Performance evaluation of different optimal smoothers for land-vehicle navigation using integrated GPS/INS systems

摘要: Integrated GPS/INS Systems are recognized as the ideal tools in the application of land-vehicle navigation. The complementary characteristics of the integration overcome the shortcomings of each system. Kalman Filter (KF) is an applicable optimal estimation method to provide real-time navigation solutions. However, in the GPS/INS integration, the accuracy of the KF navigation solutions degrades rapidly with time during GPS measurement gaps. Therefore, optimal smoothing methods are required to accommodate for this problem. In this paper, two fixed-interval smoothers, namely Two-Filter Smoother (TFS) and the Rauch-Tung-Striebel Smoother (RTSS) will be discussed and utilized. The details on the revised TFS algorithm will be investigated. Two-land vehicle field tests are used to evaluate the KF performance and the smoothing efficiency. The position errors accumulated during GPS outages are expected to be substantially improved using any of the two implemented smoothers. In general, the accuracy enhancement level was above 90% in the obtained results for the two utilized land-vehicle tests.