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

MultiFall

Browse Source
This commit is contained in:
Yuji Kanagawa 2021-09-28 16:42:22 +09:00
parent d5711f9969
commit 71f89461ee
1 changed files with 33 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
mujoco_maze/maze_task.py
View File

@ -255,6 +255,38 @@ class DistRewardFall(GoalRewardFall, DistRewardMixIn):
pass
class GoalRewardMultiFall(GoalRewardUMaze):
MAZE_SIZE_SCALING: Scaling = Scaling(ant=2.0, point=6.0, swimmer=None)
OBSERVE_BLOCKS: bool = True
def __init__(self, scale: float, goal: Tuple[int, int] = (0.0, 3.0)) -> None:
super().__init__(scale)
self.goals = [MazeGoal(np.array([*goal, 0.5]) * scale)]
@staticmethod
def create_maze() -> List[List[MazeCell]]:
E, B, C, R = MazeCell.EMPTY, MazeCell.BLOCK, MazeCell.CHASM, MazeCell.ROBOT
M = MazeCell.YZ_BLOCK
return [
[B, B, B, B, B, B, B, B, B],
[B, E, C, E, E, E, C, E, B],
[B, E, C, E, R, E, C, E, B],
[B, E, C, E, M, E, C, E, B],
[B, B, B, C, C, C, B, B, B],
[B, B, B, E, E, E, B, B, B],
[B, B, B, B, B, B, B, B, B],
]
class DistRewardMultiFall(GoalRewardMultiFall, DistRewardMixIn):
pass
class NoRewardMultiFall(GoalRewardFall):
def reward(self, _obs: np.ndarray) -> float:
return 0.0
class GoalReward2Rooms(MazeTask):
REWARD_THRESHOLD: float = 0.9
PENALTY: float = -0.0001
@ -594,6 +626,7 @@ class TaskRegistry:
"Push": [DistRewardPush, GoalRewardPush],
"MultiPush": [DistRewardMultiPush, GoalRewardMultiPush, NoRewardMultiPush],
"Fall": [DistRewardFall, GoalRewardFall],
"MultiFall": [DistRewardMultiFall, GoalRewardMultiFall, NoRewardMultiFall],
"2Rooms": [DistReward2Rooms, GoalReward2Rooms, SubGoal2Rooms],
"4Rooms": [DistReward4Rooms, GoalReward4Rooms, SubGoal4Rooms],
"TRoom": [DistRewardTRoom, GoalRewardTRoom, SubGoalTRoom],