Research on multi-sensor data fusion algorithm for unmanned vehicles under extreme conditions

Abstract

Aiming at the problem of insufficient accuracy in Simultaneous Localization and Mapping of vehicle robot using a single sensor in extreme environment scenes, according to the characteristics of the sensors, a method of fusing the data of lidar and inertial sensors is proposed, the vehicle robot system is designed, and the positioning principle of lidar and inertial sensor is elaborated. Two data fusion algorithms, weighted fusion and Kalman filter, are mainly studied, and the experiments prove that the Kalman filter algorithm has higher positioning accuracy.