Content area
Full text
Introduction
A greatly improved capability for the Global Navigation Satellite System (GNSS) is expected to be available in the near future. It will include multiple-frequency civil signals and multiple constellations, such as Galileo, a renewed GLONASS, and Compass. The improved accuracy and reduced threshold for failure detection will bring a new definition of failure. As a result, the prior probability of a failure can be determined to be greater than what is used now. Furthermore, an increased number of ranging sources due to the new GNSS constellation necessitates considering the possibility of multiple simultaneous failures ([12] Pervan et al. , 1998; [9] Macabiau et al. , 2005; [1] Angus, 2006; [2] Blanch et al. , 2007; [3] Ene et al. , 2007).
The conventional weighted least squares (WLS) failure detection and exclusion (FDE) algorithm uses a range residual vector to determine whether the current range measurement is normal or not ([13] Walter and Enge, 1995). However, because the range residual is a projection of the range error, it contains only partial information regarding the error. Therefore, the test statistic cannot be used as an absolute indication of satellite failure, and this algorithm cannot be applied in the most general cases. It can only be applied in the case of a single satellite failure. This paper proposes a new receiver autonomous integrity monitoring (RAIM) algorithm, which can handle simultaneous multiple failures as well as a single failure. The proposed algorithm uses range residuals and satellite observation matrices of several consecutive epochs for FDE. This concept of FDE was originally proposed by [10] Martini et al. (2006). It provides a test statistic for comparison with a threshold analogous to the conventional WLS algorithm, but is able to detect all kinds of failures. However, because the magnitude of the minimum detectable bias (MDB) is on the order of 5 km with a detection latency of 2-5 s, it is hard to implement in safety of life (SOL) applications. In order to achieve MDBs acceptable for practical applications, the proposed algorithm adopts the relative receiver autonomous integrity monitoring (RRAIM) scheme. In addition, the proposed algorithm combines the backward time difference residual vectors and observation matrices, and it directly estimates the error vector in order to detect failure(s) in...