bam.mujoco#

Classes#

MujocoController

A MujocoController is a class allowing to apply the torque and update frictions

Functions#

load_config(→ tuple)

Loads a BAM configuration file and returns the list of controllers and the mapping dicts.

Module Contents#

class bam.mujoco.MujocoController(model: bam.model.Model, actuator: str, mujoco_model: mujoco.MjModel, mujoco_data: mujoco.MjData, vin_drop_gain: float | None = None, vin_min: float | None = None, max_current: float | None = None)#

A MujocoController is a class allowing to apply the torque and update frictions from the computed model during a simulation.

Parameters:

  • model (bam.Model) – Model to use (can be loaded using load_model)

  • actuator (str) – Actuator to control. The actuated joint properties will be updated. This can be a list of actuators

  • mujoco_model (mujoco.MjModel) – The mujoco model

  • mujoco_data (mujoco.MjData) – The mujoco data

  • vin_drop_gain (float | None) – The voltage drop gain, if not None the voltage will be reduced by vin_drop_gain * load, where load is the sum of the absolute value of the motor torques

  • vin_min (float | None) – the minimum voltage, if not None the voltage will not go below this value

  • max_current (float | None) – Firmware current limit [A]. If not None, the motor current is clipped to [-max_current, max_current] (equivalently the motor torque is clipped to ±max_current * kt), reproducing the firmware-side current saturation of the motor.

model#
actuator#
mujoco_model#
mujoco_data#
vin_drop_gain = None#
vin_min = None#
max_current = None#
dofs = []#
q_target#
dof_to_q_target#
last_ts#
act_indexes#
joint_indexes#
qpos_indexes#
dof_indexes#
get_q_target(name: str) → float#

Return the current target position for a named actuator [rad].

Parameters:

name – Actuator name as passed to the constructor.

set_q_target(name: str, q_target: float)#

Set the target position for a named actuator.

Parameters:

  • name – Actuator name as passed to the constructor.

  • q_target – Desired joint angle [rad].

reset(qpos)#

Reset the controller to a given joint position state.

Should be called after every mujoco.mj_resetData to clear the internal target and voltage-drop state.

Parameters:

qpos – Full mj_data.qpos array. The controller extracts the positions of its controlled joints.

update()#

Update the controlled actuator(s) data: - Torque to apply - Friction parameters - Damping

bam.mujoco.load_config(path: str, mujoco_model: mujoco.MjModel, mujoco_data: mujoco.MjData, kp: float, vin: float) → tuple#

Loads a BAM configuration file and returns the list of controllers and the mapping dicts.

Parameters:

  • path (str) – path to the configuration file

  • mujoco_model (mujoco.MjModel) – the mujoco model

  • mujoco_data (mujoco.MjData) – the mujoco data

  • kp (float) – the proportional gain

  • vin (float) – the input voltage

Returns:

list of controllers, dofs to model mapping, dofs to id mapping

Return type:

list