Reliability testing for multiple GNSS measurement outlier detection

Abstract

Due to a rapid development of several Global Navigation Satellite Systems (GNSS), multiple constellations are available to enhance navigation performance and safety. With the growing number of satellite constellations, the task of the GNSS navigation is to deal with the differences among systems but, on the other hand, more great levels of integrity and satellite visibility can be expected. GNSS navigation applications have difficulties in signal degraded scenarios where the GNSS solution can be degraded by errors such as multipath and signals being obscured. RAIM (Receiver Autonomous Integrity Monitoring) is a method necessary for assessing integrity performance levels mainly in safety-critical applications. Classical RAIM techniques are based on the assumption model of a single outlier in the measurements, but with a future of higher satellite availability and for navigation conducted in urban canyon scenarios, the single outlier assumption is unrealistic. Therefore, reliability monitoring techniques need to be modified to be suitable for use cases with high signal degradation levels. The FDE (Fault Detection and Exclusion) schemes analysed in this research for reliability monitoring are the “Observation Subset Testing” and a modified approach based on a w-test (called in this paper Multiple Faults De-weighting-MFD). In order to improve their performance a novel Channel Quality Index (CQI) parameter was used to describe the measurement confidence and quality. To validate the proposed approaches, tests have been performed using simulated data with GPS, Galileo and BeiDou signals in a multipath environment.