A SLAM Method to Improve the Performance of Unmanned Vehicle

Intelligent vehicle infrastructure cooperative system is the latest development in intelligent transportation, and autonomous navigation technology is the key technology in the field of intelligent transportation. Research on SLAM is especially essential. However, it is difficult to establish an accurate priori noise model based on Kalman filter for SLAM algorithm. This paper proposes a fast SLAM algorithm based on the improved geese particle swarm optimization algorithm, which effectively improves the local part of the particle optimal and precocious problems. The algorithm of the geese group particle swarm optimization algorithm is improved, the position update formula is designed, and the geese position update is used instead of particle resampling to update the position of the unmanned vehicle. Through MATLAB simulation experiments and experiments with real datasets, the improved algorithm improves the accuracy of positioning and mapping compared with the standard PSO-FASTSLAM, and verifies the superiority of the algorithm.