Application of RBF neural network aided kalman filtering to GPS/SINS integrated navigation
Xiyuan Chen · Infrared and Laser Engineering · 2008
In order to overcome the dependency that the measured noise must be white noises with zero mean value in the routine Kalman filtering equations, The real environment of a moving vehicle′s integrated navigation system (aircraft or ship) is worse, a method of using the neural network′s capability of self-taught, self-organized and self-adapted to aid Kalman filter is proposed. The convergence speeds of back-propagation (BP) neural network (NN) and radial basis function(RBF) NN based on orthogonal least squares (OLS) are contrasted. The RBFNN is chosen since its faster convergence speed ,the simulation results indicate that the algorithm is effective in restraining divergence and improving navigation precision.