A dynamic programming approach to robotic swarm navigation using relay markers

Abstract

This work presents a dynamic programming based search and acquires method for application to robotic swarms. The method provides local rules of motion for swarm members that adhere to a global principle for both searching and converging on a stationary target in an unknown and complex environment via the use of immobile relay markers. By serving as approximations of the cost-to-go during the acquire phase, the marker locations direct the global motion of the swarm. This work presents those rules during a search phase, a dynamic programming phase, and an acquire phase which adhere to the principle of optimality. A simulation of a swarm locating a stationary target in a two-dimensional maze demonstrates the effectiveness of the method.