PRM

Table of contents

1. Core idea
2. Why use PRM
3. Common variants

---

Probabilistic Roadmap Methods (PRM) build a reusable graph of collision-free configurations, then search the graph for a feasible path.

Core idea

1. Sample many collision-free configurations.
2. Connect nearby samples with local planners.
3. Build a roadmap graph offline.
4. For each query, connect start/goal and run graph search.

Why use PRM

• Good for multi-query planning problems
• Scales well in high-dimensional configuration spaces
• Reuses the same roadmap across tasks

Common variants

• Basic PRM
• Lazy PRM
• PRM*