FR3050830A1 - METHOD AND DEVICE FOR CALIBRATING GNSS SIMULATORS - Google Patents

Abstract

The invention relates to a method and a device (10) for determining an internal bias (12), with respect to a reference clock signal (14), of a GNSS simulator (16) making available as an output a radio navigation signal (GNSS signal) in an RF band (18). The method includes executing a scenario by the simulator in which the theoretical bias between the GNSS signal and the reference clock signal is known. The method also includes digitizing the GNSS signal, generating a replica of the GNSS signal code, evaluating a correlation function between the GNSS signal and the code replica, identifying a correlation peak. of the correlation function and the deduction of the internal bias of the simulator. The evaluation of the correlation function is performed by Fourier transformation and inverse Fourier transformation. The deduction of the internal bias of the simulator is based on the correlation peak identified and the theoretical bias. The invention can be used to calibrate GNSS simulators and thus, indirectly, to calibrate GNSS receivers.

Description

TECHNICAL FIELD [0001] In general, the invention relates to a radionavigation satellite constellation simulator, more particularly to a method and a device for calibrating such a simulator.

BACKGROUND [0002] The radionavigation (GNSS) signals emitted by the satellites of a Global Navigation Satellite System (GNSS) are in the form of a carrier radio frequency (RF) modulated by a spreading waveform containing a pseudo-random binary code ("spreading code"). Incidentally, the GNSS signals carry useful data (including the navigation message containing the ephemeris data) in the form of a digital sequence (at a rate significantly lower than the pseudo-random code) further modulated on the carrier. In a simplified manner, a radionavigation signal receiver (GNSS receiver) determines its position by trilateration by measuring the time (and thus the distances) of the course of several GNSS signals from different GNSS satellites.

[0003] A radionavigation satellite constellation simulator (GNSS simulator) emulates the radiofrequency environment as it would appear to a GNSS receiver at a given place, date and time. In other words, a GNSS simulator produces, based on input parameters including the location, date and time of reception, all the radio frequency (RF) signals that a GNSS receiver would receive from the constellation real radionavigation satellites if it were placed in the said place at the said date and time. The simulated receiving location does not need to be invariable but may change over time in the context of a mobile receiver scenario. The simulator notably models the movement of the radionavigation satellites (GNSS satellites) and the time of travel of the signals between the different satellites and the location of reception. According to the improvement of the simulator, it can also simulate degradation of the GNSS signals, such as, for example, disturbances of the signals due to the ionosphere, the troposphere, GNSS clock drifts, multipath effects, etc. .

[0004] GNSS simulators are used to test GNSS receivers, prototypes and receiver components under perfectly controlled conditions.

In a simulator, before combining the signals on the RF output, each GNSS signal is generated in a separate channel. These channels include electronic and RF components that are physically independent. Separate generation of GNSS signals may result in slight differences between GNSS signals produced. In particular, the GNSS signals can be assigned different (temporal) bias (inter-channel bias) uncontrolled. These unknown biases limit the accuracy of the simulation because they are reflected in the measurements made by the GNSS receiver to be tested.

[0006] GB 2524340 A discloses a method for aligning radionavigation signals of a GNSS simulator with one another. In the method, a first bias is determined between a first code signal delivered by a first channel and a reference code signal. Other biases are determined between code signals provided by other channels and the continuation of the reference code signal. The biases are then applied to the different channels by control logic so as to align the channels with each other.

Technical Problem [0007] An object of one aspect of the present invention is to improve existing solutions for calibrating GNSS simulators.

General Description of the Invention [0008] A first aspect of the invention relates to a (calibration) method for determining an internal bias of a radionavigation satellite constellation simulator (GNSS simulator), with respect to a signal reference clock, the GNSS simulator providing at the output a radionavigation signal (GNSS signal) in RF band. The method comprises executing a scenario by the simulator in which the theoretical bias between the radionavigation signal and the reference clock signal is known. The method also includes digitizing the radionavigation signal (possibly after transposition to an intermediate frequency), generating a replica of the radionavigation signal code, evaluating a correlation function between the radionavigation signal and the replica of code, the identification of a correlation peak of the correlation function and the deduction of the internal bias of the simulator. The evaluation of the correlation function is performed by Fourier transformation and inverse Fourier transformation, ie. by determining the Fourier transforms of the radionavigation signal and the code replica and by inverse Fourier transformation of the product of the Fourier transforms. The deduction of the internal bias of the simulator is based on the correlation peak identified and the theoretical bias.

In the context of this document, the term "scenario" means the set of parameters, describing a situation and possibly its evolution over time, provided to the GNSS simulator so that it can calculate and generate the RF environment. of the receiver as if it were in the situation described. The setting options available may depend on the type of simulator. However, as this is equipment dedicated to specialists, the freedom to define scenarios is very broad.

Note that the scenario of the simulation can be chosen so that the theoretical bias is equal to a certain value, eg 0, an integer number of code periods or an integer number of chips (in English: "chips") of code. In this context, however, it should be noted that the theoretical bias does not need to be known in advance and / or when the scenario is run. For the invention, it is indeed sufficient that the simulation scenario makes it possible to deduce the theoretical bias by purely theoretical considerations and that the theoretical bias is known at the time of calculating the internal bias.

As stated above, the simulator can be used to model radio frequency environments, in a controlled and traceable manner, for a given scenario, at a receiver. For a given scenario, it is therefore possible to determine and / or know the bias generated by the propagation of a GNSS signal to the receiver. This bias represents the "theoretical bias" mentioned above. The measurable bias in a receiver test consists of theoretical bias, internal simulator bias, and receiver internal bias. Therefore, in order to be able to calibrate the receiver accurately, one must know not only the theoretical bias (depending on the parameters of the simulation) but also the internal biases of the simulator. An interesting application is therefore to use a GNSS simulator calibrated according to the method of the present invention for calibrating a GNSS receiver. Such a calibration method would be greatly facilitated over currently used calibration methods. The use of a simulator to calibrate a receiver offers advantages over real-world calibration, ie, on radionavigation signals from one or more radionavigation satellite constellations. These advantages are, among others, the repeatability, the very good control of the test conditions, the very good accuracy of the calibration in a laboratory situation, a reduced cost for the test phase of the receiver.

The GNSS signal provided by the GNSS simulator is preferably digitized at a sampling frequency of between 0.1 GHz and 1 GHz, more preferably at a sampling frequency of between 0.1 GHz and 0.5 GHz. GHz, still more preferably at a sampling frequency of between 0.1 GHz and 0.2 GHz. It should be noted that even if such sampling frequencies do not allow the GNSS signal carrier to be sampled correctly in the sense of the Nyquist-Shannon sampling theorem, they allow the sampling code to be correctly sampled. or the data included in the GNSS signal.

Preferably, after launching the simulation scenario, the GNSS signal is digitized for a time less than or equal to a period of the code included in the GNSS signal. In this way, it is ensured that the correlation function between the GNSS signal and the code replica has a unique correlation peak (a global maximum) on the correlation time. Sampling time and sampling frequency can vary dynamically (ie over time) as needed. For example, when a radionavigation signal comprises a GPS C / A code (comprising a sequence of 1023 bits of code at a rate of 1.023 Mcps (1 Mcps = 1 "mega-chips per second" = 106 chips per second)), the scan time of the radionavigation signal is less than or equal to 1 ms. At a sampling rate of 1 GHz, the C / A code is sampled 1 million times per code period, about 1000 times per chip.

The reference signal may be a signal provided by the simulator, eg a signal 1PPS (for "1 pulses per second" in English, pulse signal per second). The reference time for calculating a bias with respect to a reference signal may be a rising edge or a falling edge of the reference signal. Alternatively, the reference signal can be provided by an external system, for example, by an external clock. The clock of the GNSS simulator is preferably slaved to an external clock, also used by the calibration device used to execute the method.

A (local) replica of the code included in the GNSS signal is generated and sampled (or alternatively directly supplied in the form of samples) for use in calculating the correlation function. The replica may, for example, be sampled at the same sampling rate as the GNSS signal. The sampling time of the replica is preferably equal to the period of the code included in the GNSS signal.

The identification of a correlation peak of the correlation function preferably comprises evaluating the module of the correlation function between the code replica and the GNSS signal from the simulator. The calculation of the correlation function module is recommended if specific measures have not been taken to ensure that the correlation function is real (or purely imaginary). If we denote by C (t) the correlation function between the code replica r (t) and the GNSS signal s (t) coming from the simulator, where t is the time, we can write:

(1) where / is the integration interval, the code replica r (t) is shifted by a time τ (skew) with respect to s (t), FT [f (t)] represents the transformation of Fourier of a function f (t) and FTI [f (t)] represents the inverse Fourier transformation of a function f (t). Each point of the correlation function therefore corresponds to the integration of the function resulting from the multiplication of the GNSS signal with a time-shifted version of the code replica. In a pictorial way, the correlation function C (j) represents the similarity between the code replica r (t) and the GNSS signal s (t) as a function of the time shift τ. The module of the correlation function is maximum when the GNSS signal code is superimposed on the code replica. The area around the maximum of the correlation function module has, generally, a triangle shape. Depending on the GNSS signal type, the module of the correlation function may comprise a succession of triangles. This is the case, for example, if the modulation is of the Binary Offset Carrier (BOC) type. The maximum of the modulus of the wave function possibly has a clipped peak due to the imperfections of the GNSS signal. These imperfections may be due to, for example, a rise time or a significant descent time with respect to a chip of the code signal, overshoot ("overshoot" in English), etc.

The maximum of the module of the correlation function or peak correlation can be identified by detecting the maximum values sampled this function respectively its module. The offset in time, for which the value of the module of the correlation function is maximum, can thus be identified. Optionally, in order to increase the accuracy of the identification of the correlation peak, one of the methods described in the remainder of the document can be used to detect the maximum of the correlation function or its module. All these methods operate directly on the correlation function or its module. It will be appreciated that it is possible to combine the results of these methods (for example by calculating the average of these).

A first method comprises an adjustment (in English "fit") by a straight line of each flank of the correlation peak. The adjustment area of the correlation function or its module (ie the points of the correlation function or its module considered for adjustment) of each line is configured to contain only the linear area (flanks) around the correlation peak. The intersection of these two lines identifies both the value of the theoretical maximum (ie, without clipping) of the correlation function or its modulus and the value of the associated time offset (bias). The correlation peak is thus located precisely. Optionally constraints may be imposed during the adjustment. These constraints may include, for example, a symmetry constraint of the flanks of the correlation peak.

A second method comprises adjusting the correlation function or its module by the theoretical function (ie the expected correlation function for the modulation in question). The location of the correlation peak can be aided by the definition of a search zone of the maximum of the correlation function or its module or by the definition of initial values close to the real values of the correlation peak for the adjustment of the correlation peak. correlation function. The adjustment provides the time offset associated with the maximum of the correlation function respectively of its module.

A third method of identifying the peak correlation comprises the location of the maximum by an early-late approach (in English "early-late"). In this case, place two points separated by a given distance on either side of the culminating zone of the correlation function so that they are at the same height (value on the C axis ( τ)) and the abscissa (value on the τ axis) of the middle of the segment connecting the two points is determined. This process is repeated for different distances between the two points. In the end, the abscissa of the correlation peak is defined as the average value of the abscissae of all the environments found.

A fourth method comprises the calculation of the Fourier coefficients involved in the Fourier decomposition of the correlation function or its module. The location of the correlation peak includes determining the maximum, and the associated offset, of the resulting correlation function of the Fourier decomposition.

According to one embodiment of the invention, the possible Doppler shift of the GNSS signal is determined. The GNSS signal can effectively include a frequency offset due, for example, to a relative movement between the (simulated) satellite transmitting GNSS signals and the receiver, a drift of the satellite's internal clock (simulated) or drifting between the internal clock of the simulator with respect to the internal clock of the calibration device used to execute the method. The Doppler shift of the GNSS signal at the output of the simulator can, for example, be determined iteratively. Alternatively, to avoid an iterative method, the frequency space can be directly cut finely.

The deduction of the internal bias of the simulator is based on the correlation peak identified and the theoretical bias. The time difference between the simulator output GNSS signal and the code replica corresponds to the sum of the theoretical bias and the simulator internal bias. The internal bias of the simulator can thus be found by the difference between the measured offset and the theoretical bias.

According to one embodiment of the invention, the simulator models several GNSS signals. These GNSS signals are generated on separate channels of the simulator. Each channel has its own internal bias. These internal biases can be determined sequentially or in parallel. The multiple GNSS signals generated by the GNSS simulator can then be aligned based on several internal biases. If the determination of the different internal biases is based on scenarios involving several theoretical biases, when aligning the GNSS signals, we will take them into account.

[0025] A second aspect of the invention relates to a (calibration) device for determining an internal bias, with respect to a reference clock signal, of a GNSS simulator providing, at the output, a signal GNSS resulting from the execution of a scenario in which the theoretical bias between the radionavigation signal and the reference clock signal is known. The device includes an input for the radionavigation signal from the GNSS simulator, a digitizer for digitizing the GNSS signal, a code replica generator configured to produce a replica of the GNSS signal code, a processor, an output configured to provide the GNSS signal. internal bias deduced by the processor. The processor is configured to evaluate a correlation function between the radionavigation signal and the Fourier transform and Fourier transform inverse code replica, identify the correlation peak of the correlation function, and derive from the identified correlation peak and bias. the internal bias between the radionavigation signal and the reference clock signal. The processor may be configured for said tasks programmatically or because it is a specially designed hardware, eg a specially adapted electronic circuit ("application-specific integrated circuit" in English or ASIC). or a programmable logic circuit (eg a "field programmable gate array" or FPGA)). In the case where the processor is configured for certain tasks by programming, it comprises a memory or it has access to a memory in which the corresponding program is recorded in an interpretable or executable manner by the processor.

Preferably, the code replica generator and the processor are configured to determine a Doppler shift of the radionavigation signal.

The device for determining a bias between a reference signal and a radionavigation signal from a radionavigation satellite constellation simulator is preferably an open-loop device, i.e. a device without a feedback loop for aligning, based on one or more biases, the one or more radionavigation signals from the simulator.

BRIEF DESCRIPTION OF THE DRAWINGS [0028] Other features and characteristics of the invention will become apparent from the detailed description of certain advantageous embodiments presented below, by way of illustration, with reference to the appended drawings which show:

Fig. 1: a diagram of a calibration setup of a GNSS simulator using a calibration device;

Fig. 2: a timing diagram of a 1PPS reference clock signal and a GPS L1 C / A signal generated by a GNSS simulator;

Fig. 3: a timing diagram of a 1PPS signal and a spreading code of a GNSS signal generated at the output by a GNSS simulator;

Fig. 4: a flow chart comprising the steps of a method for determining an internal bias of a GNSS simulator;

Fig. 5: an illustration of the determination of a correlation function between a GNSS signal generated by the GNSS simulator and a pseudo-random code local replica;

Fig. 6: the graph of a correlation function obtained;

Fig. 7: an illustration of an adjustment of a correlation function by two lines;

Fig. 8: an illustration of an adjustment of a correlation function by a triangle function;

Fig. 9: an illustration of the determination of the correlation peak by an approach of advance-delay type;

Fig. 10: an illustration of the different correlation functions for different Doppler shifts;

Fig. 11: a part of a GNSS signal at the output of the GNSS simulator observed by an oscilloscope after amplification.

DETAILED DESCRIPTION OF EMBODIMENTS OF THE INVENTION [0029] FIG. 1 illustrates a calibration device 10 for determining an internal bias 12 of a GNSS simulator 16 outputting a GNSS signal 18 and a reference clock signal 14. The reference clock signal is, in the case describes a 1PPS signal. The calibration device 10 has a first input for the reference clock signal 14 and a second input for the GNSS signal 18. The calibration device 10 also has an output in order to provide the determined internal bias. In order to avoid or at least limit errors linked, for example, to a possible drift between the internal clocks of the GNSS simulator 16 and the calibration device 10, an oscillator 20 supplies a reference frequency or a reference clock signal common to the GNSS simulator 16 and to the calibration device 10.

The GNSS simulator 16 comprises different channels, made by separate electrical and / or electronic components, which are responsible for generating GNSS signals. The generation of GNSS signals by independent channels may result in differences between the GNSS signals produced. In particular, the GNSS signals can be assigned different angles which limit, if they remain unknown, the precision of the simulations carried out by the GNSS simulator 16.

The GNSS signals are generated in accordance with a simulation scenario including, inter alia, trajectories for the GNSS satellites and the GNSS receiver. Since the scenario executed by the GNSS simulator 16 can be configured virtually at will, it is possible to model realistic situations or even situations that are impossible to achieve, eg a scenario in which the GNSS satellites are on other than their orbits. real. The GNSS simulator is configured to send, during the execution of a simulation, and according to the scenario chosen by the user, instructions to the channels so that they generate the corresponding GNSS signals.

[0032] FIG. Fig. 2 shows an example of signals output from the GNSS simulator 16. The reference clock signal (14, see Fig. 1) takes the form of a 1PPS signal 22, i.e. a signal having a pulse every second but constant elsewhere. Ideally, the rising edge of the 1PPS signal is aligned at the beginning of a code period for each GNSS signal for zero propagation distance. However, for each channel of the GNSS simulator 16, there is a bias between the rising edge of the signal pulse 1PPS and the code transition and this bias is not identical (except after a possible prior calibration and unless accident) for all GNSS signals. To determine this bias, one might think of simply measuring the time between the rising edge of the 1PPS signal pulse and the code transition using an oscilloscope. To do this, it would first be necessary to amplify the GNSS signal concerned, which would require the introduction of a high gain amplifier. Since the pseudo-random code modulates a carrier, the observation that can be made on an oscilloscope is a rapidly oscillating waveform, the amplitude of which is locally reduced at the instant of the code transition (see FIG. 11). This approach does not make it possible to determine the bias concerned to less than 200 ps, which is considered insufficient for calibration purposes.

The method according to the invention relies on a correlation technique to detect the internal bias. This technique makes it possible to overcome the amplification of the GNSS signal or at least to use a less powerful amplifier.

For the illustration of a preferred embodiment of the invention, the GNSS signal output of the simulator is represented by a GPS signal C / A 24 transmitted in the frequency band L1. However, it will be understood that the method applies to any other GNSS signal. Since the period of the C / A code 24 is 1 ms, a peak of the signal 1PPS 22 will occur every thousand periods of the code.

[0035] FIG. 3 illustrates a bias between the signal 1PPS 22 (graph (a)) and the pseudo-random code 26 of a GNSS signal at the output of the simulator (graph (b)). The bias 28 between the rising edge of the signal 1 PPS and the beginning of the code 26 consists of a theoretical bias 32, due to the (simulated) propagation of the GNSS signal between the GNSS satellite and the receiver, and an internal bias 30, specific to the simulator channel that generated the signal. The durations shown in FIG. 3 are not necessarily representative of the proportions between the internal kiss 30 and the theoretical bias 32 forming the total bias 28. The theoretical bias 32 can be determined on the basis of the scenario used. It is advantageous, during the calibration of the simulator, to select a scenario, in which the theoretical bias is constant (eg 0). A scenario is therefore preferably used in which the relative movement between the GNSS satellite and the receiver is zero and in which the GNSS satellites are all in a geostationary orbit, etc.

In order to be able to visualize the internal bias more easily, a scenario can be conceived in which, in the absence of internal bias, the rising edge of the 1PPS signal would be aligned with the beginning of a fragment of the code (e.g. the first snippet of the code period). Such a scenario is now described in more detail. In order to make the scenario as simple as possible, we consider a propagation of GNSS signals in a vacuum. Note the distance between the GNSS satellite and the receiver. So that, in the absence of internal bias, the rising edge of the signal 1 PPS is aligned with the beginning of a snippet of code, it is necessary that the condition

be satisfied (c representing the speed of light in the vacuum, TCO of the code period and Tbribe the duration of a snippet of code.) To fulfill this condition in the simulation scenario, can place the GNSS satellite in an orbit geostationary (eg at 0 ° E alt alto) and the receiver at a distance just below the satellite (eg at 0 ° E, 0 ° N, and at altitude alt. d). The values of N and k can be chosen freely; eg N = 0 and k = 0 places the receiver at the transmitter location. With such a choice of the distance d, any offset between the rising edge of the signal 1 PPS and the beginning of a chip of the code is directly attributable to the internal bias of the simulator (see graph (c) of Fig. 3).

[0037] FIG. 4 illustrates the method according to one aspect of the invention for determining the internal bias of a GNSS simulator in the form of a flow chart. The first step S10 comprises the calibration (in phase, amplitude, TPG, etc.) of electrical and / or electronic elements between the simulator and the calibration device (for example electrical cables 32, see FIG. possibly electrical and / or electronic elements of the calibration device (for example amplifiers, etc.). This calibration is preferably performed using a vector network analyzer ("vector network analyzer" in English). In the next step S12, the scenario to be used for the determination of the internal bias is generated and encoded in the simulator. Advantageously, the scenario described above is used.

The scenario is then executed by the GNSS simulator in step S14 and the calibration device digitizes the output signal of the simulator. Digitization can be done, for example, at sampling frequencies between 0.1 GHz and 1 GHz. The carrier of the GNSS signal will thus be undersampled and the GNSS signal code will be correctly sampled (ie at a frequency high enough to meet the requirements of the Nyquist-Shannon sampling theorem). . For easy determination of the internal bias by the correlation between the GNSS signal and the 1PPS signal, a replica of the code included in the GNSS signal is generated and sampled at the same frequency by the step device S16. The generation of the code replica is made in accordance with the value of k chosen in step S12. In this way, a correlation peak at 0 on the abscissa means that the GNSS signal code is aligned with the generated code replica. The GNSS signal from the GNSS simulator is sampled for a time less than or equal to a period of the signal so that only one correlation peak is present in the correlation function. The digitized and sampled GNSS signal is completed with zero-padding values in order to obtain the same number of points between the replica of the sampled code and the sampled GNSS signal.

A correlation function is calculated at step S18 according to Eq. 1. In order to avoid the phase of the carrier, the module of the correlation function will be considered in the following. Fig. 5 represents the correlation between the pseudo-random code 26 of the GNSS signal and the code replica 31 of the GNSS signal. For each time offset value τ of the code replica 31 of the GNSS signal, a value of the correlation function is associated. In other words, for each given translation (time offset τ) of the code replica (denoted by the full arrow in Fig. 5), a value of the correlation function is determined. The value of the maximum of the correlation function gives the time offset between them and is related to the desired value of the internal bias 30 of the GNSS simulator. It should be noted that the peak width of the correlation function shown in FIG. 5 is not representative of a real case.

An example of a correlation function is given in FIG. 6. Ideally, the correlation function is of triangular shape 36. However, the GNSS signals at the output of the GNSS simulator are tainted with imperfections induced mainly by the crossing of the signals in the constituent components of the GNSS simulator, the calibration device and the connection elements. These imperfections cause a capping 34 of the correlation peak visible in FIG. 6. The flanks 38 of the correlation peak are, however, not (or very little) affected by these imperfections.

According to a preferred embodiment of the invention and with reference to FIG. 7, the identification of the correlation peak 40 of the correlation function 44 at step S20 is advantageously accomplished by a linear (i.e. straight line) adjustment of each flank of the correlation peak. Only the points included in the adjustment zone are taken into account to make the adjustment. Correlation peak adjustment can be achieved with a symmetry constraint on both lines (ie, same absolute values of the slopes of the straight lines) as there is no source of asymmetry, such as as interference multipath type (or multipath, "multipath" in English), in the scenario considered. The intersection 40 'of these two straight lines 42 identifies the correlation peak 40 and in particular the desired time shift 48.

According to another embodiment of the invention and with reference to FIG. 8, the identification of the correlation peak 50 of the correlation function 44 in the step S20 is carried out by an adjustment of the correlation peak by a triangle function 46. In order to identify the correlation peak 50, a search zone 49 is set during the adjustment. The peak of the triangle 50 '(the maximum of the resulting triangle function of the adjustment) identifies the correlation peak 40 and its abscissa the time offset 52. Optionally, when adjusting, a higher weight can be assigned to the points. included on the flanks of the correlation peak.

According to a third embodiment of the invention and with reference to FIG. 9, the identification of the correlation peak 54 of the correlation function 56 in step S20 is performed by an approach of the advance-delay type. According to this approach, two points 58, 60 are placed to the left and right of the correlation peak 54 at the same height (i.e., the two points have the same ordinate). The middle 62 of the segment 64 between the two points 58, 60 is determined and saved. The method is repeated for a predetermined number of pairs of points 66, 68, 70, 72 in left and right selection areas 76 of the correlation peak 54. The media 62, 78, 80 of the peer-created segments of points and saved are reported in a histogram and the average of these media identifies the desired time shift 82.

According to a fourth embodiment of the invention, the correlation peak is identified by Fourier decomposition of the calculated correlation function. After determining the Fourier coefficients associated with the Fourier decomposition, the maximum of the decomposition is sought, which thus provides the time shift.

With reference to FIG. 4, the next step S22 of the method deduces the value of the internal bias from the time offset found in step S20. Given the definition of the code replica in the example discussed in detail (theoretical bias = 0), the time offset is in this case identical to the internal bias sought.

Optionally, in step S24, another determination of the internal bias following the same or another scenario encoded in the GNSS simulator is possible. The method thus enters a loop to determine the internal bias and can, for example, average the internal bias thus determined.

Once the theoretical bias and the internal bias of the known GNSS simulator, an alignment of the GNSS signal with the 1PPS signal can be achieved by, for example, a feedback loop or, preferably, by a precise correction in loop opened.

In the case where the GNSS simulator models several GNSS signals, several GNSS signal generation channels are used. The method of determining an internal bias of a GNSS simulator channel can be used to determine the internal bias of each channel of the simulator. After determining the internal biases, an alignment of the GNSS signals to the 1PPS signal is performed.

Optionally, the possible Doppler shift of the GNSS signal is determined. The GNSS signal can effectively include a frequency offset due to, for example, a scenario comprising a relative movement between the GNSS satellite and the receiver or a drift between the internal clock of the GNSS simulator relative to the internal clock of the device calibration used to perform the process. The Doppler shift of the GNSS signal can, for example, be determined iteratively. In this case and with reference to FIG. 10, the frequency space (or Doppler space), corresponding to the space of the possible Doppler shifts of the GNSS signal, is divided into several Doppler classes 84 ("Doppler bins" in English). At first, the frequency space is roughly cut (ie fairly wide Doppler classes as shown in Fig. 10). For each Doppler class 84, the GNSS signal is multiplied by a carrier whose Doppler frequency is that associated with the Doppler class 84; The correlation function between the multiplied GNSS signal and the replica of the code included in the signal is determined according to the Eq. 1; The module of the correlation function 86 is calculated; • the global maximum 88 of the correlation function (ie, time shift and Doppler frequency) of all correlation functions is identified

The iterative process saves the time shift Tmax and the doppler frequency fo.max associated with the global maximum. At the next iteration, the Doppler class for which the overall maximum has been obtained is redrawn into finer classes and the iterative process is repeated. The iterative process is repeated until a desired accuracy is obtained for the Doppler shift and the time offset. As an alternative to the iterative method, a fine division of the frequency space can be envisaged from the beginning so as not to resort to an iterative method.

While particular embodiments have just been described in detail, those skilled in the art will appreciate that various modifications and alternatives to these can be developed in light of the overall teaching provided by the present disclosure of the invention. the invention. Therefore, the specific arrangements and / or methods described herein are intended to be given by way of illustration only, with no intention of limiting the scope of the invention.

Claims (13)

claims 1. A method for determining an internal bias, with respect to a reference clock signal, of a radionavigation satellite constellation simulator providing an output of a radionavigation signal, characterized by the execution of a scenario by said simulator in which the theoretical bias between the radionavigation signal and the reference clock signal is known; digitizing said radionavigation signal; generating a replica of the code of the radionavigation signal; evaluating a correlation function between the radio navigation signal and said code replica by Fourier transform and inverse Fourier transform; identifying a correlation peak of the correlation function; and the deduction, based on the identified correlation peak and the theoretical bias, of said simulator internal bias. 2. The method of claim 1, wherein the digitization of said radionavigation signal is at a sampling frequency between 0.1 GHz and 1 GHz. 3. The method according to any of claims 1 to 2, wherein said radionavigation signal is digitized for a time less than or equal to a period of the code signal of said radionavigation signal. 4. The method according to any one of claims 1 to 3, comprising determining a Doppler shift of said radionavigation signal. The method according to any one of claims 1 to 4, wherein identifying a correlation peak comprises adjusting by a straight line of each flank of the correlation peak and the location of the intersection of these two. straight. The method according to any one of claims 1 to 4, wherein identifying a correlation peak comprises adjusting the correlation peak by a theoretical function and locating the overall maximum of the resulting theoretical function of adjustment. The method of any one of claims 1 to 4, wherein identifying a correlation peak comprises locating the maximum by an early-late approach. The method according to any one of claims 1 to 4, wherein the identification of a correlation peak comprises the Fourier decomposition of the correlation function and the location of the maximum of the resultant function of the Fourier decomposition. . The method according to any one of claims 1 to 8, wherein the radionavigation satellite constellation simulator outputs a plurality of radionavigation signals, the method comprising determining a plurality of internal biases of the satellite constellation simulator. radionavigation. The method of claim 9, wherein the plural radionavigation signals from the radionavigation satellite constellation simulator are aligned based on the several internal biases. 11. A device for determining an internal bias, with respect to a reference clock signal, of a radionavigation satellite constellation simulator providing, at the output, a radionavigation signal, the simulator configured to execute a scenario wherein the theoretical bias between the radionavigation signal and the reference clock signal is known, the device being characterized by an input for said radionavigation signal; a digitizer for digitizing said radionavigation signal; a code replica generator configured to produce a replica of the code of said radionavigation signal; a processor configured to evaluate a correlation function between the radio navigation signal and said code replica by Fourier transform and inverse Fourier transform; o identify the peak correlation of the module of the correlation function; deriving from the identified correlation peak and the theoretical bias said internal bias between said radionavigation signal and the reference clock signal; and an output configured to provide said internal bias deduced by the processor. The device of claim 11, wherein the code replica generator and the processor are configured to determine a Doppler shift of the radionavigation signal. The device of any one of claims 10 to 12, wherein the device is an open loop device.