Kalman filtering parameter optimization techniques based on genetic algorithm

Kalman filter is widely used to restrain noise existing in flight control system of UAV due to its many merits. However, the effect is very sensitive to Kalman filter parameters, whose choice depends on operator’s experience extremely. A GA-based filter parameters optimization approach is presented. In this approach, GA is employed to find out the optimal Kalman filter parameters by way of minimizing objective function which includes such terms as variance of model uncertainty, variance of measurement noise, covariance of estimate error in initial states. The simulation results show that the approach can improve accuracy and stability of Kalman filter.