Research on AGV composition algorithm based on variable and complex environment

Abstract

Aiming at the traditional Gmapping algorithm with constant number of particles, which lead to the failure of composition in variable and complex environments, based on SLAM mapping technology, this paper proposed an Adaptive Sampling (AS) algorithm, which increased the number of sampling particles when the fluctuation of the 2D laser point cloud is larger than a certain threshold, and the experimental results shown that the algorithm was able to reasonably utilize the system resources, and effectively enhanced the algorithm’s ability to compose maps in a variable and complex environment. In addition, in order to make the system have a better effect of building maps in complex environments, this paper also integrated the firefly algorithm (AF) with it, and utilized AF’s high aggregation ability to improve the distribution of sampling particles, which then improved the estimation ability of the filter. Verified by offline real environment experiments, the results shown that the optimized algorithm significantly improves the map building ability of the system.