GPS/INS integration utilizing dynamic neural networks for vehicular navigation

Abstract

Recently, methods based on Artificial Intelligence (AI) have been suggested to provide reliable positioning information for different land vehicle navigation applications integrating the Global Positioning System (GPS) with the Inertial Navigation System (INS). All existing AI-based methods are based on relating the INS error to the corresponding INS output at certain time instants and do not consider the dependence of the error on the past values of INS. This study, therefore, suggests the use of Input-Delayed Neural Networks (IDNN) to model both the INS position and velocity errors based on current and some past samples of INS position and velocity, respectively. This results in a more reliable positioning solution during long GPS outages. The proposed method is evaluated using road test data of different trajectories while both navigational and tactical grade INS are mounted inside land vehicles and integrated with GPS receivers. The performance of the IDNN – based model is also compared to both conventional (based mainly on Kalman filtering) and recently published AI – based techniques. The results showed significant improvement in positioning accuracy especially for cases of tactical grade INS and long GPS outages.