Actions

Manager: ActionManager · Module: env/mdp/actions.py

Default action

ReferenceJointPositionActionCfg — the policy outputs residual joint deltas Δq that are added to reference joint positions from the motion command:

q_target = q_ref(motion command) + scale · action

This keeps the action space centered on the current clip frame rather than absolute joint targets.

Per-robot wiring

Each robot entity sets actions["joint_pos"].scale to a per-DoF or grouped scale tensor (e.g. <ROBOT>_ACTION_SCALE in the robot package).

Alternative

make_base_wbc_env_cfg(use_reference_residual_action=False) switches to mjlab’s absolute JointPositionActionCfg with default joint offsets.

API: MDP terms (Actions).

Related: Motion command, The robot entity.