Sequential fusion state estimation algorithm for multi-sensor nonlinear systems with cross-correlated noise

Abstract

This paper is concerned the problem of multi-sensor state estimation with cross-correlated noise, this paper adopts sequential fusion to estimate the state. The statistical characteristics of measurement noise of different sensors in the multi-sensor system is related, and also related to the system noise in one step. Firstly, based on the estimation of observation noise, a global optimal fusion filter based on sequential fusion and Cubature Kalman filter is proposed for the first time. Secondly, the algorithm proposed in this paper is simulated by numerical method. The method of numerical realization is Cubature Kalman filter based on deterministic sampling. Finally, the effectiveness of the proposed algorithm is demonstrated by a simulation example.