Design of AGV combined positioning system based on characteristic error model

This paper puts forward an AGV combined positioning method based on laser and dead reckoning. First, the kinematics model of differentially driven AGV is established in this paper, and the system equation of state is obtained. Then, the data obtained by the lidar is used as the observation of the system, and the prior state is corrected by the Extended Kalman Filter (EKF) to obtain the pose estimation of the system. Considering that the accuracy of EKF is closely related to the accuracy of the noise matrix, this paper implements modeling of process noise and observation noise based on environmental characteristics. Finally, EKF is used for data fusion to achieve high-precision AGV pose estimation. The experimental result shows that the AGV positioning combining lidar data and dead reckoning data using EKF can effectively improve the accuracy of AGV positioning.