Multi-sensor weighted combination fusion algorithm considering time registration
Longfei Wang, Xuerong Yang, Yajun Yang, Wenqi Guo · 2022 International Conference on Machine Learning and Intelligent Systems Engineering (MLISE) · 2022
This article addresses the problem of multi-sensor data fusion and time registration in the case of large data transmission delay. Kalman filter method has good filtering and prediction performance, so based on the time registration method of Kalman filter, a prediction-fusion-update-fusion target location algorithm combined with weighted combination fusion method is proposed. Firstly, the algorithm aligns the sampling time through the Kalman filter registration algorithm, and uses the weighted fusion to get the fusion estimation at the sampling time. Then, the fusion estimation is updated to the current time by using the Kalman filter. Finally, the real-time fusion estimation and the real-time prediction estimation of a single sensor are combined with weights to realize the high-precision real-time positioning of the target. Simulation results show that the algorithm can effectively reduce the positioning error of the target under the condition of large time delay, and has good robustness to improve the positioning accuracy of the target in the case of complex maneuvering.