dm_robotics.panda.environment module

RL environment builder for the Panda MoMa model.

class dm_robotics.panda.environment.PandaEnvironment(robot_params: ~dm_robotics.panda.parameters.RobotParams | ~typing.Sequence[~dm_robotics.panda.parameters.RobotParams], arena: ~dm_control.composer.arena.Arena = <dm_robotics.moma.models.arenas.empty.Arena object>, control_timestep: float = 0.1, physics_timestep: float = 0.002)

Bases: object

Adds Panda robots to an arena and creates a subtask environment.

__init__(robot_params: ~dm_robotics.panda.parameters.RobotParams | ~typing.Sequence[~dm_robotics.panda.parameters.RobotParams], arena: ~dm_control.composer.arena.Arena = <dm_robotics.moma.models.arenas.empty.Arena object>, control_timestep: float = 0.1, physics_timestep: float = 0.002) → None

add_entity_initializers(initializers: Sequence[Initializer])

Add entity initializers.

add_extra_effectors(extra_effectors: Sequence[Effector])

Add extra effectors.

add_extra_sensors(extra_sensors: Sequence[Sensor])

Add extra sensor.

add_props(props: Sequence[Prop])

Add props as free entities to the arena.

add_scene_initializers(initializers: Sequence[Callable[[RandomState], None]])

Add scene initializers.

add_timestep_preprocessors(preprocessors: Sequence[TimestepPreprocessor])

Add timestep preprocessors.

build_task_environment() → SubTaskEnvironment

Builds a subtask environment.

The environment includes the configured robots as well as any added components such as props, initializers, preprocessors etc.

property robots: Dict[str, Robot]

A map of the configured robots indexed by name.