Dynamic Window Approach: Local Obstacle Avoidance

A differential-drive robot navigates a field of static obstacles toward a goal using the Dynamic Window Approach — the classic real-time local planner behind most mobile-robot obstacle avoidance. At every control step the planner samples the (linear speed, turn rate) pairs the robot can physically reach given its acceleration limits, forward-simulates a short arc for each, discards any arc that would collide, and scores the rest on how well they point at the goal, how much clearance they keep from obstacles, and how fast they move — then drives the winning arc. Tune the speed limit, angular acceleration, and clearance weight to see the trade-off between a robot that charges straight at the goal and one that gives obstacles a wide berth.