Expand description
Closed-Loop RRT* (CL-RRT*) path planner
Extends RRT*-Reeds-Shepp by adding a forward-simulation verification step. After the tree is grown, candidate goal paths are tracked using a pure-pursuit controller on a unicycle (bicycle-kinematic) model. Only paths that are dynamically feasible — the controller reaches the goal, the tracked path does not collide, and the travel distance is reasonable — are accepted.
The planner returns the full simulated trajectory (states over time) of the best feasible path, not just the geometric waypoints.
§References
- Kuwata, Y. et al. (2009). “Real-Time Motion Planning With Applications to Autonomous Urban Driving.” IEEE T-CST.
- PythonRobotics
ClosedLoopRRTStar/by Atsushi Sakai.
Structs§
- Closed
LoopRRT Star Config - Configuration for the Closed-Loop RRT* planner.
- Closed
LoopRRT Star Planner - Closed-Loop RRT* planner.
- Closed
LoopRRT Star Result - Result of a successful CL-RRT* planning run.
- Pure
Pursuit Params - Parameters for the pure-pursuit path tracker.
- Simulation
Result - Result of a closed-loop forward simulation.
- Unicycle
Params - Parameters for the bicycle-kinematic (unicycle) vehicle model.
- Vehicle
State - Vehicle state for the unicycle model.