Investigation of Factors Influencing the Accuracy of Multi-Sensor Fusion with Extended Kalman Filter

An extended Kalman filter-convex combination fusion algorithm with time synchronization and a variable sampling rate extended Kalman filter fusion algorithm have been designed. Based on these algorithms, simulation analysis has been conducted. The impact of time synchronization on accuracy, the influence of constant velocity motion model and the “current” statistical model on system performance, performance comparison of the two systems under the “current” statistical model, the effect of low-quality inputs on system performance, and the influence of sensor spatial relationships on system performance have been investigated. This work provides insights into system construction and strategies for improving system tracking accuracy.