Multi-sensor Data Fusion of UAV Landing System

Reliable landing guide system plays an increasingly important role in modern Unmanned aerial vehicle (UAV). Since the performance of sensor customly varies with time and distance during the landing process, and the measurement signal error is non-stationary, it is difficult for the single-sensor guide system to maintain high positioning accuracy in all application scenarios. In this paper, a sliding window adaptive fusion algorithm based on iterative filter is proposed to fuse multi-source data from satellite, photoelectric, radar, and machine vision. The introduction of filtering iteration and sliding window can improve the robustness of fusion and solve the problems of multi-sensor information asynchrony and non-stationary observation error. The algorithm can also make adaptive adjustments for different stages of UAV landing. Experimental results show that the fusion algorithm can achieve precise navigation and provide a reliable basis for the accurate and safe landing of UAV.