Comparison of nonlinear ltering techniques for inertial sensors error identi cation in INS/GPS integration

Nonlinear ltering techniques are used to fuse the Global Positioning System (GPS) with Inertial Navigation System (INS) to provide a robust and reliable navigation system with a performance superior to that of either INS or GPS alone. Prominent nonlinear estimators in this eld are Kalman Filters (KF) and Particle Filters (PF). The main objective of this research is the comparative study of the well-established ltering methods of EKF, UKF, and PF based on EKF and UKF in an INS-GPS integrated navigation system. Dierent features of INS-GPS integrated navigation methods in the state estimation, bias estimation, and bias/scale factor estimation are investigated using these four ltering algorithms. Both ground-vehicle experimental test and ight simulation test have been utilized to evaluate the lters performance.