grutopia.core.robot

robot

class grutopia.core.robot.robot.BaseRobot(config: RobotCfg, scene: Scene)

Base class of robot.

apply_action(action: dict)

Apply actions of controllers to robot.

Parameters:

action (dict) – action dict. key: controller name. value: corresponding action array.

get_obs() → dict

Get observation of robot, including controllers, sensors, and world pose.

Raises:

NotImplementedError – _description_

get_robot_articulation() → Robot

Get isaac robots instance (articulation).

Returns:

robot articulation.

Return type:

Robot

get_robot_base() → RigidPrim

Get base link of robot.

Returns:

rigid prim of robot base link.

Return type:

RigidPrim

get_robot_scale() → ndarray

Get robot scale.

Returns:

robot scale in (x, y, z).

Return type:

np.ndarray

post_reset()

Set up things that happen after the world resets.

classmethod register(name: str)

Register a robot class with its name(decorator).

Parameters:

name (str) – name of the robot class.

set_up_to_scene(scene: Scene)

Set up robot in the scene.

Parameters:

scene (Scene) – scene to set up.

grutopia.core.robot.robot.create_robots(runtime: TaskRuntime, scene: Scene) → Dict[str, BaseRobot]

Create robot instances in runtime.

Parameters:

• runtime (TaskRuntime) – task runtime.

• scene (Scene) – isaac scene.

Returns:

robot instances dictionary.

Return type:

Dict[str, BaseRobot]

controller

class grutopia.core.robot.controller.BaseController(config: ControllerCfg, robot: BaseRobot, scene: Scene)

Base class of controller.

abstract action_to_control(action: ndarray | List) → ArticulationAction

Convert input action (in 1d array format) to joint signals to apply.

Parameters:

action (Union[np.ndarray, List]) – input control action.

Returns:

joint signals to apply

Return type:

ArticulationAction

cleanup()

Operations that need to be cleaned up before switching scenes (or resetting)

get_joint_subset() → ArticulationSubset | None

Get the joint subset controlled by the controller.

Returns:

joint subset.

Return type:

ArticulationSubset

get_obs() → Dict[str, Any]

Get observation of controller.

Returns:

observation key and value.

Return type:

Dict[str, Any]

classmethod register(name: str)

Register a controller with its name(decorator).

Parameters:

name (str) – name of the controller

grutopia.core.robot.controller.create_controllers(robot_cfg: RobotCfg, robot: BaseRobot, scene: Scene) → Dict[str, BaseController]

Create all controllers of one robot.

Parameters:

• robot_cfg (RobotCfg) – config of the robot.

• robot (BaseRobot) – robot instance.

• scene (Scene) – scene from isaac sim.

Returns:

dict of controllers with controller name as key.

Return type:

Dict[str, BaseController]

grutopia.core.robot.controller.inject_sub_controllers(parent: BaseController, configs: List[ControllerCfg], robot: BaseRobot, scene: Scene)

Recursively create and inject sub-controllers into parent controller.

Parameters:

• parent (BaseController) – parent controller instance.

• configs (List[ControllerParams]) – user configs of sub-controllers.

• robot (BaseRobot) – robot instance.

• scene (Scene) – scene from isaac sim.

sensor

class grutopia.core.robot.sensor.BaseSensor(config: SensorCfg, robot: BaseRobot, scene: Scene)

Base class of sensor.

cleanup()

Operations that need to be cleaned up before switching scenes (or resetting)

abstract get_data() → Dict

Get data from sensor.

Returns:

data dict of sensor.

Return type:

Dict

post_reset()

Post reset operations.

classmethod register(name: str)

Register a sensor class with the given name(decorator). :param name: name of the sensor class. :type name: str

grutopia.core.robot.sensor.create_sensors(robot_cfg: RobotCfg, robot: BaseRobot, scene: Scene) → Dict[str, BaseSensor]

Create all sensors of one robot.

Parameters:

• robot_cfg (RobotCfg) – config of the robot.

• robot (BaseRobot) – robot instance.

• scene (Scene) – scene from isaac sim.

Returns:

dict of sensors with sensor name as key.

Return type:

Dict[str, BaseSensor]