A neuro-wavelet method for multi-sensor system integration for vehicular navigation

The last two decades have shown an increasing trend in the use of navigation technologies in several applications including land vehicles and automated car navigation. Navigation systems incorporate the global positioning system (GPS) and the inertial navigation system (INS). While GPS provides position information when there is direct line of sight to four or more satellites, INS utilizes the local measurements of angular velocity and linear acceleration to determine both the vehicle's position and attitude. Both systems are integrated together to provide reliable navigation solutions by overcoming each of their respective shortcomings. The present integration schemes, which are predominantly based on Kalman filtering, have several inadequacies related to sensor error models, immunity to noise and observability. This paper aims at introducing a multi-sensor system integration approach for fusing data from an INS and GPS hardware utilizing wavelet multi-resolution analysis (WMRA) and artificial neural networks (ANN). The WMRA is used to compare the INS and GPS position outputs at different resolution levels. The ANN module is then trained to predict the INS position errors in real time and provide accurate positioning of the moving vehicle. The field-test results have demonstrated that substantial improvements in INS/GPS positioning accuracy could be obtained by applying the proposed neuro-wavelet technique.