dodox/mujoco_maze
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

fix stuff

Browse Source
This commit is contained in:
Dominik Moritz Roth 2024-04-04 15:08:45 +02:00
parent a19f232c03
commit 9b3a5ccca5
3 changed files with 10 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
mujoco_maze/ant.py
View File

@ -98,7 +98,7 @@ class AntEnv(AgentModel):
low=-0.1,
high=0.1,
)
qvel = self.init_qvel + self.np_random.randn(self.model.nv) * 0.1
qvel = self.init_qvel + self.np_random.random(self.model.nv) * 0.1
# Set everything other than ant to original position and 0 velocity.
qpos[15:] = self.init_qpos[15:]

2
mujoco_maze/assets/swimmer.xml
View File

@ -13,7 +13,7 @@
<material name='geom' texture="texgeom" texuniform="true" />
</asset>
<worldbody>
<light cutoff="100" diffuse="1 1 1" dir="-0 0 -1.3" directional="true" exponent="1" pos="0 0s 1.3" specular=".1 .1 .1" />
<light cutoff="100" diffuse="1 1 1" dir="-0 0 -1.3" directional="true" exponent="1" pos="0 0 1.3" specular=".1 .1 .1" />
<geom conaffinity="1" condim="3" material="MatPlane" name="floor" pos="0 0 -0.1" rgba="0.8 0.9 0.8 1" size="40 40 0.1" type="plane" />
<!-- ================= SWIMMER ================= /-->
<body name="torso" pos="0 0 0">

16
mujoco_maze/swimmer.py
View File

@ -24,7 +24,7 @@ class SwimmerEnv(AgentModel):
"rgb_array",
"depth_array",
],
"render_fps": 10
"render_fps": 25
}
def __init__(
@ -41,12 +41,12 @@ class SwimmerEnv(AgentModel):
super().__init__(file_path, 4, self.observation_space)
def _forward_reward(self, xy_pos_before: np.ndarray) -> Tuple[float, np.ndarray]:
xy_pos_after = self.sim.data.qpos[:2].copy()
xy_pos_after = self.data.qpos[:2].copy()
xy_velocity = (xy_pos_after - xy_pos_before) / self.dt
return self._forward_reward_fn(xy_velocity)
def step(self, action: np.ndarray) -> Tuple[np.ndarray, float, bool, dict]:
xy_pos_before = self.sim.data.qpos[:2].copy()
xy_pos_before = self.data.qpos[:2].copy()
self.do_simulation(action, self.frame_skip)
forward_reward = self._forward_reward(xy_pos_before)
ctrl_cost = self._ctrl_cost_weight * np.sum(np.square(action))
@ -58,8 +58,8 @@ class SwimmerEnv(AgentModel):
)
def _get_obs(self) -> np.ndarray:
position = self.sim.data.qpos.flat.copy()
velocity = self.sim.data.qvel.flat.copy()
position = self.data.qpos.flat.copy()
velocity = self.data.qvel.flat.copy()
observation = np.concatenate([position, velocity]).ravel()
return observation
@ -79,9 +79,9 @@ class SwimmerEnv(AgentModel):
return self._get_obs()
def set_xy(self, xy: np.ndarray) -> None:
qpos = self.sim.data.qpos.copy()
qpos = self.data.qpos.copy()
qpos[:2] = xy
self.set_state(qpos, self.sim.data.qvel)
self.set_state(qpos, self.data.qvel)
def get_xy(self) -> np.ndarray:
return np.copy(self.sim.data.qpos[:2])
return np.copy(self.data.qpos[:2])