there is little evidence if smartphone apps that target shared rides have any influence on reducing traffic levels . We study the effects of a shared mobility service offered by a start up in Mexico City Jetty which is used by travelers to book a shared ride in a car van or bus . A large scale user survey was conducted to study trip characteristics reasons for using the platform and the general travel choices of Jetty users . We calculate travel distance per trip leg for the current choices and for the modes that riders would have chosen if the platform was not available . We find that the effect of the platform on vehicle kilometers traveled depends on the rate of empty kilometers introduced by the fleet of vehicles the substitution of public versus private transport modes the occupancy rate of Jetty vehicles and assumptions on the occupancy rate of substituted modes . Following a sensitivity analysis approach for variables with unavailable data we estimate that shared rides in cars increase VKT shared vans are able to decrease VKT whereas buses are estimated to increase VKT in our preferred scenarios . These results stem from the tradeoff between the effects of the occupancy rates per vehicle and the attractiveness of the service for car users . Our findings point to the relevance of shared rides in bigger vehicles such as vans as competitors to low occupancy car services for the future of mobility in cities and to the improvement of public transportation services through the inclusion of quality attributes as provided by new shared mobility services . | Ridesourcing has been shown to likely increase traffic. The actual effect of apps for shared rides on traffic levels is unknown. The impact of shared rides on empty kilometers and traffic levels is studied. Shared rides in cars increase VKT but shared rides in vans reduce VKT. |
S0968090X20306239 | Real time traffic crash prediction has been a major concern in the development of Collision Avoidance Systems along with other intelligent and resilient transportation technologies . There has been a pronounced progress in the use of machine learning models for crash events assessment by the transportation safety research community in recent years . However little attention has been paid so far to evaluating real time crash occurrences within information fusion systems . The main aim of this paper is to design and validate an ensemble fusion framework founded on the use of various base classifiers that operate on fused features and a Meta classifier that learns from base classifiers results to acquire more performant crash predictions . A data driven approach was adopted to investigate the potential of fusing four real time and continuous categories of features namely physiological signals driver maneuvering inputs vehicle kinematics and weather covariates in order to systematically identify the crash strongest precursors through feature selection techniques . Moreover a resampling based scheme including Bagging and Boosting is conducted to generate diversity in learner combinations comprising Bayesian Learners k Nearest Neighbors Support Vector Machine and Multilayer Perceptron . To ensure that the proposed framework provide powerful and stable decisions an imbalance learning strategy was adopted using the Synthetic Minority Oversampling TEchnique to address the class imbalance problem as crash events usually occur in rare instances . The findings show that Boosting depicted the highest performance within the fusion scheme and can accomplish a maximum of 93.66 F1 score and 94.81 G mean with Nave Bayes Bayesian Networks k NN and SVM with MLP as the Meta classifier . To the best of our knowledge this work presents the first attempt at establishing a fusing framework on the basis of data from the four aforementioned categories and fusion models while accounting for class imbalance . Overall the method and findings provide new insights into crash prediction and can be harnessed as a promising tool to improve intervention efforts related to traffic intelligent transportation systems . | Design and validate a fusion framework for real time crash prediction. Information fusion strategy based on four distinct categories of features. Diversity generation using four learners BL kNN SVM and MLP. Boosting and Bagging with Meta Classifier for more robust outcomes. An Imbalance learning founded on SMOTE for crash events prediction. |
S0968090X20306252 | Autonomous vehicles are inevitably entering our lives with potential benefits for improved traffic safety mobility and accessibility . However AVs benefits also introduce a serious potential challenge in the form of complex interactions with human driven vehicles . The emergence of AVs introduces uncertainty in the behavior of human actors and in the impact of the AV manufacturer on autonomous driving design . This paper thus aims to investigate how AVs affect road safety and to design socially optimal liability rules in comparative negligence for AVs and human drivers . A unified game is developed including a Nash game between human drivers a Stackelberg game between the AV manufacturer and HVs and a Stackelberg game between the lawmaker and other users . We also establish the existence and uniqueness of the equilibrium of the game . The game is then simulated with numerical examples to investigate the emergence of human drivers moral hazard the AV manufacturers role in traffic safety and the lawmakers role in liability design . Our findings demonstrate that human drivers could develop moral hazard if they perceive their road environment has become safer and an optimal liability rule design is crucial to improve social welfare with advanced transportation technologies . More generally the game theoretic model developed in this paper provides an analytical tool to assist policy makers in AV policymaking and hopefully mitigate uncertainty in the existing regulation landscape about AV technologies . | Develops a hierarchical game for a mixed AV HV market. Investigates how AVs impact road safety designs. Socially optimal liability rules for AVs and human drivers. |
S0968090X20306276 | Reporting and hypothetical biases are inherent to canonical methods of transportation data collection and had implied that analysis in this field has often neglected aspects that are strong behavioral drivers such as uncertainty physical effort or stress . Granular information on these aspects would allow measuring their valuation and or addressing a pervasive source of endogeneity . Recent advances in miniaturization and data processing as well as evidence that indicators from biosensors correlate with psychophysiological states and emotions suggest that there is an opportunity to close this gap by collecting a new type of data from transportation users . This research works on leveraging this opportunity by putting forward illustrating and testing a methodological framework to incorporate psychophysiological indicators gathered from biosensors into transportation choice behavioral modeling . The proposed framework adapts the integrated choice and latent variable approach by incorporating the psychophysiological responses as additional indicators of a latent psychophysiological state that covariates with utility . For the practical implementation of the proposed framework we also consider a specific form of aggregation of the indicators across time to avoid the curse of dimensionality arising from the unmanageably large number of folds for integration . The proposed framework is illustrated and validated using Monte Carlo simulations . Besides a prototype field experiment was designed and performed to confirm the validity of three crucial components of the proposed framework the relation between transportation markers and emotions the possibility of measuring those emotions through biosensors installed on travelers and the validity of the proposed aggregation needed for practicality . In the experiment a public transportation user travelled wearing a Printed Circuit Board that integrated tiny biosensors to capture electrodermal activity heart rate variation temperature and acceleration . Results provide positive evidence for the research questions suggesting the convenience of developing larger data collection efforts in the future to take full advantage of the new framework . | Reporting and hypothetical bias are inherent to canonical transportation data. Biosensors can provide representative granular onsite non falsifiable new data. We propose a framework to incorporate biosensors data into transportation modeling. We illustrate and test the framework with Monte Carlo and prototype field experiment. |
S0968090X20306288 | Disastrous events have been drastically increasing both in frequency and destructive capacity over the past few years . While advance notice events have received a great deal of attention in the literature of disaster management not much attention so far has been given to the no notice events mainly because of the scarcity of data . As an attempt to address this critical gap the current study proposes a disaggregate evacuation demand framework to understand evacuees travel behavior in case of no notice emergency events . The proposed framework comprises four main steps of evacuation decision evacuation planning tour formation and activity schedule update . This article is dedicated to the introduction of the framework structure and elaboration on the tour formation step . In this step we first estimate the total number of intermediate stops travel time and distance of the evacuation tours for those who decide to evacuate through a joint modeling structure and then determine the type of each intermediate stop . It is found that a broad range of factors including evacuees demographic profiles built environment attributes and characteristics of the disastrous event plays a significant role in peoples evacuation behavior during no notice emergency events . The findings of this study can assist responsible agencies in understanding evacuees complex behavior and consequently in devising effective strategies to alleviate economic damages and casualties resulted by such events . | An evacuation demand framework for no notice emergency events is introduced. The steps are evacuation decision planning tour formation and schedule update. The focus of this article is on the evacuation tour formation step. Tour travel time distance and number of stops are estimated in a joint structure. |
S0968090X2030629X | Network traffic congestion is known to be partially caused by vehicles cruising for parking . In this paper we quantify and assess the effect of cruising for parking by developing a macroscopic parking dynamics model for a parking dense neighborhood with limited parking supply where cruising for parking is explicitly considered in conjunction with the interactions between on and off street parking . The model is mainly built upon the system dynamics of different families of vehicles in the neighborhood which is governed by mass conservation equations utilizing the concept of macroscopic or network fundamental diagram . To reduce parking congestion and improve the overall system performance two real time parking pricing strategies are developed and integrated with the parking model a feedback based reactive pricing strategy driven by the parking occupancy and a model based predictive or proactive pricing strategy that explicitly aims to minimize the expected aggregate cruising delay . Extensive numerical experiments have been conducted to compare the performance of the two strategies applied to both on and off street parking . The results provide new insights into how a parking system shall be better managed with key implications for policy making summarized . | A field survey was conducted on the relationship between the parking occupancy and the cruising time. A macroscopic parking dynamics for both on and off street parking operations with limited capacity. Information provision is considered with different levels of driver compliance. Two real time parking pricing strategies are developed and compared. |
S0968090X20306306 | This article develops a deep reinforcement learning framework for dynamic pricing on managed lanes with multiple access locations and heterogeneity in travelers value of time origin and destination . This framework relaxes assumptions in the literature by considering multiple origins and destinations multiple access locations to the managed lane | Dynamic pricing formulated as a partially observable Markov decision process. Deep reinforcement learning Deep RL algorithms used as solution methods. Deep RL algorithms outperform feedback control heuristic on different objectives. Policies trained using Deep RL algorithms transfer well to new input distributions. Deep RL algorithms suitable for multi objective optimization using reward shaping. |
S0968090X20306318 | We present an approach to assess the risk taken by on road vehicles within the framework of artificial field theory envisioned for safety analysis and design of driving support automation applications . Here any obstacle to the subject vehicle is treated as a finite scalar risk field that is formulated in the predicted configuration space of the subject vehicle . The driving risk estimate is the strength of the risk field at the subject vehicles future location . This risk field is formulated as the product of two factors collision probability and expected crash energy . The collision probability with neighboring vehicles is estimated based on probabilistic motion predictions . The risk can be assessed for a single time step or over multiple future time steps depending on the required temporal resolution of the estimates . We verified the single step approach in three near crash situations from a naturalistic dataset and in cut in and hard braking scenarios with simulation and showed the application of the multi step approach in selecting the safest path in a lane drop section . The risk descriptions from the proposed approach qualitatively reflect the narration of the situation and are in general consistent with Time To Collision . Compared to current surrogate measures of safety the proposed risk estimate provides a better basis to assess the driving safety of an individual vehicle by considering the uncertainty over the future ambient traffic state and magnitude of expected crash consequences . The proposed driving risk model can be used as a component of intelligent vehicle safety applications and as a comprehensive surrogate measure for assessing traffic safety . | A probabilistic approach for highway driving risk assessment. Static objects formulated as potential fields and surrouding vehicles as kinematic fields. Uncertainties of surrounding vehicles captured by longitudinal and lateral acceleration distributions. Model validated with empirical data and simulation. Potential applications in safety impact assessment and online trajectory planning of intelligent vehicle systems. |
S0968090X2030632X | Vehicle to vehicle communication enabled cooperation of multiple connected vehicles improves the safety and efficiency of our transportation systems . However the joining and leaving of vehicles and unreliability of wireless communication channels will cause the switching of communication topology among vehicles thus affecting the performance of multi vehicle systems . To address this issue a distributed model predictive control method is proposed for multi vehicle system control under switching communication topologies . First an open loop optimization problem is formulated within which neighbor deviation and self deviation penalties and constraints are incorporated to ensure stability . Then a DMPC algorithm is designed for multi vehicle systems subject to switching communication topologies . For the closed loop system the convergence of predicted terminal states is proved based on the neighbor deviation constraint . After that closed loop system stability is analysed based on a common Lyapunov function defined using a joint neighbor set . It is proved that asymptotic stability of the closed loop system can be achieved through a sufficient condition on the weight matrices of the open loop optimization problem . Numerical simulations are conducted to demonstrate the effectiveness of the proposed DMPC controller . | A DMPC platoon controller is proposed to address switching communication topology. The convergence of predicted terminal states is strictly proved. A sufficient asymptotic stability condition on weight matrices is derived. |
S0968090X20306409 | Recent advances in the network level traffic flow modelling provide an efficient tool for analyzing traffic performance of large scale networks . A relationship between density and flow at the network level is developed and widely studied namely the macroscopic fundamental diagram . Nevertheless few empirical studies have been dedicated on the empirical evidence on the properties of the MFD for multiple modes of transport and to the best knowledge not yet at the scale of a megacity . This work combines rich but incomplete data from multiple sources to investigate the vehicle and passenger MFDs for cars and buses in the road network of Shenzhen . A novel algorithm is proposed for partitioning bimodal network considering the homogeneous distribution of link level car speeds and bus speeds . | Combining empirical data from multiple sources to estimate bimodal MFDs. New partitioning algorithm of bimodal network is proposed. Bus passenger density is estimated by fusing smart card and GPS data. Bus pMFD can experience double hysteresis loops even if bus vMFD does not. The three dimension vehicle and passenger MFDs of real data are analyzed. |
S0968090X20306410 | Urban transportation systems are subject to a high level of variation and fluctuation in demand over the day . When this variation and fluctuation are observed in both time and space it is crucial to develop line plans that are responsive to demand . A multi period line planning approach that considers a changing demand during the planning horizon is proposed . If such systems are also subject to limitations of resources a dynamic transfer of resources from one line to another throughout the planning horizon should also be considered . A mathematical modelling framework is developed to solve the line planning problem with a cost oriented approach considering transfer of resources during a finite length planning horizon of multiple periods . We use real life public transportation network data for our computational results . We analyze whether or not multi period solutions outperform single period solutions in terms of feasibility and relevant costs . The importance of demand variation on multi period solutions is investigated . We evaluate the impact of resource transfer constraints on the effectiveness of solutions . We also study the effect of period lengths along with the problem parameters that are significant for and sensitive to the optimality of solutions . | Line planning with a multi period planning horizon takes into account the changes of demand in time. Line plans with a multi period approach outperform the plans combining single period solutions. Inter period constraints for multi period line planning integrate resource allocation decisions. Necessity for resource transfer constraints is implied by the tightness of resource levels. The choice of period length affects solution accuracyand effectiveness in resource utilization. |
S0968090X20306422 | Crowding is one of the most common problems for public transportation systems worldwide and extreme crowding can lead to passengers being left behind when they are unable to board the first arriving bus or train . This paper combines existing data sources with an emerging technology for object detection to estimate the number of passengers that are left behind on subway platforms . The methodology proposed in this study has been developed and applied to the subway in Boston Massachusetts . Trains are not currently equipped with automated passenger counters and farecard data is only collected on entry to the system . An analysis of crowding from inferred origindestination data was used to identify stations with high likelihood of passengers being left behind during peak hours . Results from North Station during afternoon peak hours are presented here . Image processing and object detection software was used to count the number of passengers that were left behind on station platforms from surveillance video feeds . Automatically counted passengers and train operations data were used to develop logistic regression models that were calibrated to manual counts of left behind passengers on a typical weekday with normal operating conditions . The models were validated against manual counts of left behind passengers on a separate day with normal operations . The results show that by fusing passenger counts from video with train operations data the number of passengers left behind during a days rush period can be estimated within | Measuring crowding through archived and real time data using object detection tools. Object detection tools with surveillance video to quantify transit performance. Logistic regression models with automatic passenger counts and train operations data. Estimation of left behind passengers at stations with farecard use only at entrance. Estimation and evaluation of waiting time as a transit service reliability measure. |
S0968090X20306446 | Shipping is one of the major transportation approaches around the world . With the growing demands for global shipping service vessel destination prediction has shown its significant role in improving the efficiency of decision making in industry and ensuring a safe and efficient maritime traffic environment . Currently most vessel destination prediction methods focus on regional destination prediction which has restrictions on destinations and regions . Thus this paper proposes a general AIS data driven model for vessel destination prediction . In this random forest based model the similarity between the vessels traveling and historical trajectories are measured and utilized to predict the destination . The destination of the historical trajectory which shares the highest similarity with the traveling trajectory is predicted as the vessels destination . The method is different from previous work which used maritime records as input to predict the destination . In our method a historical trajectory database was generated from more than 141 million AIS records which covers 534 824 traveling patterns between ports and more than 5.9 million historical trajectories . Comparative studies were carried out to validate the performance of the proposed model where eleven state of the art trajectories similarity measurement methods combined with two different decision strategies were implemented and compared . The experimental results demonstrate that the proposed model combined with the port frequency based decision strategy achieves the best prediction accuracy on 35 937 testing trajectories . | Create a database from more than 141 million AIS records. Propose a random forest based model to measure the similarity between trajectories. Develop a decision strategy for vessel destination prediction. |
S0968090X20306458 | With the advent of intelligent transportation systems spatiotemporal traffic data has gained growing importance in real time monitoring prediction and control of traffic . However in practical implementations data collection devices are often faced with malfunctions caused by various unpredictable disruptions thereby resulting in the so called missing value problems . In realistic cases the disruptions to the data collection devices are often associated with some key events in addition along with other disruptions the missing value problem could be in a complicated manner with both randomly and completely missing patterns . To perform the imputation task with such complicated missing patterns we propose a hybrid spatiotemporal method which utilizes the time series properties by prophet model and captures the spatial residuals information by iterative random forest model . The spatiotemporal method first applies the temporal part to fill the missing value and then adopts the spatial part to acquire the residual component of the missing values . The results of the two components are integrated into the final imputations . Based on the PeMS freeway dataset and an urban road dataset under extensive artificially designed scenarios like randomly clustered non completely and completely missing patterns we test our proposed approach with some existing techniques such as K Nearest Neighbor Seasonal Trend decomposition using Loess Bayesian tensor decomposition Denoising AutoEncoder . The test results indicate that the hybrid method achieves the best imputation quality for most missing patterns particularly for those with completely or hybrid missing patterns . Furthermore the hybrid model still performs well under extreme missing rates as high as 0.9 which validates the robustness of the model in extreme situations . | A hybrid spatiotemporal approach for traffic missing data imputation. Tackling multiple missing patterns including randomly clustered non completely and clustered completely missing. Extensive numerical tests with real world dataset. Applicable even when the missing rate approaches 90 . |
S0968090X2030646X | The intercontinental liner shipping services transport containers between two continents and they are crucial for the profitability of a global liner shipping company . In the daily operations of an intercontinental liner shipping service however container slot bookings from customers can be freely cancelled during a booking period which causes loss of revenue and low utilization of ship capacity . Though a pain point of the liner shipping industry the container slot cancellation problem has not yet been well investigated in the literature . To fill this research gap this study aims to estimate the probability for the cancellation of container slot booking in the long haul transports of the intercontinental liner shipping service by considering the primary influential factors of cancellation behavior . To achieve the objective a container slot booking data driven model is developed by means of a time to event modeling technique . To incorporate the effect of booking region on the cancellation probability we introduce the frailty term in the model to capture the regionality of the container shipping market . Our case study with real slot booking data shows that the developed model performs well in forecasting the loaded containers of the slot booking requests . In addition we shed light on how the internal factors of slot booking and external factors of shipping market influence the probability of cancellation . | First study on the cancellation probability of slot booking. Developing a model to estimate cancellation probability. Introducing frailty term to reflect rationality of market. Discussion about the influential factors of cancellation. |
S0968090X20306471 | This study aims to enrich autonomous vehicle adoption research and practice by being the first study to systematically review empirical studies on behavioural intention to use AVs a key element in the adoption process . This review of the extant literature provides a synthesized overview of the current state of knowledge develops a | Hedonic motivation is the strongest predictor of autonomous vehicle use intention. Personal information technology innovativeness moderates AV adoption drivers. Utilitarian motivation increases AV adoption intent among innovative users only. Overall technological anxiety reduces intention to adopt AVs among laggard users. Data privacy concerns reduces intention to adopt AVs among innovative users. |
S0968090X20306483 | Recent research on airport ground movement introduced an Active Routing framework to support multi objective trajectory based operations . This results in edges in the airport taxiway graph having multiple costs such as taxi time fuel consumption and emissions . In such a graph multiple edges exist between two nodes reflecting different trade offs among the multiple costs . Aircraft will have to choose the most efficient edge from multiple edges in order to traverse from one node to another respecting various operational constraints . In this paper we introduce a multi objective routing and scheduling algorithm based on the enumerative approach that can be used to solve such a multi objective multi graph problem . Results using the proposed algorithm for a range of international airports are presented . Compared with other routing and scheduling algorithms the proposed algorithm can find a representative set of optimal or near optimal solutions in a single run when the sequence of aircraft is fixed . In order to accelerate the search heuristic functions and a preference based approach are introduced . We analyse the performance of different approaches and discuss how the structure of the multi graph affects computational complexity and quality of solutions . | A multi objective routing and scheduling algorithm is proposed for taxiing aircraft. The problem is modelled using a multi objective multi graph. Heuristic functions and a preference based approach are introduced. The multi graph structure is analysed w.r.t. complexity and quality of solutions. |
S0968090X20306495 | Flight arrival and departure scheduling problem one of the critical tasks in terminal air traffic operation faces new challenges as the terminal air traffic management has transformed to adapt to the performance based navigation environment beside of that the terminal system uncertainties which are usually due to the time varying interference of the convective weather and flight arrival time will also exacerbate the difficulty to realize an efficient terminal air traffic operation . In order to effectively address the above issues we propose the formulation and solution approach for a stochastic terminal FAADS problem under PBN environment . The proposed FAADS problem formulation combines the flexible 4 dimensional trajectory requirement under PBN environment and the stochastic quantifications both from the convective weather and flight arrival time uncertainties which can finally make the FAADS results realize the avoidance of the convective weather and immunity of the flight arrival time variations within tolerable risk probabilities . We provide an efficient solution approach to tackle this problem with complex mixed integer and nonlinear programming basics and illustrate the capabilities of the solution approach by a test case on real terminal system in Shanghai Metroplex . Numerical results demonstrate the effectiveness of proposed problem formulation and solution approach . | A flight arrival and departure scheduling problem is considered. A stochastic optimization model with weather and flight arrival time uncertainties. A 4 dimensional trajectory complying with the PBN requirements. |
S0968090X20306501 | With the significant improvements in drone technology and the popularization of drones among their hobbyists the incidents of drones intruding airports have resulted in a large number of flight delays and temporary closure of runways . To minimize the interference of drones on normal operations in the airports a collision probability evaluation scheme based on collision course trajectories modeling is proposed in this work . Firstly a trajectory planning model of drones intruding restricted airspace is derived based on a given trajectory of the commercial aircraft and the collision course scenario of the drone . Subsequently according to the trajectories of the drone and CA a probabilistic model based on the stochastic kinematic model is developed to implement the collision risk evaluation . The proposed method is first comparatively demonstrated with the Monte Carlo simulation and several special cases with known drones trajectories . Subsequently the cases covering different drones initial positions positions updates and different collision zones are simulated and analyzed using the proposed collision course based model . The simulation results show that the established model can be employed to evaluate the collision probability even if the trajectory information of the intruding drone is limited . | The intruding drone trajectory is developed based on the collision course assumption. The relative position prediction model is introduced by considering uncertainties. Model validation by Monte Carlo simulation and in several cases. Parametric test results demonstrate the sensitivity and adaptability of the model. |
S0968090X20306513 | A large amount of data is produced every day by stakeholders of the Air Traffic Management system in particular airline operators airports and air navigation service providers . Most data is kept private for many reasons including commercial and security concerns . More than data shared information is precious as it leverages intelligent decision making support tools designed to smoothen daily operations . | Detecting significant events in large volumes of trajectory data. Understanding sources of deviation in air traffic management. Dimension reduction may create clusters in the latent space. |
S0968090X20306525 | A large portion of passenger requests is reportedly unserviced partially due to vacant for hire drivers cruising behavior during the passenger seeking process . This paper aims to model the multi driver repositioning task through a mean field multi agent reinforcement learning approach that captures competition among multiple agents . Because the direct application of MARL to the multi driver system under a given reward mechanism will likely yield a suboptimal equilibrium due to the selfishness of drivers this study proposes a reward design scheme with which a more desired equilibrium can be reached . To effectively solve the bilevel optimization problem with upper level as the reward design and the lower level as a multi agent system a Bayesian optimization algorithm is adopted to speed up the learning process . We then apply the bilevel optimization model to two case studies namely e hailing driver repositioning under service charge and multiclass taxi driver repositioning under NYC congestion pricing . In the first case study the model is validated by the agreement between the derived optimal control from BO and that from an analytical solution . With a simple piecewise linear service charge the objective of the e hailing platform can be increased by | Multi agent reinforcement learning is applied to driver repositioning contexts. A multi class MARL is applied to a case study using NYC taxi data. The first to propose an optimal reward design to guide drivers selfish behavior towards social optima. |
S0968090X20306537 | Cycle based traffic volume estimation is important for the dynamic evaluation and optimization of signal control schemes at intersections . With the development of intelligent mobility and connected vehicle technologies massive high resolution vehicle trajectory data have become available which can provide a rich source of information for estimating traffic flow parameters at signalized intersections . Existing methods for traffic volume estimation using sampled vehicle trajectories are commonly driven by shockwave or probabilistic models in which the traffic volume estimation is usually modelled as a parameter estimation problem . These methods commonly require certain assumptions for the vehicle arrival distribution and a First In First Out queuing rule . However these model driven methods have limitations when the vehicle arrival distribution is unknown or when there is more than one lane for the same movement at an intersection approach as the FIFO rule is no longer true . In addition the accuracy and stability of these models remain challenging under low penetration rates . In this paper we propose a tensor decomposition method to estimate cycle based traffic volume at signalized intersections using sampled vehicle trajectories . Unlike the existing model driven methods the proposed method is purely data driven and does not rely on any prerequisite assumptions for the arrival distributions and queuing rules . In the proposed method the traffic volume of each cycle is first divided into a known part and an unknown part based on the queuing position of the last queued sample vehicle . Then these two parts of the volume as a whole along with two relevant traffic observations are then integrated into a three dimensional tensor . This tensor can effectively preserve the temporal correlations between adjacent cycles and the interrelationships among the traffic observations . Finally the problem of cycle based traffic estimation is transformed into a tensor completion problem and the Tucker decomposition method is adopted to solve the completion problem . The proposed method is evaluated using both empirical and simulation data . The results indicate that the proposed method is capable of producing accurate and reliable estimates for cycle based traffic volumes even under low penetration rates i.e . less than 5 . | A novel data driven method for cycle based volume estimation using sampled vehicle trajectories. Our method does not rely on any prerequisite assumptions for the arrival distributions and queuing rules. Our method can effectively adapt to the time series fluctuations of traffic flow. Our method is capable of producing accurate and reliable estimates even under low penetration rates. Our method outperforms two other existing methods. |
S0968090X20306549 | The real time crash risk analyses were proposed to establish the relationships between crash occurrence probability and pre crash traffic operational conditions . Given its great application potentials that link with Active Traffic Management System for proactive safety management it has become an important research area . Currently researchers mainly developed the real time crash risk analysis models with traffic flow descriptive statistics employed as explanatory variables and with re sampled balanced dataset which hold the limitations of insufficiently capturing the temporal spatial traffic flow characteristics and failing to provide classification capabilities when deal with the imbalanced datasets . In this study a Convolutional Neural Network modelling approach with refined loss functions has been first time introduced to the real time crash risk analyses . The primary objectives of the proposed CNN models are utilizing the tensor based data structure to explore the multi dimensional temporal spatial correlated pre crash operational features and optimizing the loss functions to overcome the low classification accuracy issue brought by the imbalanced data . Data from the Shanghai urban expressway system were utilized for the empirical analysis . And a total of three types of loss functions including traditional binary cross entropy the weighted cross entropy and the focal loss were introduced and being tested with varying ratios of crash and non crash datasets . The modeling results show that the CNN model has better classification performance compared to the traditional Multi layer Perceptrons model with the tensor based structure data . Besides the developed CNN model with focal loss function has substantial classification enhancement under the imbalanced datasets . Finally the | Developed a tensor based structure to represent crash precursors. Applied CNN model for multi dimensional traffic flow features extraction. Proposed refined focal loss functions for the imbalanced data issue. Obtained 66.8 sensitivity with 3.8 FAR result under 1 100 imbalanced ratio. |
S0968090X20306550 | The current study shows the impact of distraction in designing the intelligent in vehicle systems for assisting the drivers to make stop cross decision at the onset of yellow signal . In total 83 participants drove through a simulated urban scenario with six signalised intersections . Firstly the time taken to execute the stop cross decision was statistically modelled by using Weibull AFT models . The results showed that compared to the baseline eating and drinking tasks reduced the stopping time by 6 and 7 respectively . For the crossing encounters the eating task caused 12 increment in crossing time compared to the baseline . This analysis was followed by modelling the success rates of the executed decisions with binary logistic models . The success rates of the stopping decision showed that reduction in the time lapsed in executing the decision led to failure in stopping the vehicle before the stop line . Similarly an increment in the time taken to execute the crossing decision led to reduced probability of successfully crossing the stop line within the yellow duration . Hence the results suggest that the design of the smart assistance system for decision making at the onset of yellow signal should be based on the success rate of the decision which is dependent on the time lapsed in executing the decision . Moreover the presence of distracting activities should be considered as an input parameter while designing the assistance system . | Crash risk at unsignalised intersections increased during eating and drinking tasks. No compensatory measure was adapted for eating and drinking tasks while crossing intersections. Texting task resulted in deteriorated situation awareness while approaching the intersections. Drivers who drove frequently kept larger safety margins while crossing the intersections. |
S0968090X20306562 | In the paper a routing algorithm of two sub algorithms is proposed for consideration of tactical decisions in a microscopic crowd simulation . One sub algorithm works in the high or macroscopic level and is used to determine the departure door which would suffer a later change whenever pedestrians firstly enter rooms . The sub algorithm is based on a tailored Dijkstra shortest path algorithm with introduction of the concept of impedance conduction for considering congestion along paths . The other sub algorithm works in the low or microscopic level and is supposed to work with any chosen microscopic crowd dynamics model . It is mainly used to model dynamic changes of pedestrians departure doors while they are walking inside rooms . Many mechanisms are introduced for considering practical factors affecting pedestrians routing behaviors for example consideration of congestion of segments between pedestrians and doors consideration of impact of re selection of previous doors etc . According to conducted numerical experiments the proposed algorithm can well model crowds tactical decisions in a microscopic simulation under various scenarios thus is a suitable method for study of crowd dynamics in a complicated environment . | A new crowd routing algorithm is proposed. The concept of impedance conduction is invented for an even distribution of pedestrians. An integrated framework is proposed for a systematic consideration of various factors affecting pedestrian routing behaviors. A new approach for modeling crowd in a single room of quite obstacles is proposed. |
S0968090X20306574 | The Macroscopic Fundamental Diagram which exhibits the relationship between average flow and average density of an urban network is a promising framework for monitoring and controlling urban traffic networks . Given that monitoring resources are limited in real world networks acquiring adequate data to estimate an MFD is of crucial importance . This study presents a novel network wide approach to identifying critical links and estimating average traffic flow and density . The proposed model estimates the MFD using flow and density measurements from those critical links which constitute only a small subset of all the links in the network . To find the critical links we rely on historical probe vehicle data and propose a model that builds on Principal Component Analysis a dimensionality reduction and a feature selection method . Essentially using PCA a large number of possibly interrelated variables in a dataset can be reduced to a set of smaller uncorrelated variables while maintaining as much information as possible in the dataset . The resulting uncorrelated variables or the principal components indicate the major patterns or the dominating features of the original dataset . Additionally PCA enables the reconstruction of the full scale dataset from the selected features . In this work we apply PCA in order to identify the main traffic features from a probe vehicle dataset then we find the links that are associated with these features then we locate loop detectors on those links to collect flow and density data and finally we reconstruct the full scale data building on the PCA mechanism . This gives us the flow and density of all links from which we can effectively estimate the MFD . | Developing a method based on Principal Component Analysis PCA to identify critical links. Reconstructing network scale traffic measurements from critical links. Estimating Macroscopic Fundamental Diagram MFD for large scale urban networks using only critical links. |
S0968090X20306586 | The mixed traffic flow has complex dynamics by nature . The kinematic differences between automobiles and motorcycles result to distinct driving behaviors . Traditional automobile based traffic flow theory is not always suitable for mixed traffic streams . The purpose of this study is to observe from actual data a clearance boundary called Safety Space drivers maintain from other vehicles and use it as a spatial filter to determine conflicts in mixed traffic flows . Image data are collected from an Unmanned Aerial Vehicle and microscopic characteristics such as vehicle type position velocity and trajectory are extracted through computer vision techniques . The Histogram of Oriented Gradients feature and the Support Vector Machine classifier are utilized for the vehicle detection while the Kalman Filter is employed for the derivation of vehicle trajectories . The Safety Space is then determined based on those trajectories . Validation data are collected at intersections in Taipei Taiwan Bangkok Thailand and Mumbai India . The vehicle detection and tracking are satisfactory and the Safety Space surrogate reveals risk zones caused by spatial proximity between vehicles . | Image traffic data are collected from the UAV and trajectories are extracted through object based detection and tracking. The Safety Space boundary in proximity of the vehicle is observed from data. This Safety Space forms the boundary of the comfort area drivers perceive as safe when traveling. Invasions of the Safety Space by other vehicles potentially reveal traffic conflicts. |
S0968090X20306598 | Understanding the impact of aircraft speed heterogeneity on air traffic operation is crucial for airspace design and air traffic flow management . Speed heterogeneity is recognized as a causal factor for complexity in air traffic operations through qualitative or statistical analysis . Quantitative metrics on how it affects current and future Trajectory Based Operation is lacking however . In this paper we present an in depth investigation of the impact of speed heterogeneity by defining air traffic robustness at microscopic and macroscopic levels in nominal situations derived mathematically and validated through fast time computer simulations . A human in the loop study follows investigating six 4D en route operation scenarios where operators were instructed to resolve a large disturbance in a sector using a novel interface . Results confirm the negative impact of speed heterogeneity on air traffic controller performance in terms of flow efficiency and workload . The mechanism of such impact is substantiated through analyzing several speed based robustness metrics . Although the simulated traffic scenarios have similar baseline robustness those with mixed speeds lead to significantly lower robustness and operational performance . This emphasizes the need to incorporate speed heterogeneity in robustness evaluations of air traffic control in current and future TBO environments . | The aircraft robustness with heterogeneous speeds in normal operations is analyzed. The negative impact of speed mix in contingency is validated through HITL experiments. A speed based robustness measure is proposed to interpret the underlying mechanisms. The individual differences in performing cooperative conflict resolutions are discussed. |
S0968090X20306604 | Transportation networks are unprecedentedly complex with heterogeneous vehicular flow . Conventionally vehicles are classified by size the number of axles or engine types e.g . standard passenger cars versus trucks . However vehicle flow heterogeneity stems from many other aspects in general e.g . ride sourcing vehicles versus personal vehicles human driven vehicles versus connected and automated vehicles . Provided with some observations of vehicular flow for each class in a large scale transportation network how to estimate the multi class spatio temporal vehicular flow in terms of time varying Origin Destination demand and path link flow remains a big challenge . This paper presents a solution framework for multi class dynamic OD demand estimation in large scale networks that work for any vehicular data in general . The proposed framework cast the standard OD estimation methods into a computational graph with tensor representations of spatio temporal flow and all intermediate features involved in the MCDODE formulation . A forward backward algorithm is proposed to efficiently solve the MCDODE formulation on computational graphs . In addition we propose a novel concept of tree based cumulative curves to compute the exact multi class Dynamic Assignment Ratio matrix . A Growing Tree algorithm is developed to construct tree based cumulative curves . The proposed framework is examined on a small network a mid size network as well as a real world large scale network . The experiment results indicate that the proposed framework is compelling satisfactory and computationally plausible . | A general framework for estimating multi class dynamic O D demand. A computational graph formulation with efficient forward backward solution algorithm. A tree based cumulative curves method to estimate demand gradients. State of the art deep learning techniques can be adopted in this framework. The methods are examined in large scale real world networks. |
S0968090X20306616 | Autonomous driving technologies are advancing rapidly and determining when consumers ride in driverless vehicles on a daily basis is becoming essential . Using choice experiments first we elicit consumers willingness to pay for autonomous driving systems in Japan along with their WTP for hybrid and electric engines . We found in this study that the Japanese consumers WTP is on average lower than the estimates for the US market and is not sufficient to enable autonomous vehicles to capture a meaningful share of the existing car market . Second compared with a previous study in the US we propose and discuss an expected social dilemma caused by the ethical problems that AVs will face known as the Trolley Problems . We find that social dilemma may occur because people tend not want to buy moral AVs . Third we explored the determinants of the WTP and social dilemma to find that the credibility of AVs is determined to be a critical factor for the social dilemma . | WTP in Japan for an auto driving system is estimated insufficient in car markets. A social dilemma will occur people may not buy AVs that they consider moral. Credibility regarding autonomous vehicles plays a key role in the social dilemma. |
S0968090X20306628 | Rail transit delays are generally discussed in terms of on time performance or problems at individual stops . Such stop scale approaches ignore the fact that delays are also caused and perpetuated by network wide factors . The objective of this paper is to develop a network model and metrics that can quantify the delay dependencies between transit network stops and identify local sources of network wide issues . For this purpose Bayesian network learning was utilized . Based on the calculated Bayesian networks network metrics | Bayesian network was used to identify delay interactions of stops of complex transit system. Metrics proposed to provide a tangible index for quantifying the transit stops. Crowdsourced data gathered from a real time transit information app onTime. Study showed capability of proposed approaches in identifying network wide problems of transit systems. |
S0968090X2030663X | Railway capacity is a scarce and expensive resource that has to be utilised in the best way possible . Many methods exist for medium and short term planning and analysis of railway capacity . Yet there is a lack of methods for long term planning that provides high quality estimates of capacity in complex railway networks . In this paper we define capacity of a railway network by a maximal set of trains that can be operated in a predefined period . This maximal set distinguish between different train types and the number of each type and thus the heterogeneity of the operation . | New approach determines the maximal set of trains to be operated on a network. Heterogeneous train operation is handled without the need of a fixed timetable. Branch and bound tabu search used to solve defined capacity determination problem. Ideal for early planning as little input is required while maintaining precision. Case study on Danish rail network reveals capacity improvements from upgrades. |
S0968090X20306719 | Traffic conflict points cause travel delay stop and go traffic and excessive energy consumption . Efforts have been taken to improve traffic conflict point performance via trajectory control of connected automated vehicles as the CAV technology emerges . One major challenge to these efforts is the complexity in optimization of CAV trajectories particularly with joint signal timing optimization . This challenge poses barriers to real time application requirements scaling them up to address network level problems and drawing analytical insights into problem structures . To overcome this challenge this paper aims to seek for an efficient and analytical solution to a joint vehicle trajectory and signal timing optimization problem . This problem simultaneously optimizes CAV trajectories and signal timing to minimize travel delay and energy consumption at a conflicting point with two traffic approaches . This study modifies the original complex formulation in two ways . First the vehicle trajectory shape is simplified into a piece wise quadratic function with no more than five segments . Second instead of using the highly non linear instantaneous fuel consumption function a simplified macroscopic measure is proposed to approximate fuel consumption as an analytical quadratic function of signal red interval . These simplifications provide elegant theoretical properties that enable solving an analytical exact solution to this complex problem with parsimonious analytical insights . Numerical examples reveal that the proposed model can significantly reduce travel delay and fuel consumption . Moreover it is demonstrated that the presented algorithm is highly efficient and appropriate for real world traffic applications . | Proposing an analytical near optimum solution approach to a joint vehicle trajectory and signal timing optimization problem. Analytical construction of near optimum multi trajectories. Considering macroscopic fuel consumption function consideration. Proposing theoretical properties into problem structure and analytical solution formulation. Conducting numerical examples for multiple types of conflict points. |
S0968090X20306732 | The bus bridging service has always faced problems in severe emergencies or catastrophes that require the large scale evacuation of passengers . This paper provides an alternative evacuation scheme which uses the urban bus network in the case of common metro service disruptions this is modeled by minimizing the total cost of the affected metro passengers through jointly selecting the bus lines and frequencies . The uncertain recovery time of the service disruption and the heterogeneous risk taking behavior of the affected metro passengers are incorporated in the scheme . Therefore we build a linkage between the evacuation service design and the risk taking behavior of passengers . A heuristic algorithm is proposed to calculate the optimal evacuation scheme . A numerical experiment using a real world network is conducted to illustrate the validity of the model and algorithm . | A coordination evacuation scheme under common metro disruption was proposed. Behavior of the metro passengers under uncertain recovery time was considered. Bus evacuation network and bus frequencies were jointly optimized. A heuristic algorithm for a general metro and urban bus network was designed. |
S0968090X20306744 | The interest in using drones in various applications has grown significantly in recent years . The reasons are related to the continuous advances in technology especially the advent of fast microprocessors which support intelligent autonomous control of several systems . Photography construction and monitoring and surveillance are only some of the areas in which the use of drones is becoming common . Among these last mile delivery is one of the most promising areas . In this work we focus on routing problems with drones mostly in the context of parcel delivery . We survey and classify the existing works and we provide perspectives for future research . | We present a review of the recent contributions on drone aided routing problems. We provide a description of the technological background. We focus on problems arising in parcel delivery and provide a new classification. We give some insights into current research trends. We outline possible future research directions. |
S0968090X20306756 | Obtaining accurate information about future traffic flows of all links in a traffic network is of great importance for traffic management and control applications . This research studies two particular problems in traffic forecasting capture the dynamic and non local spatial correlation between traffic links and model the dynamics of temporal dependency for accurate multiple steps ahead predictions . To address these issues we propose a deep learning framework named | A spatiotemporal deep learning model for multi step ahead traffic forecasting. Pattern aware adjacency matrix defined for dynamic non local spatial correlation capture. CNN RNN coupled seq2seq architecture with attention for temporal dependency modeling. Extensive evaluation on two large scale publicly available datasets. Quantitative and qualitative analysis for validation and interpretation of the model. |
S0968090X20306768 | The rapid conceptual development and commercialization of connected automated vehicle has led to the problem of mixed traffic i.e . traffic mixed with CAVs and conventional human operated vehicles . The paper studies cooperative decision making for mixed traffic . Using discrete optimization a CDMMT mechanism is developed to facilitate ramp merging and to properly capture the cooperative and non cooperative behaviors in mixed traffic . The CDMMT mechanism can be described as a bi level optimization program in which state constrained optimal control based trajectory design problems are imbedded in a sequencing problem . A bi level dynamic programming based solution approach is developed to efficiently solve the problem . The proposed modeling mechanism and solution approach are generic to deterministic decisions and can guarantee system efficient solutions . A micro simulation environment is built for model validation and analysis of mixed traffic . The results show that compared to the scenario with 100 HVs ramp merging can be smoother in mixed traffic environment . At high CAV penetration the section throughput increases about 18 . With the proposed CDMMT mechanism traffic throughput can be further increased by 1015 . The proposed methods form the basis of traffic analysis and cooperative control at ramp merging sections under mixed traffic environment . | We studies cooperative decision making for mixed traffic at ramp merging sections. Cooperative and non cooperative microscopic decisions are explicitly considered. The problem is solved using a bi level dynamic programming based approach. The proposed CDMMT Ramp Merging method guarantees system efficient results. |
S0968090X2030677X | Connected and automated technologies for vehicles pave the way for major changes in traffic control methodology . A decentralized control system based on a communication network using vehicle to vehicle communications and each vehicles on board controller will be one of the predominant approaches to vehicle trajectory optimization when the deployment of connected and automated vehicles has advanced because of its scalability and fault resistance compared to the centralized system . In this study we propose a dynamic game based vehicle trajectory optimization approach that can be utilized in a decentralized control system based on CAV technologies for merging segments which is one of the bottlenecks of highways . Since the decentralized system is constructed by equal individual vehicles that decide their own optimal control in conflicting situations we introduce a game theory based approach for obtaining a satisfactory result for those vehicles . Further for achieving the optimal result at the end of the target segment we employ a dynamic game to design the whole interrelated time series trajectories of competing pairs of vehicles . To tackle the curse of dimensionality we use an efficient method to enumerate the sets of combinations a zero suppressed binary decision diagram and propose an algorithm that uses this ZDD to solve the dynamic game . The performance of the proposed approach is demonstrated via numerical experiments and is compared to a static game based model and an individual dynamic decision model . The results prove the effectiveness of the proposed approachit induces efficiently adjusted behaviors of competing vehicles and enables merging success especially in a situation where the initial gaps between vehicles are not enough for immediate lane change which is achieved by the combined effect of the game theoretic approach and the dynamic decision approach . | Proposing a utility based decentralized control framework for CAVs in merging areas. Formulating a dynamic game based merging optimization model. Optimizing the merging points and the longitudinal trajectories of vehicles. Developing an algorithm that uses a ZDD to solve the dynamic Stackelberg game. Illustrating the effect of game theory and dynamic decisions via numerical examples. |
S0968090X2030680X | Ridesourcing services provided by companies like Uber Lyft and Didi have grown rapidly over the past decade and now serve a sizable portion of trips in many metropolitan areas . An understanding of these services is critical for regulating planning and managing urban multi modal transportation systems effectively . Unfortunately little is known about ridesourcing travel because private companies providing ridesourcing services were not previously subject to data sharing requirements . Fortunately the city of Chicago recently collected and released spatially and temporally aggregated data on ridesourcing trips collected from private companies . This study analyzes the Chicago ridesourcing data to examine factors influencing ridesourcing usage . The study employs a random effects negative binomial regression approach to model ridesourcing usage . Determinants considered in the model include weekend vs. weekday and weather variables as well as census tract socio demographics and commute characteristics land use variables places of interest transit supply parking features and crime . The model results indicate ridesourcing demand is higher on days when temperatures are lower there is less precipitation and on the weekend as well as in census tracts with higher household incomes a higher percentage of workers who carpool or take transit to work a higher percentage of households with zero vehicles higher population and employment density higher land use diversity fewer parking spots and higher parking rates more restaurants and more homicides . The results also demonstrate a non linear relationship between ridesourcing demand and transit supply variables . The paper discusses the implications of these model results to inform transportation planning and policymaking as well as future research . | Ridesourcing services have a substantial and growing market share role in urban transport systems. Study analyzes Chicago ridesourcing trip data largest and newest public ridesourcing data in US. Ridesourcing usage higher in census tracts with fewer parking spots and higher parking rates. Restaurants population and employment density and land use diversity positively associated with ridesourcing usage. More bus stops associated with higher ridesourcing usage rail stations and ridesourcing usage have nonlinear relationship. |
S0968090X20306811 | Peer to peer ridesharing is a form of shared use mobility that has emerged in recent decades as a result of enabling of the sharing economy and the advancement of new technologies that allow for easy and fast communication between individuals . A P2P ridesharing system provides a platform to match a group of drivers who use their personal vehicles to travel with their peer riders who are in need of transportation . P2P ridesharing systems are traditionally categorized as two side markets with two mutually exclusive sets of agents i.e . riders and drivers . Fixing the roles of participants a priori however could come at an opportunity cost of missed social welfare revenue for the system . Consequently this paper proposes a new market game and its corresponding mathematical formulation that outputs matching role assignment and pricing . We investigate the stability properties of this market game and present a mathematical formulation that yields a stable outcome if one exists . Furthermore we propose a Lagrangian relaxation algorithm to obtain a stable solution for large scale games with empty cores through subsidizing the system . Using numerical experiments we demonstrate the benefits of the proposed methodology and its advantages over previously proposed methods for stabilizing non bipartite graphs . | P2P ridesharing market with role flexibility is modeled as a roommate problem. P2P ridesharing market with role flexibility may not admit a stable outcome. A minimum subsidy problem is proposed to stabilize the P2PRM. The minimum subsidy is bounded by the integrality gap from below. The minimum subsidy is bounded by the minimum blocking value from above. |
S0968090X20306823 | Sidewalks are a critical infrastructure to facilitate essential daily trips for pedestrian and wheelchair users . The dependence on the infrastructure and the increasing demand from these users press public transportation agencies for cost effective sidewalk maintenance and better Americans with Disabilities Act compliance . Unfortunately most of the agencies still rely on outdated sidewalk mapping data or manual survey results for their sidewalk management . In this study a network level sidewalk inventory method is proposed by efficiently segmenting the mobile light detection and ranging data using a customized deep neural network i.e . PointNet and followed by integrating a stripe based sidewalk extraction algorithm . By extracting the sidewalk locations from the mobile LiDAR point cloud the corresponding geometry features e.g . width grade cross slope etc . can be extracted for the ADA compliance and the overall condition assessment . The experimental test conducted on the entire State Route 9 Massachusetts has shown promising performance in terms of the accuracy for the sidewalk extraction value of 0.946 and the efficiency for network analysis of the ADA compliance . A case study conducted in Columbus District in Boston Massachusetts demonstrates that the proposed method can not only successfully support transportation agencies with an accurate and efficient means for network level sidewalk inventory but also support wheelchair users with accurate and comprehensive sidewalk inventory information for better navigation and route planning . | Develops a fully automated network level sidewalk inventory method that leverages the emerging mobile LiDAR and deep learning. Develops a reliable deep learning model based on PointNet for accurately and efficiently segmenting LiDAR point clouds. Develops a stripe based principal component analysis PCA for accurately extracting sidewalk and conducting the corresponding geometry measurement for ADA compliance. Advances the understanding and utilization of mobile LiDAR and deep learning in critical pedestrian infrastructure management and accessibility analysis. |
S0968090X20306847 | Automated Vehicles have gained substantial attention in recent years as the technology has matured . Researchers and policymakers envision that AV deployment will change transportation development patterns and other urban systems . Researchers have examined AVs and their potential impacts with two methods survey based studies of AV preferences and simulation based estimation of secondary impacts of varied AV deployment strategies such as Shared AVs and Privately owned AVs . While the preference survey literature can inform AV simulation studies preference study results have so far not been integrated into simulation based research . This lack of integration stems from the absence of data that measure preferences towards PAVs and SAVs at the neighborhood level . Existing preference studies usually investigate adoption likelihood without collecting appropriate information to link preferences to precise locations or neighborhoods . This study develops a microsimulation approach incorporating machine learning and population synthesizing to fill this data gap leveraging a national AV perception survey and the latest National Household Travel Survey data . The model is applied to San Francisco CA and Austin TX to test the concept . We validate the proposed model by comparing the spatial distributions of synthesized ride hailing users and observed ride hailing trips . High correlations between our synthesized user density and empirical trip distributions in two study areas to some extent verify our proposed modeling approach . | A generalizable modeling framework to map neighborhood preferences for AVs. Machine learning preferences for AVs. Impute and synthesize preferences for AVs. Framework validated using San Francisco and Austin data. |
S0968090X20306859 | Transport demand is highly dependent on supply especially for shared transport services where availability is often limited . As observed demand can not be higher than available supply historical transport data typically represents a biased or | Observability of mobility demand is inherently limited by supply. Censored regression applied to mobility demand to mitigate bias. Censored Gaussian Process formulated for time varying censorship. Experiments with synthetic and real world data demonstrate solution approach. Benefit of preserving the censored information is measured. |
S0968090X20306860 | Shared parking allows the effective use of undersupplied parking spaces and contributes to the alleviation of urban parking problems traffic congestion environmental pollution and other negative externalities of traffic . However little is known about the acceptance of shared parking by consumers of a different socio demographic profile . To understand the feasibility and potential success of shared parking this paper develops a stated choice experiment with three choice options fixed mode shared parking flexible mode shared parking and not interested to investigate parking space owners propensity to engage in shared parking under varying conditions . Because the demand for shared parking is uncertain the revenues owners may generate are uncertain . As one of the most popular theories of decision making under uncertainty the cumulative prospect theory is incorporated into a multinomial logit model to capture the decision problem in which some variables are uncertain and others are not . The revenue that owners expect shared parking can bring is used as the reference point to differentiate between gains and losses . Gains refer to outcomes that exceed the reference point while losses refer to outcomes that fall short . To examine unobserved heterogeneity a random parameter version of the model is specified to estimate the distribution of decision weights across the sample . Results show that socio demographic characteristics context variables revenues and psychological concerns are all important factors in explaining parking space owners propensity to engage in platform based shared parking schemes . Incorporating unobserved heterogeneous improves the overall goodness of fit of the model . Understanding parking space owners propensity to share their parking spaces in relation to their psychological concerns and uncertain conditions is critical to improve shared parking policies . The results of this paper may help designers and planners in the delivery of shared parking services and promote the success and future growth of the shared parking industry . | Investigated the willingness to share private parking spaces. Examined the choice between flexible and fixed modes of sharing parking places. Explored the decision making under uncertainty. Developed a hybrid random parameter logit cumulative prospect theoretic model. |
S0968090X20306872 | With the fast development of automated vehicle technologies scholars have proposed various innovative local traffic control schemes for more effective management of AV traffic especially at intersections . However due to computational intractability the investigation of network level AV control is still at the initial stage . This study proposes a space time routing framework applicable in dedicated AV zones . To relieve the computational load we establish a node based conflict point network to model realistic road networks and at each conflict point we record the space time occupations of AVs in continuous timelines . Then based on the conflict point network we develop two space time routing algorithms for each AV once it enters the dedicated AV zone to minimize its trip travel time while maintaining the non collision insurances these two algorithms can trade off between solution quality and computational load . Furthermore to enhance the network throughput for handling heavy traffic we develop a platoon strategy that forces AVs to pass through conflict points in platoons and we adopt Deep Q learning to optimize the platoon sizes at different spots dynamically . Numerical tests show that both proposed algorithms perform well in that they can execute the routing tasks with very limited computational time and the average vehicle delay approaches zero when the traffic is relatively mild . Meanwhile compared with the FCFS policy and the optimization based approach the platoon strategy can greatly reduce the average vehicle delay under congested scenarios and give a better balance between the optimality and real time performance . | Conflict free routing algorithms for automated vehicles in dedicated zones. Conflict point networks with time occupations on conflict points. Dynamic platoon strategy to reduce delay in heavy traffic. Deep reinforcement learning for optimizing the platoon sizes. |
S0968090X20306884 | The potential to implement the concept of waterborne platooning in the European short sea transportation system is currently being explored . In the concept a platoon is referred to as a Vessel Train . A VT is composed of a fully manned lead vessel and a number of follower vessels . The lead vessel takes over the navigational and situational awareness responsibilities for the follower vessels . This enables automation of the navigational tasks on these follower vessels which in turn leads to a potential reduction in crew size and associated cost . | Demonstrates an assessment that is used to determine the viability of a waterborne platoon. The max. follower vessel cost and the productivity changes are the main limiting indicators. The economic viability conditions of Vessel Train transport system are presented in the case study. 1 of vessels in the existing European short sea fleet are needed to join the platooning concept. Benefits created by the VT implementation are not large enough to guarantee a successful application. |
S0968090X20306896 | Growing awareness of the environment and new regulations of the International Maritime Organization and the European Union are forcing ship owners to reduce pollution . The use of liquefied natural gas is one of the most promising options for achieving a reduction in pollution for inland shipping and short sea shipping . However the infrastructure to facilitate the broad use of LNG is yet to be developed . We advance and analyze models that suggest LNG infrastructure development plans for refueling stations that support pipeline to ship and truck to ship bunkering specifying locations types and capacities and that take into account the characteristics of LNG such as boil off during storage and loading . We develop an effective primal heuristic based on Lagrangian relaxation for the solution of the models . We validate our approach by performing a computational study for the waterway network in the Arnhem Nijmegen region in the West European river network including among others multi year scenarios in which capacity expansion and reduction are possible . | We develop models that suggest LNG infrastructure plans for refueling stations. We consider pipeline to ship and truck to ship bunkering. We develop an effective primal heuristic based on Lagrangian relaxation. We perform a case study for the West European river network. |
S0968090X20306902 | This study aims to model two dimensional motion of an Ego Vehicle . Intelligent Driver Model is enhanced for this purpose . All the surrounding obstacles are considered as stimuli that elicit drivers reactions . A Virtual LiDAR sensor on the EV perceives the surrounding . Proximity and speed are used to compute longitudinal and lateral components of average relative velocity . A new formulation is presented to compute the effective gap considering all the surrounding obstacles . It is composed of four components front rear left and right . The former two affect longitudinal acceleration while the latter two influence lateral acceleration of the EV . A discretionary lane change model is also developed to account for the drivers deliberate lateral movements . A total of 13 model parameters are used of which eight are newly introduced . These model parameters are calibrated using 1050 human driven car trajectories from Next Generation SIMulation dataset . Taguchis fractional factorial design principle is used to optimize the parameters . Gray relational analysis indicated that the three newly introduced parameters to be the three most influential parameters . Validation using another 450 car trajectories resulted in a mean radial error of 3.97m over a horizon of 10s . The performance measures of two dimensional motion planning are found to be better than those reported in the literature . The developed human like AV IDM offers a | Classical IDM is extended for lateral and longitudinal motion planning. All the surrounding obstacles are considered to entice a drivers response. 8 new model parameters are introduced to emulate human like response. A discretionary lateral acceleration model is developed. A fractional factorial design approach is used to calibrate the model parameters. |
S0968090X20306938 | Introduction of autonomous and connected trucks is expected to result in drastic changes in operational characteristics of freight shipments which may in turn have significant impacts on highway safety vehicle fuel consumption and infrastructure durability . One such important change is the formation of truck platoons which can be defined as the convoy of trucks travelling in a very close distance . Reducing congestion regulating traffic and improving fuel efficiency are some of reported and expected benefits of platooning . Yet such platooning operations may accelerate the damage accumulation within pavement structures because the lateral position of successive trucks within a lane is expected to be similar and the time between two consecutive axle loads is expected to be reduced . Therefore this study develops a platooning control strategy for a fleet of ACTs such that the lateral position of trucks and spacing between them can be explicitly optimized to minimize damage to the pavement . Pavement damage is simulated using recently developed pavement performance models . On the other hand fluid dynamics models were developed to compute fuel cost due to aerodynamic drags . Three numerical optimization algorithms genetic algorithms particle swarm optimization and pattern search algorithm were used to solve the objective function . The proposed control strategy efficiency is demonstrated through a case study relative costs to agencies and users could be reduced by 9 . | A de centralized optimization framework leveraging V2V communication is introduced. Trade off b w pavement life cycle cost and fuel cost due to aerodynamics is addressed. Controlling the lateral positions of platoons may reduce the overall cost up to 9 . |
S0968090X2030694X | The existence of significant uncertainties in the models and systems required for trajectory prediction represents a major challenge for the Air traffic Management system . Weather can be considered as one of the most relevant sources of uncertainty . Understanding and managing the impact of these uncertainties is necessary to increase the predictability of the ATM system . State of the art probabilistic forecasts from Ensemble Prediction Systems are employed to characterize uncertainty in the wind and potential convective areas . A robust optimal control methodology to produce efficient and predictable aircraft trajectories in the presence of these uncertainties is presented . Aircraft motion is assumed to be at a constant altitude and variable speed considering BADA4 as the aircraft performance model . A set of Pareto optimal trajectories is obtained for different preferences among predictability convective risk and average cost index running a thorough parametric study on a North Atlantic crossing use case . Results show that the cost of reducing the arrival time window by 10 s. is between 100 and 200kg or 3 and 6min . depending on the cost index . They also show that reducing the exposure to convection by 50km is on the order of 5 and 10min . or 100 and 200kg . of average fuel consumption . | We find optimal aircraft routes considering the effects of atmospheric uncertainties and convective indicators. We consider convective indicators that results in areas of potential development of thunderstorms. Results show that the cost of reducing the arrival time window by 10 s is between 100 and 200 kg or 3 and 6 min. depending on the cost index. Results show that reducing the exposure to convection by 50 km is on the order of 5 and 10 min. or 100 and 200 kg of average fuel consumption. |
S0968090X20306951 | Car following behavior modeling is of great importance for traffic simulation and analysis . Considering the multi steps decision making process in human driving we propose a sequence to sequence learning based car following model incorporating not only memory effect but also reaction delay . Since the seq2seq architecture has the advantage of handling variable lengths of input and output sequences in this paper it is applied to car following behavior modeling to memorize historical information and make multi step predictions . We further compare the seq2seq model with a classical car following model and a deep learning car following model . The evaluation results indicate that the proposed model outperforms others for reproducing trajectory and capturing heterogeneous driving behaviors . Moreover the platoon simulation demonstrates that the proposed model can well reproduce different levels of hysteresis phenomenon . The proposed model is further extended with spatial anticipation which improves platoon simulation accuracy and traffic flow stability . | A car following model based on sequence to sequence seq2seq learning is proposed. The model makes multi step predictions with the consideration of reaction delay. The model outperforms other models in simulating trajectory and capturing heterogeneous driving behaviors. The model well reproduces different levels of hysteresis phenomenon. The model extended with spatial anticipation improves platoon simulation accuracy and traffic flow stability. |
S0968090X20306963 | As app based ride hailing services have been widely adopted within existing traditional taxi markets researchers have been devoted to understand the important factors that influence the demand of the new mobility . Econometric models are mainly utilized to interpret the significant factors of the demand and deep neural networks have been recently used to improve the forecasting performance by capturing complex patterns in the large datasets . However to mitigate possible traffic congestion and balance utilization rates for the current taxi drivers an effective strategy of proactively managing a quota system for both emerging services and regular taxis is still critically needed . This paper aims to systematically design an explainable deep learning model capable of assessing the quota system balancing the demand volumes between two modes . A two stage interpretable machine learning modeling framework was developed by a linear regression model coupled with a neural network layered by long short term memory . The first stage investigates the correlation between the existing taxis and on demand ride hailing services while controlling for other explanatory variables . The second stage fulfills the long short term memory network structure capturing the residuals from the first estimation stage in order to enhance the forecasting performance . The proposed stepwise modeling approach forecasts the demand of taxi rides and it is implemented in the application of pick up demand prediction using New York City taxi data . The experiment result indicates that the integrated model can capture the inter relationships between existing taxis and ride hailing services as well as identify the influence of additional factors namely the day of the week weather and holidays . Overall this modeling approach can be applied to construct an effective active demand management for the short term period as well as a quota control strategy between on demand ride hailing services and traditional taxis . | A stepwise explainable deep learning formulation using linear regression LR and a recurrent neural network. Facilitate quota based planning to balance utilization rates between for hire vehicles FHVs and traditional taxis. Data analysis from New York City Taxi Limousine Commission to observe the correlation between FHV and regular taxis. Real world data sets examined for coupled LR and long short term memory LSTM framework. |
S0968090X20306975 | Population synthesis is concerned with the generation of agents for agent based modelling in many fields such as economics transportation ecology and epidemiology . When the number of attributes describing the agents and or their level of detail becomes large survey data can not densely support the joint distribution of the attributes in the population due to the curse of dimensionality . It leads to a situation where many attribute combinations are missing from the sample data while such combinations exist in the real population . In this case it becomes essential to consider methods that are able to impute such missing information effectively . In this paper we propose to use deep generative latent models . These models are able to learn a compressed representation of the data space which when projected back to the original space leads to an effective way of imputing information in the observed data space . Specifically we employ the Wasserstein Generative Adversarial Network and the Variational Autoencoder for a large scale population synthesis application . The models are applied to a Danish travel survey with a feature space of more than 60 variables and trained and tested using cross validation . A new metric that applies to the evaluation of generative models in an unsupervised setting is proposed . It is based on the ability to generate diverse yet valid synthetic attribute combinations by comparing if the models can recover missing combinations while keeping truly impossible combinations models at a minimum . For a low dimensional experiment the VAE the marginal sampler and the fully random sampler generate 5 21 and 26 more structural zeros per sampling zero when compared to the WGAN . For a high dimensional case these figures increase to 44 2217 and 170440 respectively . This research directly supports the development of agent based systems and in particular cases where detailed socio economic or geographical representations are required . | Implicit Generative models can be tailored to the population synthesis problem. Synthesis of populations with many attributes and for rare combinations of attributes. Evaluating synthesis models measuring recovered structural and sampling zeros. The WGAN is the best performing model using both traditional and proposed measures. |
S0968090X2030704X | In pavement management systems it is beneficial to consider the economies of scale stemming from the synchronization of repairs conducted on neighboring sections within a single work zone . However finding the globally optimal solution of the repair and work zone policy for a large scale pavement network along a long term planning horizon can be computationally cumbersome . In this study as a benchmark we first propose an exact solution algorithm based on dynamic programming . Then we second propose a computationally feasible methodology a time invariant simplified rule to determine desirable policies . The proposed methodology is applied to two numerical studies Case 1 for a small scale road pavement system to compare life cycle costs and computational times between the rule based methodology and the exact solution algorithm and Case 2 for a real scale road pavement system to discuss the effectiveness of the rule based methodology . In Case 1 the rule based methodology derives a near optimal solution with a significantly shorter computational time than the exact solution algorithm . Case 2 shows that the rule based methodology can find a superior policy to the aggregation of the optimal solutions independently found for each of decomposed sub systems in a feasible computational time . Through sensitivity analyses we find that the repair and work zone policies should vary depending on the deterioration process cost factors and weight between agency and user costs for societys view or available budget for the agencys perspective . | Jointly optimize long term pavement repair and work zone policies. Consider the economies of scale stemming from the synchronization of repairs. Develop optimal solution and rule based approximation. Apply the proposed methods to two numerical studies for validation. Derive practical insights through sensitivity analyses. |
S0968090X20307051 | The establishment of dedicated automated vehicle lanes has been regarded as an effective approach for addressing heterogeneous traffic with both AVs and regular vehicles promoting both traffic efficiency and safety . However building new dedicated AV lanes in urban areas is not cost effective in the early stage of AV adoption . Fortunately dedicated bus rapid transit lanes can provide a separate right of way for AVs which is a practical and economical alternative for promoting AV development . In this paper we propose an innovative idea of allowing AVs to use dedicated BRT lanes and quantitatively analyze the stationary performance of mixed use lanes . Specifically the analysis is conducted through a cyclic spacetime model for AVs on the mixed use lane and a sequential optimization method is proposed to approximately solve the model . With the SOM providing a valid tool for performance evaluation we then develop an assignment model for the routing of AVs on a traffic corridor with both mixed use lane and general purpose lanes to minimize the total travel time of both AVs and RVs . The model is formulated as a black box nonlinear program without an explicit analytical form a successive linear programming algorithm with finite difference for gradient approximation is then utilized to solve the nonlinear program . Numerical experiments are conducted in different scenarios which reveal that the establishment of the mixed use lane can not only improve the efficiency of AVs but also alleviate the congestion on general purpose lanes . | An innovative idea of implementing a mixed use BRT AV lane by allowing AVs to use dedicated BRT lanes. A quantitative evaluation model for the performance on the BRT AV lane under different traffic inputs. An optimal assignment model of AVs on the entire corridor of general purpose lanes and the mixed use lane. Both numerical experiments and simulation experiments to evaluate the effectiveness of the proposed methods. |
S0968090X20307063 | In this paper we develop a new joint pattern recognition method that combines network motif based analysis with activity sequence based analysis . We use the advantages of both methods in creating groups of patterns that have within them distinct pattern homogeneity and across pattern heterogeneity . The first portion of the analysis here applies a more traditional approach to identify unique network motifs with 16 of them capturing 83.05 of the 2017 NHTS California workday data . Multivariate analysis of grouped motifs data shows different preference of motifs for students part time workers retirees telecommuters drivers women and younger adults . In the second portion of the analysis motifs are grouped into categories based on the number of locations a person visits in a day and their correlation with time use and travel is explored . Time use and travel are analyzed based on minute by minute time allocation pattern identification using sequence analysis and hierarchical clustering . The correlation between motifs group and sequence analysis finds substantial heterogeneity within the motif groups . The within motif group clusters of activity based sequences show typical commuting going to school and resting patterns . We also find seven patterns that are not typical but have similarities across motifs in their temporal footprint and the variety of activities in each sequence . The paper provides a summary of the analytical steps and findings as well as next steps . | A new pattern recognition method combining motif with activity sequence analysis. Apply motif to identify human mobility patterns in California travel survey. Explore the relationship between motifs and peoples characteristics. Conduct an activity sequence analysis to reveal heterogeneity in time allocation. |
S0968090X20307087 | Due to the confined span of parking location choice for human driven vehicles the spatiotemporal imbalance in parking space utilization has always been a challenging problem in many major cities leading to a substantial waste of precious land resources . However in the era of autonomous vehicles with parking autonomy there are more alternative parking options available as AV can park at more distant locations away from the trip makers destination . This paper aims to investigate the morning commute problem with consideration of AV commuters distant parking choices in a many to one network . Parking sharing scheme is explicitly considered wherein AV travellers can choose to park at city centre at public parking facilities and then lease their own parking space out park at home or park at a shared parking space . We first examine how AV commuters in different residential clusters prioritize their parking location choices as well as their willingness to share the vacant parking space . Then we investigate the travellers trip timing decisions and determine the resultant equilibrium travel pattern . Without other dynamic managing schemes the model results indicate that appropriate CBD parking supply together with differentiated parking charges subsidy can reduce the total queueing congestion significantly yet at the cost of higher total travel cost . | Morning commute problem with parking is studied in many to one network in an AV future. Multiple parking options CBD parking home parking and shared parking are considered. The behaviours of AV commuters as both parking users and suppliers are examined. Parking prioritization of AV commuters is investigated. Appropriate CBD parking provision and differentiated pricing alleviate total bottleneck congestion. |
S0968090X20307099 | Drivers cruising for scarce on street parking in city centers create negative externalities including congestion and pollution . We apply a serious game PARKGAME to understand and model drivers two intertwined instantaneous parking choices | We employ a serious game to reveal parking search behavior. AFT and MNL regression models are applied to analyze parking choices. Driver behavior is suboptimal myopic and risk averse. Drivers search path choice complies with a biased random walk model. Formalized decision making rules can be used to characterize agents in a parking ABM. |
S0968090X20307105 | While activity based travel demand generation has improved over the last few decades the behavioural richness and intuitive interpretation remain challenging . This paper argues that it is essential to understand why people travel the way they do and not only be able to predict the overall activity patterns accurately . If one can not understand the why then a models ability to evaluate the impact of future interventions is severely diminished . Bayesian networks provide the ability to investigate causality and is showing value in recent literature to generate synthetic populations . This paper is novel in extending the application of BNs to daily activity tours . Results show that BNs can synthesise both activity and trip chain structures accurately . It outperforms a frequentist approach and can cater for infrequently observed activity patterns and patterns unobserved in small sample data . It can also account for temporal variables like activity duration . | Bayesian network approach to synthesise daily activity and trip chains. Network structure includes demographic travel and temporal variables. Rigorous and disaggregate regime to test accuracy. Bayesian networks is both behaviourally rich and intuitively interpretable. |
S0968090X20307117 | The range anxiety has been a major factor that affects the market acceptance of electric vehicles . Even with the recent development of battery technologies a lack of charging stations and range anxiety are still significant concerns specifically for intercity trips . This calls for more investments in building charging stations and advancing battery technologies to increase the market share of electric vehicles and improve sustainability . This study suggests a configuration for plug in electric vehicle charging infrastructure to support long distance intercity trips of electric vehicles at the network level . A model is proposed to minimize the total system cost including infrastructure investment and travel time delays . This study fills existing gaps in the literature by capturing realistic patterns of travel demand and considering flow dependent charging delays at charging stations . Furthermore the proposed model which is formulated as a mixed integer program with nonlinear constraints solves the optimization problem at the network level . At the network level impacts of charging station locations on the traffic assignment problem with a mixed fleet of electric and conventional vehicles need to be considered . To this end a traffic assignment module is integrated with a simulated annealing algorithm . The numerical experiments show a satisfactory application of the model for a full scale case study . The solution quality and efficiency of the proposed solution algorithm are evaluated against those of an enumeration approach for a small case study . The results suggest that even for the current market share and charging stations setting a significant investment is needed to support intercity trips without range anxiety issues and with acceptable delays . Furthermore through sensitivity analyses the required infrastructure and battery investments to support intercity trips with acceptable delays are established for hypothetical increased market shares and battery size in the future . | Optimizing charging station locations and number of chargers for inter city trips. Minimizing infrastructure cost and users detour waiting and charging delays. Tracking state of charge and allowing multiple recharging for long distance trips. Proposing a meta heuristic solution algorithm based on the Simulated Annealing. Practical insights by sensitivity analyses on market share and battery size. |
S0968090X20307129 | Road traffic congestion is the result of various phenomena often of random nature and not directly observable with empirical experiments . This makes it difficult to clearly understand the empirically observed traffic instabilities . The vehicles acceleration deceleration patterns are known to trigger instabilities in the traffic flow under congestion . It has been empirically observed that free flow pockets or voids may arise when there is a difference in the speeds and the spacing between the follower and the leader increases . During these moments the trajectory is dictated mainly by the characteristics of the vehicle and the behaviour of the driver and not by the interactions with the leader . Voids have been identified as triggers for instabilities in both macro and micro level which influence traffic externalities such as fuel consumption and emissions . In the literature such behaviour is usually reproduced by injecting noise to the results of car following models in order to create fluctuations in the instantaneous vehicles acceleration . | Formalizing heterogeneity of the vehicle driver system. Free flow term of a CF model leads to the formation and propagation of oscillations. Simulations reproduce periodical oscillations with variable amplitudes and periods. The concave growth pattern of the standard deviation of the speed is observed. Validation was performed on three real world car following datasets. The proposed approach provides accurate fuel consumption estimations. |
S0968090X20307142 | This paper focuses on traffic parameters estimation at signalized intersections based on a framework combining shockwave analysis and Bayesian Network using vehicle trajectory data . Detailed queuing evolution and spillback across adjacent intersections are considered . According to shockwave analysis the analytical probability distribution of individual vehicles travel time is derived based on different initial conditions . This probability distribution is parameterized by the fundamental diagram parameters traffic volume and cycle state . A three layer recursive BN model is then proposed to construct the state evolution process as well as the relationships between traffic volume cycle state FD parameters sampled vehicles arrival times and intersection travel times . As traffic volume and initial queue can not be measured directly from sampled trajectory data the expectation maximization algorithm and particle filtering are introduced to solve this recursive BN model . By shockwave analysis such estimated traffic parameters are then used to estimate the maximum queue length and traffic volume of each cycle . The proposed method is evaluated using microscopic traffic simulation data as well as empirical data . Numerical results show that the proposed method achieves promising accuracy even under low penetration rates with the mean absolute percentage error of the estimation bounded by 15 and generally around 10 . | Traffic parameter estimation at signalized intersections considering queue spillback using vehicle trajectory data. A framework combined shockwave analysis and Bayesian Network is developed. Cycle by cycle evolution of queuing and spillback is captured in the combined framework. Different delay patterns are elaborated under both unsaturated and oversaturated conditions. Estimation results with mean absolute percentage error bounded by 15 |
S0968090X20307154 | This paper proposes a new class of extreme value distribution called compound generalized extreme value distribution for investigating the effects of monthly and seasonal variation on extreme travel delays in road networks . Since the frequency and severity of extreme events are highly correlated to the variation in weather conditions as an extrinsic cause of incidents and long delays monthly and seasonal changes in weather contributes to extreme travel time variability . The change in driving behavior which itself varies according to road weather conditions also contributes to the monthly and seasonal variation in observed extreme travel times . Therefore it is critical to model the effect of monthly and seasonal changes on observed extreme travel delays on road networks . Based on the empirically revealed linear relationship between mean and standard deviation of extreme travel delays for both monthly and seasonal levels two multiplicative error models are formulated . The CGEV distribution is then obtained by linking the two multiplicative error models and forming a compound distribution that characterizes the overall variation in extreme travel delay . The CGEV distribution parameters are calibrated and the underlying assumptions that are used to derive the CGEV distribution are validated using multi year observed travel time data from the City of Calgary road network . The results indicate that accounting for the seasonality by identifying seasonal specific parameters provides a flexible and not too complex CGEV distribution that is shown to outperform the traditional GEV distribution . Finally the application of the proposed CGEV distribution is evaluated in the context of road network vulnerability taking into account the observed probability of extreme event occurrences and the link importance . This derived data driven vulnerability index incorporates a wealth of information related to both network topography in terms of connectivity and the dynamic interaction between travel demand and supply . This new data driven vulnerability measure can thus be used as a decision support tool to inform decision makers in prioritizing improvements to critical links to enhance overall network vulnerability reliability and resilience . | A new data driven network vulnerability approach is presented. A new distribution called Compound Generalized Extreme Value CGEV is derived. The CGEV distribution models monthly and seasonal variation in extreme travel delay. The CGEV distribution is applied to derive a new measure of network vulnerability. The vulnerability measure considers probability and consequences of link failure. |
S0968090X20307178 | In this paper we proposed a new method to extract travel patterns for transit riders from different public transportation systems based on temporal motif which is an emerging notion in network science literature . We then developed a scalable algorithm to recognize temporal motifs from daily trip sub sequences extracted from two smart card datasets . Our method shows its benefits in uncovering the potential correlation between varying topologies of trip combinations and specific activity chains . Commuting different types of transfer and other travel behaviors have been identified . Besides varying travel activity chains like Home | Developing a method based on Temporal Motif to identify individual travel patterns. Inferring the correlation between travel activity chains and different travel motifs. Extracted individual travel motifs provide extra information for trip prediction. Bridging the concepts of travel regularity and temporal network complexity. |
S0968090X20307191 | In recent years there has been a rapid growth of smart apps that could interact with users and implement personalized rewards to coordinate and change user behavior . Understanding user behavior is an enabling factor for the success of these promising apps . However existing statistical models for modeling user behavior encounter limitations . Choice models based on Random Utility Maximization commonly assume that the data collection is independent with the human behavior . However when users interact with the apps the real potential and also the real challenge for modeling user behavior is that the apps not merely are data collection tools but also change users behaviors . In this work we model the user behavior as a graphical model examine our hypothesis that existing choice models are not suitable and develop an interesting computational strategy using max margin formulation to overcome the learning challenge of the our proposed graphical model that is named the Latent Decision Threshold model . | Latent Decision Threshold model characterizes the userapp interaction process. LDT model provides a characterization of decision making behavior. Max margin learning algorithm can efficiently estimate the parameters. The LDT model can help discover users behavior patterns. |
S0968090X20307208 | This study focuses on single variable optimization approaches which determine the holding time of a vehicle when it is ready to depart from a bus stop . Up to now single variable optimization methods resort to rule based control logics to equalize the inter departure headways or adhere to the target headway values . One of them is the two headway based control logic which determines the holding time of a bus based on its headway with its preceding and following bus without addressing other implications such as overcrowding . To rectify this we introduce a new model for the single variable bus holding problem that considers the passenger demand and vehicle capacity limits . Then we reformulate this problem to an easier to solve program with the use of slack variables and introduce an analytic solution that can determine the holding time of a vehicle at the respective bus stop . Our analytic solution does not add a computational burden to the two headway based control logic and can be applied in real time . The operational benefit of our bus holding approach compared to other analytic solutions that do not consider the vehicle capacity is investigated using actual data from bus line 302 in Singapore . | Analytic solution for the single variable bus holding problem considering capacity limits. The holding time of a vehicle can be determined in real time. Significant reduction of stranded passengers in the expense of a slight regularity deterioration. The service regularity performance is relatively insensitive to demand variations. |
S0968090X2030721X | Lane changing is one of the complex driving tasks that depends on the number of vehicles objectives and lanes . A driver often needs to respond to a lane changing request of a lane changer which is a function of their personality traits and the current driving conditions . A connected environment is expected to assist during the lane changing decision making process by increasing situational awareness of surrounding traffic through vehicle to vehicle communication and vehicle to infrastructure communication . Although the majority of lane changing decision making components in a traditional environment has been frequently investigated our understanding of drivers interactions during the lane changing decision making process in a connected environment remains elusive due to the novelty of a connected environment and the scarcity of relevant data . As such this study examines drivers responses to lane changing requests in a connected environment using the CARRS Q Advanced Driving Simulator . Seventy eight participants responded to the lane changing request of a lane changer under two randomised driving conditions baseline and connected environment . A segmentation based approach is employed to extract drivers responses to the lane changing request and subsequently estimate their response time from trajectory data . Additionally drivers response times are modelled using a random parameter accelerated failure time hazard based duration model . Results reveal that drivers tend to be more cooperative in response to a lane changing request in the connected environment compared with the baseline condition whereby they tend to accelerate to avoid the lane changing request . The AFT model suggests that on average drivers response times are shorter in the connected environment implying that drivers respond to the lane changing request faster in the presence of driving aids . However at the individual level connected environments impact on drivers response times is mixed as drivers response times may increase or decrease in the connected environment compared to the baseline condition for instance we find that female drivers have lower response times in the connected environment than that of male drivers . Overall this study finds that drivers in connected environment on average take less time to respond and appear to be more cooperative and thus are less likely to be engaged in safetycritical events . | Examined drivers responses to a lane changing request in a connected environment. Drivers are more cooperative in response to a lane changing request in a connected environment. Modelled response time using a random parameter accelerated failure time hazard based duration model. Response times on average are shorter in a connected environment. At the individual level response time may increase or decrease in a connected environment. |
S0968090X20307221 | The present study investigates the determinants of the volatility of passenger demand for paratransit services and explores the feasibility of a data driven model for medium term forecast of the daily demand . Medium term demand forecasting is a significant insight to optimise resource allocation and reduce operations costs . Using operational data from the reservation platform of the paratransit services in Toulouse France and enriching them with exogenous information the study derives statistical and deep learning models for medium term forecast . These models include a seasonal ARIMAX model with rolling forecast a Random Forest Regressor a LSTM neural network with exogenous information and a CNN neural network with independent variables . The seasonal ARIMAX model yields the best performance suggesting that when linear relationships are considered econometric models and deep learning models do not have significant differences in their performance . All the models show limited ability to grasp unique events with multi day impacts such as strikes . Albeit a highly volatile demand and limited knowledge ahead of the forecast these models suggest the volume of early reservations is a good proxy for the daily demand . | Daily demand of demand responsive paratransit services is highly volatile. Weather conditions do not seem significant to explain the variability of the daily demand. The volume of reservations 7days ahead and the type of day are significant proxy for forecasting daily demand. Statistical models with exogenous information provide similar accuracy with deep learning models such as LSTM and CNN. |
S0968090X20307233 | Two car households with one conventional and one battery electric vehicle have an opportunity to partially circumvent the range limitations of a modest range battery electric vehicle through flexible use . To investigate the extent to which real world households utilize this flexibility we used from 20 two car households in the Gothenburg area in Sweden GPS data from before and during an EV trial in which the households were asked to temporarily replace one of their two conventional cars with a short range EV for the duration of the trial . The actual household electric drive fraction i.e . the EV distance as a share of the total two car household driving distance varied between 30 and 70 with a household mean of 47 . On average this corresponds to 80 of the estimated potential household electric drive fraction during the trial . We quantify the flexibility in choosing the EV as the difference in distance between the potential and the minimum needed EV driving . For below range home to home non overlapping trips the households used 69 of that flexibility . For household trips that did overlap in time they used 56 of the flexibility . Thus the EV is the preferred vehicle but the preference is less obvious for overlapping trips . Our analysis implies an even more dominant role for the EV in weekend driving . Further although the pre trial data showed a large difference in the household shares of distances driven between a replaced first car and a replaced second car this difference disappears when an EV replaces either one . | In a trial a BEV replaced one of the conventional cars in 20 two car households. Unique study with GPS data on both cars in the trial as well as pre trial period. For below range single trips the BEV was used for over 2 3 of driving. The replaced cas driving whether 1st or 2nd car was transcended. |
S0968090X20307245 | In recent years due to the increased availability of data and improvements in computing power application of machine learning techniques to various aviation safety problems for identifying isolating and reducing risk has gained momentum . Data collected from on board recorders in commercial aircraft contain thousands of parameters in the form of multivariate time series which are used to train the machine learning models . Among the phases of flight approach and landing phases result in the most accidents and incidents . The performance and trajectory of the aircraft during the approach phase is an indicator of its landing performance which in turn affects incident or accident probability such as runway excursions . Landing performance is commonly measured using metrics such as landing airspeed vertical speed location of touchdown point on runway etc . While current applications of machine learning to aviation focus on retrospective insights to implement corrective measures they offer limited value for real time risk identification or decision making as they are inherently reactive in nature . In this work a novel offline online framework is developed for building a global predictive model offline to predict landing performance metrics online . The framework leverages flight data from the approach phase between certain approach altitudes also called | Provides a novel online predictive model of aircraft landing performance using data collected on board an aircraft during the approach phase. Demonstrates accurate prediction of the critical metrics further ahead in time than existing approaches and at a higher accuracy. Introduces innovations in generating feature vectors and target labels using a flexible approach that can be easily replicated for other metrics of interest. Significantly improves over previous similar work in literature both in terms of accuracy and applicability in an online setting. |
S0968090X20307270 | In the subway system passenger crowding in peak hours is likely to cause train delays that easily propagate to following trains resulting in a lower efficiency of the system . Consequently this paper focuses on determining a robust timetable for the trains on the one hand i.e . finding a better timetable to avoid delay propagation as much as possible in case of a crowded subway system . On the other hand this paper considers the energy efficiency i.e . reducing the total energy consumption during operations by selecting appropriate speed profiles and maximizing the utilization of regenerative braking energy . A related mathematical optimization model is formulated with the objective of maximizing the robustness and minimizing the total energy consumption . In order to solve this model an efficient algorithm i.e . simulation based variable neighborhood search algorithm is presented to obtain a good timetable in reasonable amount of time . Finally experiments are implemented to show the performance of the proposed algorithm . | Propose the adjustment strategies in case of train delays. Propose a method of evaluating the robustness of the train timetable. Provide a method of formulating a robust and energy efficient train timetable. Design a simulation based variable neighborhood search algorithm. |
S0968090X20307282 | Train timetabling and train platforming are problems of crucial importance when scheduling high speed trains . Often these problems are solved separately and in sequence . It is also not uncommon for the problems to be further decomposed by direction since the use of tracks is usually direction specific in a high speed network . In this paper we consider the optimization problem of integrating re timetabling and re routing decisions within station areas for multiple stations when scheduled maintenance renders the existing optimized schedules infeasible . We model the underlying problems using a spacetime network on a mesoscopic level and propose a 01 binary integer programming model that can simultaneously modify the timings and routes of trains from different directions . Two different solution approaches are described . The first is a commonly used Lagrangian Relaxation approach while the second utilizes the Alternating Direction Method of Multipliers concept . For both methods a time dependent dynamic programming approach is used to solve the resulting subproblems . A comparison of the two approaches on instances provided by the Chinese high speed railway indicates that the ADMM based approach provides tighter upper bounds and typically requires fewer iterations than the Lagrangian Relaxation approach . Furthermore the results show that a flexible track utilization policy provides better timetables with fewer cancellations and less total delay than a fixed dedicated direction track policy . | An integrated approach to re optimize train schedules given track maintenance tasks. Modeling at the mesoscopic level captures interdependencies between multiple stations. A flexible station track utilization policy reduces train delay and cancellations. The Alternating Direction Method of Multipliers outperforms Lagrangian Relaxation. |
S0968090X20307294 | Turning movement spillover is the result of a turning bay section not being able to accommodate all arriving vehicles so that the turning vehicle queue spills back and blocks other vehicles turning in different directions . We are not aware of any TMS estimation method that can remedy this situation or support relevant applications in real time . This research proposes a quasi real time algorithm for estimating TMS which includes triggering movement as well as duration estimation . The proposed method is based on data for connected vehicles including their trajectories and their desired turning directions . In addition a model that uses partial trajectory data is proposed . For each assumed TMS a simplified trajectory is developed by the construction of a piece wise linear curve . To minimize any deviation of the simplified trajectory from observation a TMS estimation can be made . This proposed method is effective and computationally efficient when tested against dynamic demand in two mainstream signal phase settings with varied sample sizes . Even though data for a higher number of vehicle samples is generally favorable the proposed model still makes a good estimate when only one trajectory is available . | A quasi real time estimator of turning movement spillover TMS using partial connect vehicle data is proposed. The estimator can identify the TMS triggering movements and estimate the TMS durations. A nonlinear minimization problem was formulated to minimize the deviation of the theoretical trajectory from the observed trajectory. The estimation algorithm is tested against dynamic demand and common signal settings. The results prove the accuracy and the computational efficiency. |
S0968090X20307312 | Ridesourcing services provide alternative mobility options in several cities . Their market share has grown exponentially due to the convenience they provide . The use of such services may be associated with car light or car free lifestyles . However there are growing concerns regarding their impact on urban transportation operations performance due to empty unproductive miles driven without a passenger . This paper is motivated by the potential to reduce deadhead mileage of ridesourcing trips by providing drivers with information on future ridesourcing trip demand . Future demand information enables the driver to wait in place for the next riders request without cruising around and contributing to congestion . A machine learning model is employed to predict hourly and 10 minute future interval travel demand for ridesourcing at a given location . Using future demand information we propose algorithms to assign drivers to act on received demand information by waiting in place for the next rider and match these drivers with riders to minimize deadheading distance . Real world data from ridesourcing providers in Austin TX and Chengdu China are leveraged . Results show that this process achieves 68 82 and 53 60 reduction of trip level deadheading miles for the RideAustin and DiDi Chuxing sample operations respectively under the assumption of unconstrained availability of short term parking . Deadheading savings increase slightly as the maximum tolerable waiting time for the driver increases . Further it is observed that significant deadhead savings per trip are possible even when a small percent of the ridesourcing driver pool is provided with future ridesourcing demand information . | Providing drivers with future ridesourcing trip demand predictions can reduce trip level deadheading. Drivers act on high expected demand information by waiting in place for the next rider. Machine learning for future demand prediction and trips assignment by minimizing the empty vehicle mileage. The analysis leverages sample data from ridesourcing services operating in US and China. Average deadhead mileage reduction ranges from 53 to 82 for ridesourcing operations per trip. |
S0968090X20307324 | Max pressure traffic signal control has many desirable properties . It is analytically proven to maximize network throughput if demand could be served by any signal control . Despite its network level stability properties the control itself is decentralized and therefore easily computed by individual intersection controllers . Discussions with city engineers have suggested that a major barrier to implementation in practice is the non cyclical phase actuation of max pressure control which can actuate any phase in arbitrary order to serve the queue with highest pressure . This arbitrary phase selection may be confusing to travelers expecting a signal cycle and is therefore unacceptable to some city traffic engineers . This paper revises the original max pressure control to include a signal cycle constraint . The max pressure control must actuate an exogenous set of phases in order with each phase actuated at least one time step per cycle . Each cycle has a maximum length but the length can be reduced if desired . Within those constraints we define a modified max pressure control and prove its maximum stability property . The revised max pressure control takes the form of a model predictive control with a one cycle lookahead but we prove that the optimal solution can be easily found by enumerating over phases . The policy is still decentralized . Numerical results show that as expected the cyclical max pressure control performs slightly worse than the original max pressure control due to the additional constraints but with the advantage of greater palatability for implementation in practice . | We define a modified max pressure control that follows an exogenous signal cycle with maximum cycle length. We prove the maximum stability and decentralized properties of the max pressure control. Numerical results compare performance with previous max pressure control policies |
S0968090X20307348 | This paper presents two algorithms to estimate traffic state in urban street networks with a mixed traffic stream of connected and unconnected vehicles and incorporates them in a real time and distributed traffic signal control methodology . The first algorithm integrates connected vehicles and loop detector data to estimate the trajectory of unconnected vehicles based on car following concepts . The second algorithm converts the temporal point vehicle detections to a spatial vehicle distribution on a link . The signal control methodology utilizes either algorithm to estimate traffic state on all network links at a time optimizes the signal timing parameters over a prediction period constituting several time steps implements the optimal decisions in the next time step and continues this process until the end of the study period . We applied the methodology to a real world case study network simulated in Vissim . The results show that both algorithms are effective under a wide range of CV market penetration rates in all tested demand patterns at 0 market penetration rate the proposed methodology reduced travel time by 2 to 10 and average delay by 7 to 20 compared to the existing signal timing parameters and traffic demand . At a 40 penetration rate the proposed algorithms reduced travel time by 27 to 33 | Development of two traffic state estimation algorithms for partially connected transportation networks. Developing a real time methodology for traffic signal control with partial connected vehicle information. Achieving significant improvement in traffic operations compared to existing approaches. Even at 10 CV market share the number of completed trips increased by 3.23.5 . |
S0968090X2030735X | On demand systems in which several users can ride simultaneously the same vehicle have great potential to improve mobility while reducing congestion . Nevertheless they have a significant drawback the actual realization of a trip depends on the other users with whom it is shared as they might impose extra detours that increase the waiting time and the total delay even the chance of being rejected by the system depends on which travelers are using the system at the same time . In this paper we propose a general description of the sources of unreliability that emerge in ridesharing systems and we introduce several measures . The proposed measures are related to two sources of unreliability induced by how requests and vehicles are being assigned namely how users times change within a single trip and between different realizations of the same trip . We then analyze both sources using a state of the art routing and assignment method and a New York City test case . Regarding same trip unreliability in our experiments for different fixed fleet compositions and when reassignment is not restricted we find that more than one third of the requests that are not immediately rejected face some change and the magnitude of these changes is relevant when a user faces an increase in her waiting time this extra time is comparable to the average waiting time of the whole system and the same happens with total delay . Algorithmic changes to reduce this uncertainty induce a trade off with respect to the overall quality of service . For instance not allowing for reassignments may increase the number of rejected requests . Concerning the unreliability between different trips we find that the same origin destination request can be rejected or served depending on the state of the fleet . And when it is served the waiting times and total delay are rarely equal which remains true for different fleet sizes . Furthermore the largest variations are faced by trips beginning at high demand areas . | We propose measures for the novel unreliability sources in on demand pooled ridesharing systems. Two types of unreliability sudden changes during a trip and different results for a same request. We measure these changes using a state of the art assignment method over a real dataset. At least one third of the requests face sudden changes and different repetitions are rarely equal. We measure a trade off between reliability and waiting times detours and rejection rates. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.