Design of decentralized chemotactic control law for area coverage using swarm of mobile robots

Abstract

Design of control laws for robots operating cooperatively, whereby individual tasks need to be allocated among available agents, is critical to the overall performance of the system. Some biological examples of this type of system in nature, such as ant colonies, use chemotactic, or chemical pheromone-driven, mechanisms to control the behavior of each individual agent. This work uses a set of reaction-diffusion equations to form a decentralized control law which results in emergent area coverage behavior. The influence of pheromone diffusion and evaporation, which are environmental mechanisms affecting the chemical pheromone, is studied and the resulting area coverage performance is quantified. Likewise, the impact on area coverage performance due to the addition of noise to the control law is studied. The contribution of this work is to show how these parameters affect area coverage performance. Results show that the magnitude of noise added is the most critical component affecting the performance of the algorithm in exhaustively covering the given area.