A Huber based Unscented Kalman Filter Terrain Matching Algorithm for Underwater Autonomous Vehicle

Abstract

Underwater autonomous vehicles(AUV) are effective tools for marine surveys and military applications. As the energy of the electromagnetic signal will decay rapidly in seawater, most land navigation methods cannot be implemented on underwater vehicles. Underwater terrain matching navigation is an emerging high-precision passive navigation method. It compares the measured underwater terrain data with the digital map and obtains its own position. The unscented Kalman filter(UKF) based underwater terrain matching algorithm has the advantage of high position accuracy and good engineering feasibility. However, the UKF adopts Kalman filtering as the basic framework, which leads to a weaker robustness. When there is a deviation from the actual parameters, the estimation results will be greatly affected. Therefore, based on the above reasons, this paper studies the feasibility of UKF in the AUV underwater terrain matching problem and introduces the Huber method into the UKF, and obtains a framework of underwater terrain matching algorithm based on Huber method, which effectively improves the robustness of the algorithm and makes the algorithm more suitable for AUV different works.