Enhanced Gmapping Algorithm Utilizing Improved Particle Swarm Optimization for Efficient SLAM Performance

Abstract

In recent decades, Simultaneous Localization and Mapping (SLAM) has emerged as a critical area of research within autonomous robotic systems, playing a fundamental role in navigation and environmental mapping. Traditional Rao-Blackwellized Particle Filter (RBPF)-based SLAM algorithms face challenges related to particle degradation and high computational cost, particularly as the number of particles increases. This paper addresses these issues by proposing an enhanced Gmapping algorithm based on improved particle swarm optimization (PSO). The proposed method introduces normal distribution and compression factors to optimize PSO convergence, improving localization accuracy while reducing the number of sampled particles. Experimental results demonstrate that the improved algorithm reduces localization error and enhances mapping accuracy, all while lowering hardware requirements. These findings highlight the potential for deploying this algorithm on cost-effective robotic platforms in future applications.