panda-gym includes:

  • 1 robot:
    • the Franka Emika Panda robot,

  • 6 tasks:
    • Reach: the robot must place its end-effector at a target position,

    • Push: the robot has to push a cube to a target position,

    • Slide: the robot has to slide an object to a target position,

    • Pick and place: the robot has to pick up and place an object at a target position,

    • Stack: the robot has to stack two cubes at a target position,

    • Flip: the robot must flip the cube to a target orientation,

  • 2 control modes:
    • End-effector displacement control: the action corresponds to the displacement of the end-effector.

    • Joints control: the action corresponds to the individual motion of each joint,

  • 2 reward types:
    • Sparse: the environment return a reward if and only if the task is completed,

    • Dense: the closer the agent is to completing the task, the higher the reward.

By default, the reward is sparse and the control mode is the end-effector displacement. The complete set of environments present in the package is presented in the following list.

Sparce reward, end-effector control (default setting)

  • PandaReach-v3

  • PandaPush-v3

  • PandaSlide-v3

  • PandaPickAndPlace-v3

  • PandaStack-v3

  • PandaFlip-v3

Dense reward, end-effector control

  • PandaReachDense-v3

  • PandaPushDense-v3

  • PandaSlideDense-v3

  • PandaPickAndPlaceDense-v3

  • PandaStackDense-v3

  • PandaFlipDense-v3

Sparce reward, joints control

  • PandaReachJoints-v3

  • PandaPushJoints-v3

  • PandaSlideJoints-v3

  • PandaPickAndPlaceJoints-v3

  • PandaStackJoints-v3

  • PandaFlipJoints-v3

Dense reward, joints control

  • PandaReachJointsDense-v3

  • PandaPushJointsDense-v3

  • PandaSlideJointsDense-v3

  • PandaPickAndPlaceJointsDense-v3

  • PandaStackJointsDense-v3

  • PandaFlipJointsDense-v3