fancy_gym/fancy_gym/meta
2022-11-01 22:51:43 +01:00
..
__init__.py add tests for replanning env & adapt observation space for box pushing & add max_planning_times to replanning tasks 2022-11-01 22:51:43 +01:00
base_metaworld_mp_wrapper.py updated for gym 0.25.1 2022-08-18 09:04:38 +02:00
goal_change_mp_wrapper.py renaming to fancy_gym 2022-07-13 15:10:43 +02:00
goal_endeffector_change_mp_wrapper.py renaming to fancy_gym 2022-07-13 15:10:43 +02:00
goal_object_change_mp_wrapper.py renaming to fancy_gym 2022-07-13 15:10:43 +02:00
object_change_mp_wrapper.py added black box tests and bug fixes 2022-10-21 16:16:49 +02:00
README.MD Added more tests 2022-10-14 14:02:12 +02:00

MetaWorld Wrappers

These are the Environment Wrappers for selected Metaworld environments in order to use our Movement Primitive gym interface with them. All Metaworld environments have a 39 dimensional observation space with the same structure. The tasks differ only in the objective and the initial observations that are randomized. Unused observations are zeroed out. E.g. for Button-Press-v2 the observation mask looks the following:

    return np.hstack([
        # Current observation
        [False] * 3,  # end-effector position
        [False] * 1,  # normalized gripper open distance
        [True] * 3,  # main object position
        [False] * 4,  # main object quaternion
        [False] * 3,  # secondary object position
        [False] * 4,  # secondary object quaternion
        # Previous observation
        [False] * 3,  # previous end-effector position
        [False] * 1,  # previous normalized gripper open distance
        [False] * 3,  # previous main object position
        [False] * 4,  # previous main object quaternion
        [False] * 3,  # previous second object position
        [False] * 4,  # previous second object quaternion
        # Goal
        [True] * 3,  # goal position
    ])

For other tasks only the boolean values have to be adjusted accordingly.