https://hal-enac.archives-ouvertes.fr/hal-01512447Blais, AntoineAntoineBlaisENAC - Ecole Nationale de l'Aviation CivileJulien, OlivierOlivierJulienENAC - Ecole Nationale de l'Aviation CivileAssessment of the effect of quantization on the degradation brought by interference on a GNSS receiverHAL CCSD2017[SPI.SIGNAL] Engineering Sciences [physics]/Signal and Image processingPorte, Laurence2017-04-23 12:30:442021-11-03 04:52:412017-04-23 12:30:44enConference papers10.33012/2016.145991A safety critical application needs to ensure that it can provide a given service even in worst case conditions. For a positioning service based on GNSS, one way of doing this is to compare the worst-case received C/N0 at the correlator output against the minimum required C/N0 at the correlator output to perform adequately certain key processing(s). This analysis is thus done in two steps: 1 - Assessing the minimum C/N0 threshold above which the receiver can fulfill specific requirements, 2 - Assessing the worst-case C/N0 that the user receiver will encounter in the field. In the case of civil aviation receivers, the C/N0 threshold can be set regarding specific acquisition, tracking or data demodulation requirements, while the worst case received C/N0 is computed based on the worst case interference environment encountered by an aircraft. A first assessment of the C/N0 threshold and of the worst-case received C/N0 can be obtained analytically based on mathematical models. Such assessment then requires to make assumptions on the receiver configuration, which is generally assumed to be representative of a minimum receiver used by the targeted application. This paper looks specifically at the assessment of the worst-case effective received C/N0 at the correlator output. This computation is based on the assessment of the degradation of the C/N0 due to the propagation channel, which is essentially dependent upon the presence of interference, the user antenna gain pattern, the effect of the RF front-end filter (bandlimiting, insertion losses, ADC, ...) and the desired signal type. The effects of each of these degradation sources have been thoroughly investigated and accurate mathematical models exist to represent them. The general formula to represent the effective C/N0 at the correlator output is given in [RTCA, 2006]. Typical analysis of the effective C/N0 in a given environment have looked at the effect of the ADC on the useful signal and thermal noise terms, but then only consider that the same effect applies to the interference sources. It sometimes occurs that a specific ADC-related degradation is added to the inter- and intra-GNSS system interference term, although the justification is generally not provided. This indicates that the effect of the ADC on the equivalent noise generated by interference might not be well taken into account, which might have a significant impact in the effective C/N0 assessment in a high interference environment. To account adequately for the ADC is challenging for 2 reasons: 1 - It does affect, potentially in a different way, different types of interference, 2 - It consists in a non-linear operation that might have a different effect depending upon the characteristics of the total interference (spectrum, power, etcâ€¦), These effects of the ADC were highlighted by [Betz and Schnidan, 2007] regarding the case of the inter-GNSS system interference. Most notably, the authors of this reference took the specific case of a 1-bit quantizer and showed through simulations that the actual interference effect on a GNSS signal considering the ADC could be significantly different (in better or worse) from the effect of this same interference without considering the ADC. They also showed that this effect was dependent upon the interference type. As a consequence, if this effect is not well modeled, this means that the resulting effective C/N0 assessment could be wrong by several tenths, or even even several units, of dB. Going further in this direction, [Hegarty, 2010] proposed an analytical model to represent the effect of quantization on GNSS signal reception considering any type of interference. In this reference, the author showed that the actual C/N0 loss in presence of interference due to the ADC depended upon the RF filter bandwidth, the interference Power Spectral Density (PSD), the receiver sampling frequency and the number of bits used by the ADC. The present article intends to analyze further the effect of the quantization in presence of a large number of simultaneous interferers (potentially powerful) by building on the methodology from [Hegarty, 2010]. The idea is thus to define and model the ADC contribution in the equivalent noise associated to a given interference environment and for a given receiver configuration (the case of pulsed interference is not covered in this paper). To do so, the analysis will be incremental: - First, the effect of one single interference (in presence of thermal noise) will be analyzed as a function of the number of ADC bits, the interference power with respect to the thermal noise level, the sampling frequency used by the receiver and the RF front-end filter bandwidth. Several types of interference (inter-GNSS system interference and narrow to wideband continuous interference) will be analyzed, which will allow getting an insight on the type of degradation to expect from a specific interference PSD. The investigated victim GNSS receivers will be receivers adapted to Galileo E1 OS, Galileo E5a, GPS L1 C/A and GPS L5 signals. - Second, the combination of several interferences will be investigated to assess the degree of non-linearity brought by the quantizer. This point will be assessed mostly as a function of the number of bits used by the quantizer and the interference characteristics. The results will be compared to the results obtained from the single-interference case. - Finally, a typical civil aviation worst-case environment in the L1 band will be investigated. This environment will consist in a mix of inter- and intra-GNSS system interference together with narrow band interference. All these results will be provided using a simulation tool based on analytical formulas and will be validated by extensive simulations. The conclusion of this article will provide a refined model of the degradation of the C/N0 as a function of the desired signal, receiver characteristics (number of bits of the quantizer, sampling frequency, front-end filter bandwidth) and interference type.