Measurement Quality assessment in urban environments using correlation function distortion metrics
Résumé
Multipath decreases the accuracy of the GNSS positioning by distorting the correlation function. Based on this observation it is possible to detect abnormally large multipath error by monitoring the correlation function. Indeed for the tracking of the code delay, several correlator outputs that correspond to specific locations on the correlation function are observable. Based on the available observables that are combined to form a metric, one can design a test detection to quantize the distortion. The potential metric candidates are selected based on noise resilience considerations. Afterwards, using a simplified correlator output model, two methods are compared for the determination of the detection thresholds. The value of the thresholds is related to the expected Probability of False Alarm. An analytic way to assess the performance of the coherent tests in term of sensitivity is given. The efficiency of the test can be widely improved by smoothing the correlator outputs or the test variable itself. The discussed metrics are implemented on a realistic GNSS tracking simulator processing the Land Mobile Satellite Channel model developed by the DLR. The correlation between the multipath error and the signal to noise ratio (SNR) is well known. Performances of test based on metrics and SNR are finally compared
Origine : Fichiers produits par l'(les) auteur(s)