Research On Indoor AGV Based On Simultaneous Localization And Mapping

Automated Guided Vehicle, also known as the Handling Robot, is an important link of modern intelligent logistics system. However, the current mainstreams of AGV guidance technology couldn’t make it work completely independently. Simultaneous Localization and Mapping algorithm, which has great theoretical research and practical application value, has been developing rapidly since it is firstly put forward by Smith Self and Cheese man. It has been successfully applied in the outdoor, indoor, underwater and aerial environments, etc. Using SLAM ideas to solve the AGV navigation problems, ensures the safety and positioning precision of AGV, and also lays the foundation of the efficient path planning and control decisions determine.This thesis studies the algorithm of indoor AGV simultaneous localization and mapping. Firstly, briefly introduces the most commonly used navigation mode of AGV at present, and discusses the superiority and defects of them respectively; then, analyzes details of SLAM applied in the AGV navigation field; Through a series of model establishment progress, build the foundation of SLAM for AGV; finally, focuses on the various SLAM algorithm, including the extended Kalman filter (EKF) algorithm used in nonlinear system, the particle filter (PF) algorithm based on the probability, FastSLAM algorithm based on RBPF. Aiming at the shortcomings of the traditional method, this paper proposes an improved particle filter method based on improved particle swarm optimization algorithm. The method is a kind of FastSLAM algorithm based on the Rao-Blackwellized (RBPF), taking into account the influence of optimal solution of both the individual particles and group particle. Mutation operators of genetic algorithm are used to replace the smaller values particles by new ones and adjust the obtained particle sets. The simulation results show that the new algorithm effectively improve the positioning precision of AGV; keep the diversity of the particles and improve the global search ability of the algorithm. At last chapter, the full text is summarized and further research direction in the future is prospected.