Multi-sensor Fusion SLAM Algorithm Considering Wheel Slip

With the continuous development of automated guided vehicle (AGV) technology, simultaneous localization and mapping (SLAM) has become a research hotspot. Currently, domestic AGVs mainly construct environmental maps based on laser radar scanning, while inertial measurement unit (IMU) sensors and wheel odometry are used to provide attitude estimation and motion compensation, speed and displacement information respectively. However, when the drive wheels of the AGV slip, the wheel odometry can produce significant errors, leading to positioning errors and map drift. This paper uses an error state Kalman filter (ESKF) to fuse wheel-IMU odometry and a slip detection method based on slip rate. When the AGV slips, the wheel odometry data is promptly discarded, and the IMU pre-integration is used directly to replace the odometry, effectively reducing the impact of slipping. Through comparative experiments, it can be found that the algorithm can reduce the error caused by slip by about 78%, which has certain stability and accuracy, and can further improve the robustness of the system when applied to AGV.