PERFORMANCE COMPARISON OF A FIXED-INTERVAL SMOOTHER AND A KALMAN FILTER FOR GNSS/INS INTEGRATION

摘要

This paper compares several approaches of GNSS/INS integration. Tightly coupled and loosely coupled approaches are discussed as well as an approach that tightly integrates differential GNSS and inertial navigation with carrier phase ambiguity resolution. Hence, centimetre accuracy can be achieved with a tightly coupled system. The approach can be combined with a smoother which improves the overall system performance by means of post-processing. Measurement results show that the presented smoothing algorithm can overcome problems of poor satellite availability. It is possible to recover the trajectory during GNSS outages even together with a low-cost MEMS inertial navigation system.