Approach for Adaptive Filter of Systems with Random Changing Structures

An approach was given for adaptive filter of systems with random changing structures containing correlative or non-correlative characteristic parameters. An improved bootstrap filtering algorithm based on random sampling, which adapts quite well to the nonlinear/non-Gaussian problems, was adopted to avoid "basic structure distortion" caused by conventional filtering algorithm and to increase the estimation precision of the total filtering algorithm. Simulation results about missile guidance problems where measurement information is randomly interrupted in "twinkling" noise environment proved the validity of the approach.