Abstract We present a collision avoidance technique developped in the case of a wheeled vehicle with tactile or telemetric exteroceptive sensors. The method don't assume any model of the obstacle; its principle consists in mapping all the inequalities concerning the predicted sensor-obstacle distances into linear constraints in the two-dimensional control space, yelding thus a convex polytope A of admissible displacements. A possible displacement is chosen as the projection of the planned elementary next motion into A , completed, in case of failure, by a secondary strategy that consists to go around the obstacle by its right side.

This collision avoidance technique for wheeled vehicles uses linear constraints in control space to map sensor-obstacle distances into admissible displacements, with a secondary strategy to go around obstacles if necessary.

A Collision Avoidance Method for Mobile Robots

Key Takeaway: This collision avoidance technique for wheeled vehicles uses linear constraints in control space to map sensor-obstacle distances into admissible displacements, with a secondary strategy to go around obstacles if necessary.