An Improved Adaptive Unscented Kalman Filter With Application in the Deeply Integrated BDS/INS Navigation System

The integration of the BeiDou Navigation Satellite System(BDS) and the Inertial Navigation System(INS) can provide a more reliable and accurate navigation service than either system alone. The deeply coupled architecture for BDS and INS integration has more superior performance than the loosely coupled or the tightly coupled. Owing to the complicated dynamic scenario and the nonlinear system's noise uncertainty, the adaptive Kalman filter(AKF) algorithm is often adopted in the deep integration(DI) system. The adaptive Sage window methods including innovation-based adaptive estimation(IAE) and residual-based adaptive estimation(RAE) are widely applied in AKF algorithms, but they have several limitations. We propose an improved adaptive unscented Kalman filter(AUKF) based on forgetting-factor-weight smoothing and multi-factor adaptation to overcome these limitations. Compared with the Sage window methods, the improved AUKF algorithm is immune to the quantity change of the satellites concerning integration and more sensitive to present dynamic. Furthermore, it can reduce the computation and storage burden in implementation. A simulation test based on a software platform and the deeply integrated BDS/INS navigation system is carried out to evaluate the performance of the improved AUKF. Simulation results show that the improved AUKF algorithm outperforms the extended Kalman filter(EKF) and has a similar performance with the RAE-AUKF.

Abstract licence: CC BY-NC-ND

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.