import numpy as np
from gym import utils
from gym.envs.mujoco import MujocoEnv
from gym.spaces import Box
class HumanoidStandupEnv(MujocoEnv, utils.EzPickle):
"""
### Description
This environment is based on the environment introduced by Tassa, Erez and Todorov
in ["Synthesis and stabilization of complex behaviors through online trajectory optimization"](https://ieeexplore.ieee.org/document/6386025).
The 3D bipedal robot is designed to simulate a human. It has a torso (abdomen) with a
pair of legs and arms. The legs each consist of two links, and so the arms (representing the
knees and elbows respectively). The environment starts with the humanoid laying on the ground,
and then the goal of the environment is to make the humanoid standup and then keep it standing
by applying torques on the various hinges.
### Action Space
The agent take a 17-element vector for actions.
The action space is a continuous `(action, ...)` all in `[-1, 1]`, where `action`
represents the numerical torques applied at the hinge joints.
| Num | Action | Control Min | Control Max | Name (in corresponding XML file) | Joint | Unit |
| --- | ---------------------------------------------------------------------------------- | ----------- | ----------- | -------------------------------- | ----- | ------------ |
| 0 | Torque applied on the hinge in the y-coordinate of the abdomen | -0.4 | 0.4 | hip_1 (front_left_leg) | hinge | torque (N m) |
| 1 | Torque applied on the hinge in the z-coordinate of the abdomen | -0.4 | 0.4 | angle_1 (front_left_leg) | hinge | torque (N m) |
| 2 | Torque applied on the hinge in the x-coordinate of the abdomen | -0.4 | 0.4 | hip_2 (front_right_leg) | hinge | torque (N m) |
| 3 | Torque applied on the rotor between torso/abdomen and the right hip (x-coordinate) | -0.4 | 0.4 | right_hip_x (right_thigh) | hinge | torque (N m) |
| 4 | Torque applied on the rotor between torso/abdomen and the right hip (z-coordinate) | -0.4 | 0.4 | right_hip_z (right_thigh) | hinge | torque (N m) |
| 5 | Torque applied on the rotor between torso/abdomen and the right hip (y-coordinate) | -0.4 | 0.4 | right_hip_y (right_thigh) | hinge | torque (N m) |
| 6 | Torque applied on the rotor between the right hip/thigh and the right shin | -0.4 | 0.4 | right_knee | hinge | torque (N m) |
| 7 | Torque applied on the rotor between torso/abdomen and the left hip (x-coordinate) | -0.4 | 0.4 | left_hip_x (left_thigh) | hinge | torque (N m) |
| 8 | Torque applied on the rotor between torso/abdomen and the left hip (z-coordinate) | -0.4 | 0.4 | left_hip_z (left_thigh) | hinge | torque (N m) |
| 9 | Torque applied on the rotor between torso/abdomen and the left hip (y-coordinate) | -0.4 | 0.4 | left_hip_y (left_thigh) | hinge | torque (N m) |
| 10 | Torque applied on the rotor between the left hip/thigh and the left shin | -0.4 | 0.4 | left_knee | hinge | torque (N m) |
| 11 | Torque applied on the rotor between the torso and right upper arm (coordinate -1) | -0.4 | 0.4 | right_shoulder1 | hinge | torque (N m) |
| 12 | Torque applied on the rotor between the torso and right upper arm (coordinate -2) | -0.4 | 0.4 | right_shoulder2 | hinge | torque (N m) |
| 13 | Torque applied on the rotor between the right upper arm and right lower arm | -0.4 | 0.4 | right_elbow | hinge | torque (N m) |
| 14 | Torque applied on the rotor between the torso and left upper arm (coordinate -1) | -0.4 | 0.4 | left_shoulder1 | hinge | torque (N m) |
| 15 | Torque applied on the rotor between the torso and left upper arm (coordinate -2) | -0.4 | 0.4 | left_shoulder2 | hinge | torque (N m) |
| 16 | Torque applied on the rotor between the left upper arm and left lower arm | -0.4 | 0.4 | left_elbow | hinge | torque (N m) |
### Observation Space
The state space consists of positional values of different body parts of the Humanoid,
followed by the velocities of those individual parts (their derivatives) with all the positions ordered before all the velocities.
**Note:** The x- and y-coordinates of the torso are being omitted to produce position-agnostic behavior in policies
The observation is a `ndarray` with shape `(376,)` where the elements correspond to the following:
| Num | Observation | Min | Max | Name (in corresponding XML file) | Joint | Unit |
| --- | --------------------------------------------------------------------------------------------------------------- | ---- | --- | -------------------------------- | ----- | -------------------------- |
| 0 | z-coordinate of the torso (centre) | -Inf | Inf | root | free | position (m) |
| 1 | x-orientation of the torso (centre) | -Inf | Inf | root | free | angle (rad) |
| 2 | y-orientation of the torso (centre) | -Inf | Inf | root | free | angle (rad) |
| 3 | z-orientation of the torso (centre) | -Inf | Inf | root | free | angle (rad) |
| 4 | w-orientation of the torso (centre) | -Inf | Inf | root | free | angle (rad) |
| 5 | z-angle of the abdomen (in lower_waist) | -Inf | Inf | abdomen_z | hinge | angle (rad) |
| 6 | y-angle of the abdomen (in lower_waist) | -Inf | Inf | abdomen_y | hinge | angle (rad) |
| 7 | x-angle of the abdomen (in pelvis) | -Inf | Inf | abdomen_x | hinge | angle (rad) |
| 8 | x-coordinate of angle between pelvis and right hip (in right_thigh) | -Inf | Inf | right_hip_x | hinge | angle (rad) |
| 9 | z-coordinate of angle between pelvis and right hip (in right_thigh) | -Inf | Inf | right_hip_z | hinge | angle (rad) |
| 10 | y-coordinate of angle between pelvis and right hip (in right_thigh) | -Inf | Inf | right_hip_y | hinge | angle (rad) |
| 11 | angle between right hip and the right shin (in right_knee) | -Inf | Inf | right_knee | hinge | angle (rad) |
| 12 | x-coordinate of angle between pelvis and left hip (in left_thigh) | -Inf | Inf | left_hip_x | hinge | angle (rad) |
| 13 | z-coordinate of angle between pelvis and left hip (in left_thigh) | -Inf | Inf | left_hip_z | hinge | angle (rad) |
| 14 | y-coordinate of angle between pelvis and left hip (in left_thigh) | -Inf | Inf | left_hip_y | hinge | angle (rad) |
| 15 | angle between left hip and the left shin (in left_knee) | -Inf | Inf | left_knee | hinge | angle (rad) |
| 16 | coordinate-1 (multi-axis) angle between torso and right arm (in right_upper_arm) | -Inf | Inf | right_shoulder1 | hinge | angle (rad) |
| 17 | coordinate-2 (multi-axis) angle between torso and right arm (in right_upper_arm) | -Inf | Inf | right_shoulder2 | hinge | angle (rad) |
| 18 | angle between right upper arm and right_lower_arm | -Inf | Inf | right_elbow | hinge | angle (rad) |
| 19 | coordinate-1 (multi-axis) angle between torso and left arm (in left_upper_arm) | -Inf | Inf | left_shoulder1 | hinge | angle (rad) |
| 20 | coordinate-2 (multi-axis) angle between torso and left arm (in left_upper_arm) | -Inf | Inf | left_shoulder2 | hinge | angle (rad) |
| 21 | angle between left upper arm and left_lower_arm | -Inf | Inf | left_elbow | hinge | angle (rad) |
| 22 | x-coordinate velocity of the torso (centre) | -Inf | Inf | root | free | velocity (m/s) |
| 23 | y-coordinate velocity of the torso (centre) | -Inf | Inf | root | free | velocity (m/s) |
| 24 | z-coordinate velocity of the torso (centre) | -Inf | Inf | root | free | velocity (m/s) |
| 25 | x-coordinate angular velocity of the torso (centre) | -Inf | Inf | root | free | anglular velocity (rad/s) |
| 26 | y-coordinate angular velocity of the torso (centre) | -Inf | Inf | root | free | anglular velocity (rad/s) |
| 27 | z-coordinate angular velocity of the torso (centre) | -Inf | Inf | root | free | anglular velocity (rad/s) |
| 28 | z-coordinate of angular velocity of the abdomen (in lower_waist) | -Inf | Inf | abdomen_z | hinge | anglular velocity (rad/s) |
| 29 | y-coordinate of angular velocity of the abdomen (in lower_waist) | -Inf | Inf | abdomen_y | hinge | anglular velocity (rad/s) |
| 30 | x-coordinate of angular velocity of the abdomen (in pelvis) | -Inf | Inf | abdomen_x | hinge | aanglular velocity (rad/s) |
| 31 | x-coordinate of the angular velocity of the angle between pelvis and right hip (in right_thigh) | -Inf | Inf | right_hip_x | hinge | anglular velocity (rad/s) |
| 32 | z-coordinate of the angular velocity of the angle between pelvis and right hip (in right_thigh) | -Inf | Inf | right_hip_z | hinge | anglular velocity (rad/s) |
| 33 | y-coordinate of the angular velocity of the angle between pelvis and right hip (in right_thigh) | -Inf | Inf | right_hip_y | hinge | anglular velocity (rad/s) |
| 35 | angular velocity of the angle between right hip and the right shin (in right_knee) | -Inf | Inf | right_knee | hinge | anglular velocity (rad/s) |
| 36 | x-coordinate of the angular velocity of the angle between pelvis and left hip (in left_thigh) | -Inf | Inf | left_hip_x | hinge | anglular velocity (rad/s) |
| 37 | z-coordinate of the angular velocity of the angle between pelvis and left hip (in left_thigh) | -Inf | Inf | left_hip_z | hinge | anglular velocity (rad/s) |
| 38 | y-coordinate of the angular velocity of the angle between pelvis and left hip (in left_thigh) | -Inf | Inf | left_hip_y | hinge | anglular velocity (rad/s) |
| 39 | angular velocity of the angle between left hip and the left shin (in left_knee) | -Inf | Inf | left_knee | hinge | anglular velocity (rad/s) |
| 40 | coordinate-1 (multi-axis) of the angular velocity of the angle between torso and right arm (in right_upper_arm) | -Inf | Inf | right_shoulder1 | hinge | anglular velocity (rad/s) |
| 41 | coordinate-2 (multi-axis) of the angular velocity of the angle between torso and right arm (in right_upper_arm) | -Inf | Inf | right_shoulder2 | hinge | anglular velocity (rad/s) |
| 42 | angular velocity of the angle between right upper arm and right_lower_arm | -Inf | Inf | right_elbow | hinge | anglular velocity (rad/s) |
| 43 | coordinate-1 (multi-axis) of the angular velocity of the angle between torso and left arm (in left_upper_arm) | -Inf | Inf | left_shoulder1 | hinge | anglular velocity (rad/s) |
| 44 | coordinate-2 (multi-axis) of the angular velocity of the angle between torso and left arm (in left_upper_arm) | -Inf | Inf | left_shoulder2 | hinge | anglular velocity (rad/s) |
| 45 | angular velocitty of the angle between left upper arm and left_lower_arm | -Inf | Inf | left_elbow | hinge | anglular velocity (rad/s) |
Additionally, after all the positional and velocity based values in the table,
the state_space consists of (in order):
- *cinert:* Mass and inertia of a single rigid body relative to the center of mass
(this is an intermediate result of transition). It has shape 14*10 (*nbody * 10*)
and hence adds to another 140 elements in the state space.
- *cvel:* Center of mass based velocity. It has shape 14 * 6 (*nbody * 6*) and hence
adds another 84 elements in the state space
- *qfrc_actuator:* Constraint force generated as the actuator force. This has shape
`(23,)` *(nv * 1)* and hence adds another 23 elements to the state space.
- *cfrc_ext:* This is the center of mass based external force on the body. It has shape
14 * 6 (*nbody * 6*) and hence adds to another 84 elements in the state space.
where *nbody* stands for the number of bodies in the robot and *nv* stands for the number
of degrees of freedom (*= dim(qvel)*)
The (x,y,z) coordinates are translational DOFs while the orientations are rotational
DOFs expressed as quaternions. One can read more about free joints on the
[Mujoco Documentation](https://mujoco.readthedocs.io/en/latest/XMLreference.html).
**Note:** HumanoidStandup-v4 environment no longer has the following contact forces issue.
If using previous HumanoidStandup versions from v4, there have been reported issues that using a Mujoco-Py version > 2.0 results
in the contact forces always being 0. As such we recommend to use a Mujoco-Py version < 2.0
when using the Humanoid environment if you would like to report results with contact forces
(if contact forces are not used in your experiments, you can use version > 2.0).
### Rewards
The reward consists of three parts:
- *uph_cost*: A reward for moving upward (in an attempt to stand up). This is not a relative
reward which measures how much upward it has moved from the last timestep, but it is an
absolute reward which measures how much upward the Humanoid has moved overall. It is
measured as *(z coordinate after action - 0)/(atomic timestep)*, where *z coordinate after
action* is index 0 in the state/index 2 in the table, and *atomic timestep* is the time for
one frame of movement even though the simulation has a framerate of 5 (done in order to inflate
rewards a little for faster learning).
- *quad_ctrl_cost*: A negative reward for penalising the humanoid if it has too large of
a control force. If there are *nu* actuators/controls, then the control has shape `nu x 1`.
It is measured as *0.1 **x** sum(control2)*.
- *quad_impact_cost*: A negative reward for penalising the humanoid if the external
contact force is too large. It is calculated as *min(0.5 * 0.000001 * sum(external
contact force2), 10)*.
The total reward returned is ***reward*** *=* *uph_cost + 1 - quad_ctrl_cost - quad_impact_cost*
### Starting State
All observations start in state
(0.0, 0.0, 0.105, 1.0, 0.0 ... 0.0) with a uniform noise in the range of
[-0.01, 0.01] added to the positional and velocity values (values in the table)
for stochasticity. Note that the initial z coordinate is intentionally selected
to be low, thereby indicating a laying down humanoid. The initial orientation is
designed to make it face forward as well.
### Episode End
The episode ends when any of the following happens:
1. Truncation: The episode duration reaches a 1000 timesteps
2. Termination: Any of the state space values is no longer finite
### Arguments
No additional arguments are currently supported.
```
env = gym.make('HumanoidStandup-v4')
```
There is no v3 for HumanoidStandup, unlike the robot environments where a v3 and
beyond take gym.make kwargs such as xml_file, ctrl_cost_weight, reset_noise_scale etc.
### Version History
* v4: all mujoco environments now use the mujoco bindings in mujoco>=2.1.3
* v3: support for gym.make kwargs such as xml_file, ctrl_cost_weight, reset_noise_scale etc. rgb rendering comes from tracking camera (so agent does not run away from screen)
* v2: All continuous control environments now use mujoco_py >= 1.50
* v1: max_time_steps raised to 1000 for robot based tasks. Added reward_threshold to environments.
* v0: Initial versions release (1.0.0)
"""
metadata = {
"render_modes": [
"human",
"rgb_array",
"depth_array",
],
"render_fps": 67,
}
def __init__(self, **kwargs):
observation_space = Box(
low=-np.inf, high=np.inf, shape=(376,), dtype=np.float64
)
MujocoEnv.__init__(
self,
"humanoidstandup.xml",
5,
observation_space=observation_space,
**kwargs
)
utils.EzPickle.__init__(self, **kwargs)
def _get_obs(self):
data = self.data
return np.concatenate(
[
data.qpos.flat[2:],
data.qvel.flat,
data.cinert.flat,
data.cvel.flat,
data.qfrc_actuator.flat,
data.cfrc_ext.flat,
]
)
def step(self, a):
self.do_simulation(a, self.frame_skip)
pos_after = self.data.qpos[2]
data = self.data
uph_cost = (pos_after - 0) / self.model.opt.timestep
quad_ctrl_cost = 0.1 * np.square(data.ctrl).sum()
quad_impact_cost = 0.5e-6 * np.square(data.cfrc_ext).sum()
quad_impact_cost = min(quad_impact_cost, 10)
reward = uph_cost - quad_ctrl_cost - quad_impact_cost + 1
if self.render_mode == "human":
self.render()
return (
self._get_obs(),
reward,
False,
False,
dict(
reward_linup=uph_cost,
reward_quadctrl=-quad_ctrl_cost,
reward_impact=-quad_impact_cost,
),
)
def reset_model(self):
c = 0.01
self.set_state(
self.init_qpos + self.np_random.uniform(low=-c, high=c, size=self.model.nq),
self.init_qvel
+ self.np_random.uniform(
low=-c,
high=c,
size=self.model.nv,
),
)
return self._get_obs()
def viewer_setup(self):
assert self.viewer is not None
self.viewer.cam.trackbodyid = 1
self.viewer.cam.distance = self.model.stat.extent * 1.0
self.viewer.cam.lookat[2] = 0.8925
self.viewer.cam.elevation = -20