Processing Technique and Performance of the Observation of Evil Waveform in the Chip Domain
Résumé
This paper focuses on the Chip Domain Observable, a particular observation of the GNSS signal that is done in the time-domain, directly on the digitized samples at the output of the RF front-end of a GNSS receiver. With such observable, it is possible to observe in details the average chip transitions of the received GNSS signal, and to monitor some anomalies affecting the signal. This observable is obtained through a particular signal processing technique, which is presented in details in the paper. Then, a particular type of signal deformation, called the Evil Waveforms and modelled by the ICAO, is applied to this observable. The paper gives the analytical expression of each ICAO Threat Model observed by the Chip Domain Observable, as well as the measurement noise affecting the observation, in function of the receiving conditions and some signal processing parameters. This work may serve in the future to investigate some novel signal quality monitors, working in the time domain, rather than in the correlation domain. INTRODUCTION The basic principle of GNSS signal processing assumes that the received signal is undistorted. For example, a code delay lock loop aims at equaling an early and a late correlator in order to find the location of the prompt correlator, which should correspond to the maximum of the correlation function of the GNSS signal, which is then turned into a pseudo-range measurement. This assumes that the correlation peak has a triangular shape. However, in reality, the received signal is affected by distortions, for example due to some incorrect behavior of the signal generation payload onboard GNSS satellite (also called Evil Waveform) or due to the local environment of the receiving antenna which may create multipath. Both effects are translated into deformation of the correlation function, which in turn creates some error on the pseudo-range measurements and the subsequent position. In order to detect or mitigate errors coming from signal distortion, a family of signal processing techniques has been derived, usually based on the monitoring of the correlation function of the received signal thanks to multiple correlators. The correlation function is therefore observed at different delays, and combinations of different correlators permit to detect anomalies on the correlation function such as asymmetry or slope change. Such multi-correlator techniques are used in high integrity systems, for example Ground-Based Augmentation System, in order to detect an Evil Waveform and exclude a potentially failing ranging source [7]. Another application is the detection of multipath for receivers roaming in urban environment, in order to either exclude some measurements and improve the overall quality of the position solution, or to estimate the multipath parameters and mitigate their impact [5][1]. More recently, signal deformation monitoring has turned to the observation of the signal in the time domain, rather than in the correlation domain. Indeed, by observing the signal in the time domain, thanks to the particular properties of GNSS signals (recalled in [3]) and adequate signal processing, it is possible to observe the chip transitions with high fidelity. This observation is believed to be used by Novatel under the name of Vision Correlator. The aim of this paper, and resulting structure, is to detail the advantages of the Chip Domain Observable (CDO), to explain what signal processing steps are required for CDO computation, to illustrate the use of this technique on the Evil Waveform monitoring for GPS L1 C/A, and finally to detail the accuracy of such observation, in terms of standard deviation of the observed CDO values.
Origine : Fichiers produits par l'(les) auteur(s)