Autonomous Task Allocation for Swarm Robotic Systems Using Behavioral Decomposition

Abstract

Swarm robotic systems (SRS) are a type of multi-robot systems, in which robots operate without any form of centralized control. In SRS, the generation of a complex swarm behavior resulting in robots being dynamically distributed over different sub-tasks requires an autonomous task allocation mechanism. It has been well recognized that evolutionary robotics with an evolving artificial neural network is a promising approach for generating collective swarm behavior. However, the artificial evolution often suffers from the bootstrap problem, especially when the underlying task is very complex. On the other hand, the behavioral decomposition, which is based on the divide-and-conquer thinking, has been reported to be effective for overcoming the bootstrap problem. In this paper, we describe how a behavioral decomposition based evolutionary robotics approach can be applied to synthesize a composite artificial neural network based controller for a complex task. The simulation results show the hierarchical strategy based evolutionary robotics approach is effective for generating autonomous task allocation behavior for a swarm robotic system.