Path planning computes a route for a robot to reach a goal. It considers obstacles, the robot's shape, and the cost of different routes. The output is a sequence of waypoints or a continuous path that the robot can follow. Motion planning adds timing and dynamics. Path planning is about where, not when.

The classic algorithms are Dijkstra and A*. Dijkstra finds the shortest path in a graph. A* adds a heuristic to speed up the search. On a grid map, A* is fast and effective. On a continuous map, sampling-based planners like RRT and PRM explore the space by randomly sampling configurations and connecting them.

Path planning considerations

• Shortest path: minimize distance.
• Safest path: maximize clearance from obstacles.
• Smoothest path: minimize turns and acceleration.
• Energy-efficient path: minimize uphill travel or braking.
• Dynamic obstacles: replan as the environment changes.

A path that is short may be unsafe. A path that is safe may be long. The planner optimizes a cost function that weights these factors. In a warehouse, the cost might be time. In a surgical robot, the cost might be deviation from a planned trajectory. In a planetary rover, the cost might be risk of getting stuck. Path planning is the bridge between a map and a movement. Without it, the robot knows where it wants to go but not how to get there. With it, the robot can navigate complex environments without human intervention.