Graph-Optimization Online Calibration Method of Vision-Inertia-Wheel Odometer

Online estimation and optimization of external parameters are crucial to the fusion performance of the camera-IMU-wheel odometer integrated navigation and positioning system. The low estimation of traditional online calibration methods limits the positioning performance of multi-sensor fusion. Therefore, this paper proposes a graph-optimization online calibration method of vi-sion-inertia-wheel odometer based on preintegration theory. First, a comprehensive IMU-odometry preintegration model is formulated, incorporating the odometer scale factor as well as the relative pose deviations between the IMU and the odometry sensor. Then, based on the graphical model optimiza-tion theory, the preintegration model is optimized together with the visual factors, and the external parameters between the cameras, IMU and odometry are estimated. Finally, the effectiveness of this method is verified through the KITTI dataset. The results show that the calibration accuracy of our method is higher than other methods. At the same time, the estimated external parameter results are brought into the navigation system to obtain the trajectory closest to the true value.