mirror of
https://github.com/Farama-Foundation/Gymnasium.git
synced 2025-08-01 06:07:08 +00:00
47 lines
1.7 KiB
Python
47 lines
1.7 KiB
Python
import numpy as np
|
|
from gym import utils
|
|
from gym.envs.mujoco import mujoco_env
|
|
|
|
class AntEnv(mujoco_env.MujocoEnv, utils.EzPickle):
|
|
def __init__(self):
|
|
mujoco_env.MujocoEnv.__init__(self, 'ant.xml', 5)
|
|
utils.EzPickle.__init__(self)
|
|
self.finalize()
|
|
|
|
def _step(self, a):
|
|
xposbefore = self.get_body_com("torso")[0]
|
|
self.do_simulation(a, self.frame_skip)
|
|
xposafter = self.get_body_com("torso")[0]
|
|
forward_reward = (xposafter - xposbefore)/self.dt
|
|
ctrl_cost = .5 * np.square(a).sum()
|
|
contact_cost = 0.5 * 1e-3 * np.sum(
|
|
np.square(np.clip(self.model.data.cfrc_ext, -1, 1)))
|
|
survive_reward = 1.0
|
|
reward = forward_reward - ctrl_cost - contact_cost + survive_reward
|
|
state = self._state
|
|
notdone = np.isfinite(state).all() \
|
|
and state[2] >= 0.2 and state[2] <= 1.0
|
|
done = not notdone
|
|
ob = self._get_obs()
|
|
return ob, reward, done, dict(
|
|
reward_forward=forward_reward,
|
|
reward_ctrl=-ctrl_cost,
|
|
reward_contact=-contact_cost,
|
|
reward_survive=survive_reward)
|
|
|
|
def _get_obs(self):
|
|
return np.concatenate([
|
|
self.model.data.qpos.flat[2:],
|
|
self.model.data.qvel.flat,
|
|
np.clip(self.model.data.cfrc_ext, -1, 1).flat,
|
|
])
|
|
|
|
def _reset(self):
|
|
self.model.data.qpos = self.init_qpos + np.random.uniform(size=(self.model.nq,1),low=-.1,high=.1)
|
|
self.model.data.qvel = self.init_qvel + np.random.randn(self.model.nv,1)*.1
|
|
self.reset_viewer_if_necessary()
|
|
return self._get_obs()
|
|
|
|
def viewer_setup(self):
|
|
self.viewer.cam.distance = self.model.stat.extent * 0.5
|