Bio-Inspired Geomagnetic Navigation Algorithm Based on Segmented Search for AUV

摘要

Classical geomagnetic navigation mainly adopts geomagnetic matching technology with a priori magnetic map. In this paper, we research the navigation problem of the autonomous underwater vehicle (AUV) without a priori magnetic map. Meanwhile, our research found that the parameters of geomagnetic field can converge in one direction at the same time in some local areas and this direction will not change until the consistency of parameters convergence is spoiled. Given this, our paper proposes an optimized navigation algorithm based on the segmented search strategy. Firstly, probability evolution strategy (PES) is used to search for navigation. When multiple geomagnetic parameters converge at the same time, the current heading motion is maintained in a straight line, thus shortening the navigation time. However, when multiple geomagnetic parameters can not converge at the same time, we use search algorithm to complete the navigation task finally. This algorithm could make full use of the magnetotaxis in the stage of searching. This improved algorithm can change the navigation strategy in real-time by combing distribution characteristics of geomagnetic field and the search method in the subsequent process. The simulation results verify the feasibility and effectiveness of our optimized algorithm, which has indicated potential application in the future.