A nonlinear suboptimal fusion algorithm weighted by matrices based on LMI and BP networks
In order to solve the fusion estimation problem for multi-sensor nonlinear systems with unknown cross-covariance,an improved suboptimal fusion algorithm weighted by matrices is proposed.Under the sense of linear minimum variance,the simplest constraints based on fusion weighted by matrices are derived by Shure complement theorem.These constraints can ensure the positive definiteness of the fusion estimation error variance,and the consistency of the proposed suboptimal fusion estimation.Further,a suboptimal fusion estimation weighted by matrices is proposed based on linear matrix inequality(LMI).Considering the time-consuming problem in the optimization process of LMI algorithm,the optimal value is obtained by the BP networks.A nonlinear suboptimal fusion algorithm weighted by matrices based on LMI and BP networks is proposed in combination with the Cubature Kalman filter algorithm(CKF).Simulation analyses verify the effectiveness of the algorithm applied to nonlinear systems.
NETL
NSTL
万方数据
