An Improved Particle Filtering Strategy for Terrain Aided Navigation Based on MBES Information

Abstract

Terrain aided navigation is widely used in underwater positioning. The problem of positioning error accumulation is difficult to be completely solved by strapdown inertial navigation system (SINS) / doppler velocity log (DVL) integrated navigation system. Single beam sounder and DVL provide less depth information and have lower real-time navigation filtering accuracy. Therefore, a filtering strategy for terrain aided navigation based on multi-beam sounding information is proposed. A tightly integrated framework of SINS, DVL and multibeam echosounder (MBES) is designed, which fully considers the sensor installation error, reduces the number of subsystems, and improves the system stability. An improved particle filter algorithm is proposed, the index of particle dispersal range and particle diversity loss rate are proposed, and a secondary resampling method is designed to solve the problem of particle degradation and particle diversity loss. The simulation experiment based on the local topographic map of Qiandao Lake proves that the proposed algorithm has higher positioning accuracy, better convergence of positioning errors during a long time voyage, and better positioning effect in flat areas.