A Voting-Based Robust Estimator Aided by INS Redundancy for Tightly Coupled GNSS/INS Integration in Urban Environment

The combination of Global Navigation Satellite System (GNSS) and Inertial Navigation System (INS) is widely applied for many ground transportation applications thanks to its accurate and continuous navigation service. However, with the advent of the multipath effect and non-line-of-sight (NLOS) signals caused by obstacles in urban environments, the quality of raw GNSS measurements is heavily degraded, leading to unexpected integration solutions. For this, an improved voting-based robust estimator (VBRE) algorithm assisted by the INS-derived measurement redundancy, is proposed for reasonable evaluation of satellite observation information, ensuring reliable integrated GNSS/INS solutions. The introduction of the generalized observation model allows the direct involvement of the short-term accurate INS-derived resolution in monitoring faulty GNSS measurements. Inspired by the voting theory, a residual-based collective decision framework is developed, where a group of INS-associated voters evaluates traversally each satellite observation candidate by the distance measure calculation and the agreement indicator mapping. Based on this, to determine the final GNSS measurement contribution, the detailed voter implementation is formulated with key steps including rejecting abnormal indicators, median voting, and multiple voting cycles. The above behaviors facilitate the rationality and fairness of each GNSS measurement weight allocation for integrated navigation. The proposed algorithm has been validated on two challenging open-source vehicular integration datasets in hybrid urban areas of Tokyo, Japan. The resultant 3D accuracy regarding position and velocity of Dataset I is around 60% higher than that of the Kalman filter and outperforms other existing methods. Similar performances are also achieved on more challenging Dataset II, with improvement reaching 80%.Meanwhile, this proposed algorithm presents remarkable flexibility against the various abrupt GNSS faults caused by different harsh urban areas.