|
LIRRA
1.0.0
|
Public Member Functions | |
| __init__ (self, LevelTables T, PatrolPD nominal, np.ndarray track_weight, float gamma0, int|None n_cost, float effort_weight, np.ndarray obs_mean, np.ndarray obs_std, PlanEnvConfig cfg=PlanEnvConfig()) | |
| int | observation_size (self) |
| int | action_size (self) |
| nominal_rollout (self, x0, idx) | |
| clock_reference (self, x0, idx, int length) | |
| mjx_env.State | reset (self, jax.Array rng) |
| mjx_env.State | step (self, mjx_env.State state, jax.Array action) |
Public Attributes | |
| T | |
| cfg | |
| obs_mean | |
| obs_std | |
| A | |
| B | |
| C | |
| cost_w | |
| track_w | |
| effort_w | |
| pos | |
| vel | |
| inp | |
| kp | |
| kd | |
| n_ax | |
| n_wp | |
| reg | |
| wps | |
| wp_tol | |
| leg_a | |
| leg_ab | |
| u_lo | |
| u_hi | |
| clock | |
Protected Member Functions | |
| _advance (self, x, idx) | |
| _u_nom (self, x, idx) | |
| _u_stop (self, x) | |
| _reference (self, s) | |
| _patrol_s (self, p) | |
| _waypoint_pass (self, p0, p1) | |
| _obs (self, s) | |
| _plan (self, s, action) | |
Minimal Playground-compatible env (``reset``/``step``, ``observation_size``/``action_size``).
| rl_stl.envs.patrol_plan.PatrolPlanEnv.__init__ | ( | self, | |
| LevelTables | T, | ||
| PatrolPD | nominal, | ||
| np.ndarray | track_weight, | ||
| float | gamma0, | ||
| int | None | n_cost, | ||
| float | effort_weight, | ||
| np.ndarray | obs_mean, | ||
| np.ndarray | obs_std, | ||
| PlanEnvConfig | cfg = PlanEnvConfig() |
||
| ) |
|
protected |
|
protected |
|
protected |
Path parameter of ``p`` on the closed polyline through the waypoints, in ``[0, n_wp]`` (as :func:`rl_stl.evaluation.patrol._patrol_projection`).
|
protected |
|
protected |
``(x_nom (N+1, n), u_nom (N, m), waypoint index)`` for the plan rooted at ``s``.
|
protected |
|
protected |
Stopping input past the end of the cached plan: damp the planar velocity, hold set-points.
|
protected |
``(forward pass, backward pass, miss [m])`` of the step ``p0 -> p1`` across a waypoint (at most one per step; the waypoint legs are much longer than one step).
| int rl_stl.envs.patrol_plan.PatrolPlanEnv.action_size | ( | self | ) |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.clock_reference | ( | self, | |
| x0, | |||
| idx, | |||
| int | length | ||
| ) |
Clock reference from the episode start: ``X (length+1, n)``, ``U (length, m)``, and the waypoint index ``I (length,)`` used at each step (as :class:`ClockReference`).
| rl_stl.envs.patrol_plan.PatrolPlanEnv.nominal_rollout | ( | self, | |
| x0, | |||
| idx | |||
| ) |
| int rl_stl.envs.patrol_plan.PatrolPlanEnv.observation_size | ( | self | ) |
| mjx_env.State rl_stl.envs.patrol_plan.PatrolPlanEnv.reset | ( | self, | |
| jax.Array | rng | ||
| ) |
| mjx_env.State rl_stl.envs.patrol_plan.PatrolPlanEnv.step | ( | self, | |
| mjx_env.State | state, | ||
| jax.Array | action | ||
| ) |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.A |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.B |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.C |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.cfg |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.clock |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.cost_w |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.effort_w |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.inp |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.kd |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.kp |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.leg_a |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.leg_ab |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.n_ax |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.n_wp |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.obs_mean |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.obs_std |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.pos |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.reg |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.T |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.track_w |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.u_hi |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.u_lo |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.vel |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.wp_tol |
| rl_stl.envs.patrol_plan.PatrolPlanEnv.wps |