Hybrid Topology-Based Particle Swarm Optimizer for Multi-source Location Problem in Swarm Robots

Junqi Zhang, Yehao Lu, Mengchu Zhou

Abstract: A multi-source location problem aims to locate sources in an unknown environment based on the measurements of the signal strength from them. Vast majority of existing multi-source location methods require such prior environmental information as the signal range of sources and maximum signal strength to set some parameters. However, prior information is difficult to obtain in many practical tasks. To handle this issue, this work proposes a variant of Particle Swarm Optimizers (PSO), named as Hybrid Topology-based PSO (HT-PSO). It combines the advantages of multimodal search capability of a ring topology and rapid convergence of a star topology. HT-PSO does not require any prior knowledge of the environment, thus it has stronger robustness and adaptability. Experimental results show its superior performance over the state-of-the-art multi-source location method.