Non-linear, non-Gaussian estimation for INS/GPS integration

Mobile mapping system (MMS) has developed rapidly over the past two decades. In that system, a direct geo-referencing system using Global Positioning System (GPS) and Inertial Navigation System (INS) is applied to determine the time-variable position and orientation parameters. The Kalman filter or Extended Kalman Filter (EKF) is popularly used as the optimal estimation tool for real-time INS/GPS integrated kinematic position and orientation determination. In such those estimation strategies, linearization and assuming Gaussian distribution are utilized. However, the fact that the state model and measurement model in INS/GPS integration are originally non-linear and the noise arising during operation may be non-Gaussian distribution. This characteristic may lead to the degraded performance of the system utilizing KF or EKF in case of non-linear models and non-Gaussian noises. This paper investigates some of non-linear, non-Gaussian estimation strategies in order to improve the performance of system. Firstly, the fundamental of algorithms will be introduced, those estimation strategies is then applied on simulation scenario to find the optimal method, finally, real data from field test will be implemented for the real aspects and final conclusions.