Multi-robot foraging based on Darwin's survival of the fittest

Abstract

This paper presents a collective foraging algorithm designed to simulate natural selection in a group of swarm robots. The Robotic Darwinian Particle Swarm Optimization (RDPSO) previously proposed is improved using fractional calculus theory and evaluated on real low-cost mobile robots performing a distributed foraging task. This work aims at evaluating this novel exploration strategy, by studying the performance of the algorithm within a population of up to 12 robots, under communication constraints. In order to simulate the maximum allowed communication distance, robots were provided with a list of their teammates' addresses. Experimental results show that only 4 robots are needed to accomplish the proposed mission and, independently on the number of robots, maximum communication distance and fractional coefficient, the optimal solution is achieved in approximately 90% of the experiments.