Autonomous Navigation based on Multi-sensor CIFIMM-SCKF

Abstract

A multi-sensor SCKF algorithm based on cubature information filter (CIF) is proposed for the problem of nonlinear state estimation and multi-sensor information fusion of the spacecraft. The multi-model filtering idea is adopted to realize the state filtering by embedding spherical radial volume rules in the extended information filter (EIF) framework. It not only preserves the excellent performance of the cubature Kalman filter algorithm, but also easily extends to multi-model navigation system for the state estimation. The simulation results show that the autonomous navigation method based on cubature information filtering multi-sensor SCKF can effectively avoid the problem of filter divergence due to the linearization error of the model and overcomes the unsteady filtering value of the UKF algorithm. The algorithm has higher accuracy and can more effectively solve the problem of state estimation in the case of strong multi-sensor nonlinearity.