**Abstract** : The Code-Carrier Divergence (CCD) fault is one of the ranging source faults that should be protected against at the high-integrity augmentation systems such as the Ground-Based Augmentation System (GBAS) and the Satellite-Based Augmentation System (SBAS). Several CCD monitors have been proposed according to support the GBAS Approach Service Types (GAST). The single-frequency CCD monitor is defined within the GAST C solution which supports Category (CAT) I precision approach [1]. The performance of the single-frequency CCD monitors was somewhat limited due to the effect of the ionospheric delay in the monitor metric. The GAST F, which supports CAT II/III, is defined for multi-frequency and multi-constellation including GPS L1/L5 and Galileo E1/E5. For this future iteration of GBAS, several dual-frequency CCD monitors have been proposed which can completely eliminate the ionospheric delay in the monitor metrics using the linear combination of the pseudorange and carrier phase measurements [2]. The single- and dual-frequency CCD monitors are based on the smoothed rate of the measurement using the two cascades of the 1st order filter. The use of two sequential filters is able to reduce the standard deviation of the monitor metric drastically even if the pseudorange measurements are involved. However, its smoothing performance could be considerably limited due to the multipath errors. The multipath error is time-correlated so that it reduces the number of independent samples from the point of view of the smoothing filter. Consequently, the standard deviation of the CCD monitor metric cannot be fully reduced so that the monitor may fail to detect some of faults that might be threatening the GBAS airborne users. In other words, the dual-frequency CCD monitor might not make the most of benefits of eliminating the ionospheric delay due to the reduced performance of the smoothing filter associated with the multipath error.
In the first part of the paper, to understand the impact of the multipath error on the dual-frequency CCD monitor, the equation for the standard deviation of the CCD monitor metric is derived as a function of a time constant of two cascades of the 1st order filter, a time constant of the multipath error and the standard deviation of the measurement noise by referring the approach proposed in [3]. Here, we assume that the multipath error can be modeled as the 1st order Gauss-Markov process. From the derived equation, a theoretical standard deviation of the CCD monitor not only at the steady state but also at the transient state can be obtained for given conditions of the smoothing filter and multipath error and measurement noise. The results obtained from the derived equation are validated by using simulated data. When we estimate the standard deviation of the CCD monitor using real data, it can be validated with the derived equation if the measurement noise and the time constant of the multipath error in measurement are known from the pre-processing of the real data. In this validation step, the standard deviation of the CCD monitor as well as the time constant of the multipath in the measurement can be verified inclusively. The result of this validation process is given in the paper using the currently available GPS L1/L5 and Galileo E1/E5 real data collected from the research-oriented GBAS ground stations located at Tenerife and Barcelona.
In the second part of this study undertaken, the threat space for the CCD faults has been defined using the requirement on the Probability of the Missed Detection (PMD) of the monitor which has been defined for the GAST D in the recent ICAO standards [4]. Referring to the requirement, the integrity risk does not occur when the absolute value of the differential range error is smaller than 0.75 m regardless of which values the PMD of the monitor is. This criterion can be used to limit the threat space of the CCD faults. Using the proposed threat space, the performance of the dual-frequency CCD monitors including divergence free and ionosphere-free monitors, which have been proposed by ENAC [2], is validated with respect to the requirement defined for the GAST D standards. For the GAST F, the corresponding requirement may be more relaxed thanks to the improved satellite geometry of the multi-constellation. However, it has not been defined yet, and for the conservative analysis, we use the same requirement for the GAST D. In the analysis, the various filter initialization times of the airborne carrier smoothing filter have been tested to take into account the possibility of having various differential range errors under the same CCD fault. In addition, the optimal monitor combination among four ENAC CCD monitors is suggested.