Research on an effective integrated navigation failure-tolerance filtering technology

For an INS/multi-antenna GNSS/MHS/DNS integrated navigation system, the navigation information errors of one or more sensors may cause the positioning accuracy of the entire integrated navigation system in serious decline, and even could not make the system work as usual. In order to solve this problem, a multi-mode federated Gaussian sum particle fault-tolerant filtering technology is proposed. It adopts a no reset federated filter framework, and in the locally filtering, a Gaussian sum particle filter algorithm is utilized to improve the precision of filtering for nonlinear and non Gaussian system, meanwhile, a particle likelihood function failure detection algorithm is used to detect and isolate failure sensors. The simulation studies presented here show that the proposed multi-mode federated Gaussian sum particle fault-tolerant filtering technology is feasible, and the designed multi-mode integrated navigation filter can achieve sufficient reliability and accuracy.