Path Planners: Probabilistic Road Map

[guilyx]

Probabilistic Road Map (PRM) Path Planner

Algorithm (src/components/plan/prm/prm_path_planner.py):

• Sampling phase: Uniformly samples 500 collision-free points in world space (seeded RNG for reproducibility)
• Connection phase: Connects each node to neighbours within a 5.0m radius via collision-free straight-line local paths (20-point line sampling)
• Search phase: Dijkstra on the resulting roadmap graph from start to goal
• Clearance-aware: Treats both obstacles (1.0) and clearance zones (0.75) as impassable
• Search GIF: 5-phase animation — sampling nodes, connecting edges, Dijkstra expansion, path drawing, hold

Simulation (src/simulations/path_planning/prm_path_planning/prm_path_planning.py):

• Same obstacle layout as other planners (two obstacles)
• Generates prm_search.gif (roadmap construction + search) and prm_navigate.gif (PurePursuit car following)