Research On Anti-jamming Filtering Algorithm For Systems With Unknown Input

The state estimation problem of the system widely exists in the domains of automatic pilot,data transmission,fault diagnosis,etc.The system state equation and measurement equation usually contain unknown inputs due to the environmental impacts,improper selection of model parameters,measurement equipment failures and other reasons.The existence of unknown inputs makes the study of above state estimation problem extraordinary complicated.Therefore,for the linear system with unknown inputs and the nonlinear system with unknown inputs,the estimation problem of system state and unknown input is studied in this thesis.Firstly,a novel extended recursive three-step filter is proposed to estimate the state and unknown input of the linear system simultaneously in case that unknown input direct feed through to the measurement equation.Then,a novel recursive three-step filter is designed for the linear system when unknown input only affects the system equation.Finally,considering the nonlinear system with unknown inputs,an extended cubature Kalman filter is presented.The main contributions of this thesis are as follows:(1)Filter is designed for linear system when the coefficient matrix of unknown input in measurement equation is not of full rank.Without prior knowledge of unknown input,the classical recursive three-step filter cannot be applied when the unknown input coefficient matrix in measurement equation is not of full column rank.According to the linear minimum variance unbiased estimation criterion,a novel extended recursive threestep filter is proposed,and the specific recursive steps of this filter are summarized.The simulation experiments show that the novel filtering method can effectively reduce the state estimation error compared with other unbiased minimum variance state estimation method.(2)Filter is designed for linear system when the coefficient matrix of unknown input in state equation is not of full rank.For the linear system with unknown inputs,if the unknown input coefficient matrix in state equation is not of full column rank,the classical recursive three-step filter cannot be applied in state estimation.In order to solve this problem,we design a novel recursive three-step filter based on the linear minimum variance unbiased estimation criterion.And the recursive steps of this novel filter are gathered too.Simulation results show that this novel filter can estimate the system state and unknown input effectively.(3)Extended cubature Kalman filter is designed for nonlinear system with unknown inputs.For a nonlinear system affected by unknown inputs,the problem of estimating unknown input and nonlinear system state simultaneously is considered.An extended cubature Kalman filter is proposed to estimate the state and unknown input in case that unknown input can be any signals which affects both the nonlinear system equation and measurement equation.The effectiveness of the proposed filter is verified by simulation results.