Fusion of GPS, INS and odometric data for automotive navigation

Nowadays, accurate and reliable positioning systems are required in several applications such as safety of life or liability critical. The global positioning system (GPS) is the most common navigation system for almost any application involving localization, navigation or tracking. In particular, GPS is used intensively for automotive navigation. However, the quality of the localization task for ground vehicles is subjected to degrade in outdoor conditions. For instance, severe degradations have been observed in urban areas, when the vehicle is subjected to multipath or masking effects interfering with the satellite lines of sight. In this case, combining GPS pseudo ranges with other measurement sources is an interesting way of improving localization performance. This paper studies different fusion approaches based on differential odometry provided by wheel speed sensors (WSS) and inertial sensors such as accelerometers and gyros. Different tracking strategies based on extended and unscented Kalman filters are investigated and compared for these fusion approaches.