Application of AUV Navigation Based on Deterministic Particle Filter Algorithm
Xiaoyan Fei, Yue Shen, Tianhong Yan · 2018
Navigation algorithm is the key of AUV positioning. Particle Filtering (PF) is a Bayesian estimation based on Monte Carlo method, which is suitable for the analysis of nonlinear and non-Gaussian systems. It is widely used in the research of target tracking, navigation positioning. However, the biggest drawback of particle filters is the degradation of particles. One of the ways to mitigate the effects of this phenomenon is to resample the particles. This paper describes the basic principle of particle filter and introduces polynomial and deterministic resampling algorithms in detail. By comparing the simulation results of the Extended Kalman Filtering (EKF) algorithm, the Particle Filtering (PF) algorithm and the Deterministic PF(DPF) algorithm in AUV navigation and positioning, the feasibility and accuracy of the DPF algorithm are confirmed.