Ground Based Augmentation System: Positioning And Integrity Monitoring

It is of highly necessity to enhance the integrity monitoring performance of both the civilaviation user and local ground facility (LGF), in order to enable the user satisfy the requirednavigation performance (RNP) in the phase of vertical guidance (APV) or even higherperformance required based phases. In other words, the integrity performance of ground basedaugmentation system (GBAS) should be enhanced in terms of accuracy, integrity, availabilityand continuity. Since essence of integrity monitoring is to provide a measure of confidencethat come along with all possible potential fault sources, and GBAS is based on thedifferential positioning technology, the thesis is mainly about how to mitigate and monitor themainly existing error sources originate from GBAS positioning process.Since the navigation environment difference between the civil aviation user and the LGF,furthermore, even the same kind of error will impose different effect to the user and the LGF,it is therefore necessary to design the integrity methods in terms of the user and the LGF,respectively. As GBAS is try to enable the user within reach to satisfy the RNP of differentcivil aviation phases, the thesis managed to give an insight into the positioning technologyand integrity monitoring algorithms. The outline of the thesis is as follows:Firstly, with the purpose of suppressing the smoothed error that comes from carriersmoothing process in the single-point positioning mode, an Kalman filter is combined withthe carrier-smoothed-code algorithm, i.e., the adaptive Kalman filter based on the maximumlikelihood criterion was proposed to compensate the deficiency of traditionalcarrier-smoothed-code algorithm. The innovation sequences were utilized to estimate andadjust the variance of system noise and measurement noise in real-time, respectively, and thenan optimal smoothing-time constant in the dynamic positioning were derived from theadaptive algorithm.Second, through analyzing the time-space effect of ionosphere delay and multipath onthe differential positioning based carrier-smoothed-code, a novel adaptivecarrier-smoothed-code algorithm is proposed to minimize the non-common errors containedin the differential pseudorange corrections broadcasted from the LGF. Through the errorvariance analysis of the differential pseudorange corrections, the optimal smoothing timeconstants for the user and the LGF are obtained from the extreme value theory (EVT) tomitigate the non-common errors, which mainly induced by ionosphere delay and multipath, with the purpose of improving the accuracy of the differential pseudorange corrections. Theflight test verified the effect of the proposed algorithm on minimizing the non-common errors.Third, the Doppler measurement based adaptive clock-bias prediction model withsmoothing-window and solution separation were combined to aid the traditional receiverautonomous integrity monitoring (RAIM). With the adjustment of smoothing time window inthe adaptive clock-bias estimation model, both the positioning accuracy and the availabilitylevel of the RAIM in the horizontal direction can be enhanced as more redundancy can beobtained from the Doppler measurements. In order to satisfy the stricter RNP in the verticaldirection, the map-matching aided tightly coupled GPS/DR/DEM integrated navigationsystem is applied. A variable false alarm rate (as opposed to a constant false alarm rate in thetraditional RAIM) is considered to improve the fault detection performance in selecting thecorrect link, especially near junctions. The vehicle on-board experiment shows that the systemavailability level derived from autonomous integrity monitoring extrapolation (AIME) couldsatisfy the APV I based RNP.Forth, since the traditional identification threshold based satellite fault detection andidentification algorithm led to missed detection and false alarm, which reduced the correctidentification rate, a new RAIM algorithm was proposed for identifying double-satellitessimultaneously. The geometry relationship between the parity vector and fault feature planewas used to identify the faulty satellites. The proposed algorithm was therefore immune to theproblem caused by identification threshold and improved the correct identification rate.Fifth, as the non-perfect distribution of pseudorange correction error, i.e., the non-zeromean Gaussian error and non-Gaussian error in the correction error, the traditional sigmainflation was difficult to satisfy the stricter availability level requirement (e.g., CAT II/III).With satisfying the error-tail overbound requirement, the mean inflation and non-Gaussianinflation were combined to improve the traditional sigma inflation. By selecting the optimalinflation vector, which include Gaussian the proposed method was not only able to overboundthe predefined non-ideal error but also enhance effectively the system availability. Thesimulation result reveals that the inflation vector based sigma inflation could improve theavailability to at least95%even under the case of critical satellite failure.Sixth, since the traditional maximum-likelihood estimation criterion based B-value,which generated from multiple reference consistency check (MRCC) in range-domain haslimitations in satisfying the integrity requirement of CAT II/III for civil aviation, a Kalmanfilter-based position-domain test statistics has been developed for improving the faultdetection and isolation of multiple reference receivers. Based on the integrity risk model of multiple receivers’ failure hypothesis, the adaptive fading factor was applied in the Kalmanfilter process to keep a balance between filtering stability and integrity risk. The outdoorexperiment demonstrates that, even if different kinds of fault types are considered under thesingle-fault case, the adaptive Kalman filter could improve the fault detection and isolationperformance with an acceptable availability loss.