Nonlinear filtering techniques for geomagnetic navigation

Abstract

This article focuses on the application of different nonlinear filtering techniques for geomagnetic navigation, including extended Kalman filter, unscented Kalman filter, particle filter and extended Kalman particle filter. The research evaluates the four methods for navigation of missile during its cruise phase. The measurement equations are obtained by using a surface spline method with the real regional geomagnetic data. Simulation results show all the filters have good performance in areas with abundant geomagnetic information. Among the four filters, unscented Kalman filter tops in the convergence rate and precision with a fairly low tuning sensitivity to the flight path. As the performance of unscented Kalman filter effect is influenced by unscented conversion parameters, a method of parameters optimization using genetic algorithm is presented. The method’s feasibility is further demonstrated by robustness analysis of optimal parameters.