A New Multiple Hypothesis RAIM Algorithm : Direct Estimation of a Fault Vector with an RRAIM Concept

Abstract

In the next years, it is expected that dramatically improved capability for Global Navigation Satellite System (GNSS) will be available which will include multiple-frequency civil signals and multiple-constellations (Galileo, a renewed GLONASS, and Compass). To provide robust LPV-200 service worldwide, new Receiver Autonomous Integrity Monitoring (RAIM) algorithms are being developed by various research groups (GEAS, 2008), (Todd Walter et al, 2010), (Livio Gratton et al, 2010). Most of these algorithms assume single satellite failure condition. However the definition of failures would be changed, because the accuracy will be improved, the threshold for failure detection will be reduced. As a result, the prior probability of failures could be larger than what we used now. Furthermore, increased number of ranging sources due to new GNSS constellation makes it necessary to consider the possibility of simultaneous multiple failures. This paper develops and analyzes a new RAIM algorithm as a candidate of future architecture of RAIM algorithm which can treat not only a single failure but also simultaneous multiple failures. A proposed algorithm uses measurements residuals and satellite observation matrices of several consecutive epochs for multiple Failures Detection and Exclusion (FDE). This new concept of FDE was firstly proposed by (Martini and Hein, 2006). It can detect multiple failures without limitation of number of faulty measurements. However the magnitude of Minimum Detectable Bias (MDB) is in the order of 5km with detection latency of 2 to 5 seconds, so that it is hard to be implemented in Safety Of Life (SOL) applications. This paper proposes a new FDE algorithm modified to have no detection latency by estimating the current error vector using the measurements of current and past epochs. In order to make the MDBs have smaller values so that it can be applied to practical application, we adopted Relative RAIM (RRAIM) scheme for navigation and protecting users against system failures. In this paper, we give detailed explanation of the FDE algorithms with rigorous mathematical expression. Simulation results show that proposed algorithm can detect and exclude the multiple failures of tens of meters depending on satellite geometry.