Key points are not available for this paper at this time.
Inertial navigation system (INS) has an increasingly important role in indoor navigation, which mainly uses inertial measurement units based on a micro electro mechanical system (MEMS) to acquire data, and which is independent of the external environment. However, INS has serious accumulated errors, and thus, it was often integrated with wireless location systems (WLSs), such as ultra wideband (UWB) system, in order to enhance the position performance. Namely, MEMS-based inertial sensors have the problem of random errors. Besides, a UWB system is vulnerable to external environment conditions, such as the non-line-of-sight (NLOS) factor and multipath effects, and thus, many outliers are produced. In order to improve the overall performance of the INS/UWB system, this paper proposes the three-tier approach based on: 1) analysis and pre-filtering of random errors of MEMS-based inertial sensors, and use of a complementary filter to provide attitude information of navigation system; 2) use of the anti-magnetic ring (AMR) to eliminate the outliers from the UWB system in NLOS environment; and 3) improvement of positioning accuracy at information fusion level using the double-state adaptive Kalman filter. The proposed approach was verified by experiments that included AMR test and filter test. The obtained results have validated the proposed method efficiency.
Fan et al. (Thu,) studied this question.