nieNSrSSKrSSKJrJr SrSr"SS\RS9rg) zThe abstract robot class.N)OptionalSequencelinear_velocityangular_velocityc\rSrSrSr\R SS\\\ S\\\ SS4Sjj5r \R S5r \R S 5r \R S 5r \R S 5r\R S 5r\S 5r\\R S55r\S5r\S5r\S5r\S5r\S5rSrg) RobotBase z8The base class for all robots used in the mobility team.N base_positionbase_orientation_quaternionreturncg)a#Resets the states (e.g. pose and sensor readings) of the robot. This is called at the start of each episode by the environment. Args: base_position: Robot base position after reset. If None, robot stay where it was after reset. For robot that does not support reset with position change, a ValueError should be raised. base_orientation_quaternion: Robot base orientation after reset. If None, robot stays in pre-reset orientation. For robot that does not support reset with orientation change, a ValueError should be raised. N)selfr r s b/home/james-whalen/.local/lib/python3.13/site-packages/pybullet_envs/minitaur/robots/robot_base.pyresetRobotBase.resets" cg)zShuts down the robot.Nrrs r terminateRobotBase.terminate" rcg)aProcesses the input action before the action repeat loop. We assume that an action sent to the real robot is sticky, i.e. it will be executed until a new action is received after some time. To simulate this, we introduced the action_repeat parameter, to reflect how many time steps it takes for the policy to generate a new action. That is, for each control step, the simulation contains an inner loop: robot.pre_control_step(action) # Smooth or interpolate the action for i in range(action_repeat): robot.apply_action(action) bullet.stepSimulation(time_step) # Step the sim for one time step robot.receive_observation() # Update the sensor observations robot.post_control_step() # Update some internal variables. Args: action: Data type depends on the robot. Can be desired motor position/torques for legged robots, or desired velocity/angular velocity for wheeled robots. Nrractions rpre_control_stepRobotBase.pre_control_step's, rcg)z Applies the action to the robot.Nrrs r apply_actionRobotBase.apply_action?rrcg)z"Updates the robot sensor readings.Nrrs rreceive_observationRobotBase.receive_observationDrrcg)z6Updates some internal variables such as step counters.Nrrs rpost_control_stepRobotBase.post_control_stepIrrc[S5e)zThe action spec of the robot.zaction_space is not implementedNotImplementedErrorrs r action_spaceRobotBase.action_spaceNs ? @@rcg)a9Name of each action in the action_space. This is a structure of strings with the same shape as the action space, where each string describes the corresponding element of the action space (for example, a kinematic robot might return ("linear_velocity", "angular_velocity")). Used for logging in the safety layer. Nrrs r action_namesRobotBase.action_namesSsrc[S5e)aReturns the sensors on this robot. Sensors are the main interface between the robot class and the gym environment. Sensors can return what the robot can measure (e.g. joint angles, IMU readings), and can represent more general quantities, i.e. the last action taken, that can be part of the observation space. Sensor classes are used by the robot class to the specify its observation space. z sensors property not implementedr(rs rsensorsRobotBase.sensors^s @ AArc[S5e)aReturns the base pose as a quaternion in format (x, y, z, w). These properties differ from the sensor interfaces, as they represent the built-in measurable quantities. We assume most robots have an IMU at its base to measure the base pose. Actually, some sensor classes like the base pose sensor and joint angle sensor will call these built-in methods. In general, how these quantities can be extracted depends on the specific real robots. z.base_orientation_quaternion is not implementedr(rs rr %RobotBase.base_orientation_quaternionls N OOrc[S5e)z-Returns the base roll, pitch, and yaw angles.z&base_roll_pitch_yaw is not implementedr(rs rbase_roll_pitch_yawRobotBase.base_roll_pitch_yawzs F GGrc[S5e)Nz+base_roll_pitch_yaw_rate is not implementedr(rs rbase_roll_pitch_yaw_rate"RobotBase.base_roll_pitch_yaw_rates K LLrc[S5e)Nz base_position is not implementedr(rs rr RobotBase.base_positions @ AArr)NN)__name__ __module__ __qualname____firstlineno____doc__abcabstractmethodrrfloatrrrrr"r%propertyr*r-r0r r5r8r __static_attributes__rrrrr s}@26?C huo. $,HUO#< IM  $    .       A A    B  B  P  P H H M M B Brr) metaclass) r@rAtypingrrLINEAR_VELOCITYANGULAR_VELOCITYABCMetarrrrrKs/ %$%yB#++yBr