Artificial Neural Network Filter Algorithm Improvement for Tightly-coupled AHRS/GPS
Lijun Gu · Journal of Projectiles.Rockets.Missiles and Guidance · 2009
The parameter model for land vehicular tightly-coupled AHRS/GPS was built based on fusing pseudo range pseudo-pseudo range rate-heading angle. The radial basis function (RBF) NN aided EKF was designed to approach the nonlinearity of the integrated system,realizing the adaptive estimation for navigation parameters. The simulation indicates the nonlinear model can be approached with high accuracy and rapidity,the RMS of attitude error is decreased by 5.9%~22.8% compared with standard EKF. The accuracy and dynamic performance are all improved as expected under limited sensor accuracy.