Proposing a Robust Adaptive Unscented Particle Filtering (RAUPF) Algorithm
Li Xue,Shesheng Gao,Jianchao Wang
DOI: https://doi.org/10.3969/j.issn.1000-2758.2011.03.027
2011-01-01
Xibei Gongye Daxue Xuebao/Journal of Northwestern Polytechnical University
Abstract:Particle filtering causes degeneration and has difficulties in selecting the importance density function. Aiming at this problem, we propose a RAUPF algorithm, which is better than the existing particle filtering algorithms, through adding the concept of robust adaptive filtering to the unscented particle filtering (UPF). Sections 1 and 2 of the full paper explain our proposal. The core of section 2 consists of: (1) in selecting the importance density function, we combine the robust estimation with the UPF algorithm by utilizing the equivalent weight function and the adaptive factor as shown by eq. (19); (2) we reasonably distribute the effective model information, thus suppressing the influence of abnormal interference on the state model and the observation model. With two numerical examples, section 3 verifies our RAUPF algorithm and applies it to a SINS/SAR integrated navigation system. The simulation results, given in Figs. 1 through 5, and their analysis show preliminarily that: (1) our RAUPF algorithm outperforms the conventional extended Kalman filtering and the particle filtering in terms of filtering precision, thus being more suitable for the filtering calculation of a nonlinear and non-Gaussian model; (2) it can improve the location precision of the SINS/SAR integrated navigation system.