Range-based Collision Prediction for Dynamic Motion

Real-time proximity and collision detection via radio distance measurements has application in smart-helmets, drones, autonomous vehicles, and social distancing. In this paper we present a range-based, infrastructure-free, distributed algorithm that utilizes inter-robot range data and intra-robot acceleration data to estimate each robot's recent positions. As the next step, a collision prediction scheme is proposed that uses the derived relative kinematics of each robot to predict an impending collision between any pair of robot. The algorithm is tested and validated with the help of measurement trace-based simulation of motion involving acceleration.