Rao-Blackwellized Particle Filter Application in Deeply Integrated SINS/GPS Navigation System
Jian Xiong · Journal of Astronautics · 2009
As for data fusion in the integrated navigation system,there exists mixture linear and nonlinear system with high dimensional model.The traditional linear Kalman Filter(KF)can not treat directly and single particle filter(PF)have the drawback of computational intensive.The paper implements Rao-Blackwellized particle filter(RBPF)which divides state vector into some parts.The linear part is estimated using KF and the nonlinear part is estimated using PF.using RB technology,the filtering system dimensions decrease from 20 to 3,which makes the essential particle number decrease quickly.With the same particle number,the simulation demonstrates filtering precision increase rapidly,which insures the time demand in the real time navigation system.