A New Neural Network-based Algorithm for GPS/INS Integration

Integrated navigation systems have been becoming more and more important in many applications. Particularly, Global Positioning System (GPS) and Inertial Navigation System (INS) integration is being given much attention in the past few years, as it is widely used in several positioning and navigation fields. Both of such two systems have their unique features and shortcomings. Their integration offers systems that overcome each of their drawbacks and maximize each of their benefits. As we know, Kalman filtering theory is one of data fusion methods for incorporating of INS with GPS. But, it has some drawbacks in terms of stability, computation load, immunity to noise effects, observability, and so on . Specially, Kalman filters perform adequately only under certain predefined dynamic models. Neuron computing, one of the technologies of Artificial Neural Network (ANN), is a useful tool for solving nonlinear problems that involve mapping input data to output data without having any prior knowledge about the mathematical process involved. It is also suitable to be used for integrating GPS and INS. In this paper, a new adaptive Kalman filtering based on Back Propagation Neural Network (BPNN) is studied, whi ch is used for integration between GPS and INS data. BPNN algorithm can be employed to aid the adaptive Kalman filter to reduce the estimation error due to, among other imperfections, highly manoeuvring, model varying effect. When GPS information exists, the integrated model is established and its internal structure is tuned to mimic the present vehicle dynamics. During periods of GPS signal blockage, the studied algorithm works in the prediction mode to estimate the position changes based on the INS velocity and azimuth information. Finally, to assess the performance of the proposed algorithm, the performance of the proposed algorithm is compared against the traditional Kalman filter model. The correlative simulation results show that the performance of the BPNN-based adaptive Kalman filter is better than the normal adaptive Kalman filter for such integrated navigation system.