Tightly-coupled INS/GPS integration with magnetic aid

This paper presents a novel tightly INS/GPS integration algorithm for altitude estimation. The altitude estimation is independent to provide accurate Euler angles for INS system, and magnetic sensors reduce the growth-rate of navigation-errors. In this paper, a tightly INS/GPS model is built with 23 state variables; bias and scale factor of IMU are estimated to correct INS system; and PVT information of GPS is used as measurement model in the Kalman filter. An experimental system is built on low-cost components and tested in outdoor environment to evaluate the overall accuracy of estimated trajectory. Through experimental results, we have shown that the tightly-coupled algorithm is higher accuracy and more robustness than loosely-coupled algorithm when GPS signal is corrupted or when three or two satellite signals are received.