Calibration method for a conversion chain of an analog signal receiver and corresponding analog signal receiver
A pseudo-random sequence-based calibration method for analog signal receivers addresses the high cost and resource inefficiency of existing methods, providing accurate and efficient self-calibration with minimal resource usage.
Patent Information
- Authority / Receiving Office
- FR · FR
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-09-13
- Publication Date
- 2026-03-20
AI Technical Summary
Existing methods for calibrating analog signal receivers, particularly navigation receivers, are costly in terms of time and memory due to the need for multiple calibration signals, leading to measurement biases and errors in position determination.
A method involving the generation of a pseudo-random sequence to form a frequency-modulated analog calibration signal, which is injected into the conversion chain, allowing for the determination of a transfer function through correlation, enabling self-calibration with minimal resource usage.
The method reduces measurement biases and errors in analog channels by correcting distortions with minimal time and memory costs, enabling continuous and simultaneous calibration with signal reception.
Smart Images
Figure 00000000_0000_ABST
Abstract
Description
Title of the invention: Method for calibrating a conversion chain of an analog signal receiver and corresponding analog signal receiver
[0001] The present invention relates to a method for calibrating a conversion chain of an analog signal receiver, as well as the corresponding analog signal receiver.
[0002] Radio navigation, whether by satellites or by ground beacons, is implemented by means of a navigation receiver, configured to receive signals emitted by satellites or beacons and to determine a position based on the signals received.
[0003] The signal(s) are received on an analog antenna of the receiver and are processed by a signal processing chain generally configured to implement analog processing steps, an analog-to-digital conversion step and digital processing steps.
[0004] Imperfections in the implementation of the analog processing steps can result in transfer functions with non-constant amplitude and / or group propagation time in the receiver bandwidth, which result in the appearance of distortions on the output signals of the processing chain.
[0005] In the case where the signals to be received are not modulated in the same way and / or in the case where the receiver antenna is a Controlled Radiation Pattern Antenna (CRPA), distortions can result in different measurement biases on the different analog channels of the receiver, thus increasing the error on the position determined by the receiver.
[0006] To reduce these errors, it is known to calibrate analog channels of a navigation receiver by a method comprising determining the difference between the respective propagation times of the same signal in two different processing chains of the receiver, or of two differently modulated signals in the same processing chain.
[0007] To this end, respective calibration signals modulated in the same way as the signals whose propagation time we wish to determine, with pseudo-random spreading codes, are injected into the inputs of the processing chain, near the output of the receiver antenna, and compared with the respective demodulated signals obtained at the output of the processing chain, by correlation, This allows the different signals to be separated, thanks to the properties of pseudo-random spreading codes.
[0008] These processes are costly in terms of time and memory since they require the use of as many calibration signals as possible modulations
[0009] An object of the invention is therefore to propose a method for calibrating an analog signal receiver, in particular a navigation signal receiver, making it possible to reduce measurement biases on the analog channel(s) of this receiver and / or, where applicable, the differences between these measurement biases, with a reduced cost in time and / or memory.
[0010] To this end, the invention relates to a method for calibrating a conversion chain of an analog signal receiver, the conversion chain being configured to provide on a main output a main digital signal output from a main analog signal received on a main input, the calibration method comprising: a) the generation of a pseudo-random sequence of states with values in a set of discrete values; b) the formation of a frequency-modulated analog calibration signal from the pseudo-random sequence; c) the injection of the modulated analog calibration signal into the conversion chain; d) the frequency demodulation of an intermediate digital signal received on an input of the conversion chain in response at least to the modulated analog calibration signal, to form a digital signal to be processed; e) the determination of a correlation function between the digital signal to be processed and the pseudo-random sequence; f) determining a transfer function of the conversion chain from the correlation function; and (g) if necessary, the determination of at least one parameter of a processing step to be applied downstream of the main output on the basis of the transfer function.
[0011] In the invention, the modulated analog calibration signal, injected into the input of the analog signal receiver conversion chain, is formed from a wideband pseudo-random signal, so that the calibration signal can have a flat spectrum throughout the Shannon band and it is possible to inject it with negligible power compared to the total noise power in the receiver bandwidth.
[0012] The calibration signal can thus be injected into the analog receiving chain, possibly continuously, without degradation of the signal-to-noise ratio.
[0013] Due to the autocorrelation properties of the pseudo-random sequence used to generate the calibration signal, the correlation function between the digital signal to be processed, generated at least from the analog calibration signal injected into the conversion chain, and the pseudo-random sequence is the impulse response of the conversion chain. The transfer function of the conversion chain can then be determined by Fourier transforming this impulse response.
[0014] It is thus possible to determine a treatment to be applied to the main analog signal of interest, for example by means of an equalization filter, to correct the distortions induced by the conversion chain throughout the Shannon band on the basis of this transfer function.
[0015] The method according to the invention requires the injection of only a single calibration signal. This method is therefore inexpensive in terms of time and memory, and it allows for the self-calibration of the receiver, possibly continuously and / or simultaneously with the reception of navigation signals.
[0016] According to other advantageous aspects of the invention, the method comprises one or more of the following features, taken individually or in all technically possible combinations:
[0017] - the processing step includes a filtering step using an equalization filter, the determination of the processing step including the determination of at least one coefficient of the equalization filter;
[0018] - a first modulation frequency through which the calibration signal analog is modulated is included in a bandwidth of the receiver;
[0019] - a characteristic duration separating two successive states of the pseudo sequence random is chosen based on a sampling frequency of the conversion chain;
[0020] - the determination of the transfer function includes the calculation of a transform of Fourier of the correlation function;
[0021] - the determination of the correlation function includes subsampling of the pseudo-random sequence and the digital signal to be processed;
[0022] - the determination of at least one coefficient of the equalization filter includes the calculation of a plurality of equalization filter coefficients from the transfer function, this calculation including minimizing a distance between the transfer function and a predetermined ideal transfer function;
[0023] - the calibration process is implemented simultaneously with the reception of the main analog signal.
[0024] The invention also relates to an analog signal receiver comprising at least one conversion chain configured to provide on a main output a main digital signal output from a main analog signal received on a main input, configured to implement the calibration process as described above.
[0025] According to other advantageous aspects of the invention, the receiver comprises one or more of the following features, taken individually or in all technically possible combinations:
[0026] - the conversion chain includes:
[0027] i) an analog receiving chain comprising:
[0028] * the main entrance,
[0029] * an analog-to-digital converter downstream of the main input, and
[0030] * a first calibration unit, comprising a conversion subunit digital-to-analog and a modulation subunit, the first calibration unit being configured to inject a modulated analog calibration signal, modulated by a first modulation frequency, into the conversion chain upstream of the analog-to-digital converter,
[0031] ii) a digital processing module comprising:
[0032] * a digital signal processing unit downstream of the analog- converter digital, the processing unit comprising the main output and a filtering stage including at least one reconfigurable equalization filter upstream of the main output, and
[0033] * a second calibration unit comprising:
[0034] ** a sequence generation subunit, configured to generate a sequence pseudo-random discrete state sequences with values in a set of discrete values and provide the first calibration unit with an incoming digital calibration signal formed from said pseudo-random sequence for the formation of the modulated analog calibration signal, and
[0035] ** a computing subunit, configured for:
[0036] a) determine a correlation function between a digital signal to be processed received from the processing unit and the pseudo-random sequence, the digital signal to be processed being supplied to the computing subunit downstream of the digital-to-analog converter in response to at least said modulated analog calibration signal,
[0037] b) determine a transfer function of the conversion chain from the correlation function,
[0038] c) determine at least one coefficient of the equalization filter of the filtering stage based on the transfer function, and
[0039] d) transmit at least one coefficient to the processing unit for reconfiguration of the filtering stage;
[0040] - the receiver includes an antenna with a controlled radiation pattern in upstream of a plurality of respective main inputs of a plurality of chains conversion, the sequence generation subunit being common to several of said conversion chains.
[0041] The invention will become clearer upon reading the following description, given solely by way of non-limiting example, and made with reference to the drawings in which:
[0042] [Fig. 1] [Fig. 1] is a schematic representation of a first analog signal receiver according to the invention;
[0043] [Fig.2] [Fig.2] is a logic diagram representation of a calibration process for the receiver of [Fig.1];
[0044] [Fig.3] [Fig.3] is a schematic representation of a second analog signal receiver according to the invention;
[0045] [Fig. 4] [Fig. 4] shows two examples of pseudo-random sequences generated in step 205 of the process in [Fig. 2] (Figures 4a and 4b) and an example of an analog calibration signal formed from one of these sequences (Figure 4c), the x-axis representing time t on each of Figures 4a, 4b and 4c; and
[0046] [Fig.5] [Fig.5] represents an embodiment of part of the calculation subunit of [Fig.1].
[0047] The invention relates to an analog signal receiver 10 comprising at least one conversion chain 12 configured to provide on a main output 15 a main digital output signal NUM_p from a main analog signal ANA_p received on a main input 20.
[0048] The receiver 10 includes, for example, an antenna 22 connected to the main input 20.
[0049] Antenna 22 is, for example, configured to receive a signal emitted by a source of signal, such as a satellite or a terrestrial beacon and to provide the main analog signal ANA_p on the main input 20.
[0050] The signal emitted by the signal source is, for example, a radio navigation signal.
[0051] The signal emitted by the signal source comprises one or more components including the respective frequencies are included in at least one frequency band, referred to hereafter as the receiver bandwidth, after the modulation described in the following paragraph.
[0052] The signal emitted by the source is frequency modulated by means of a carrier signal whose frequency is referred to in the following as the "carrier frequency" f_port.
[0053] The carrier frequency f_port is for example between 1 GHz and 2 GHz, in particular on the order of 1.5 GHz.
[0054] In a particular embodiment, the antenna 22 receiving the signal emitted by the source is a CRPA antenna, that is to say, a controlled radiation pattern antenna.
[0055] The conversion chain 12 includes an analog receiving chain 25 and a digital processing module 30.
[0056] The analog receiving chain 25 includes the main input 20, an analog-to-digital converter 35 downstream of the main input 20, and a first calibration unit 40, configured to inject a modulated analog calibration signal ANA_cal,mod into the conversion chain 12 upstream of the analog-to-digital converter 35.
[0057] The modulated analog calibration signal ANA_cal,mod is modulated by a signal of frequency equal to a first modulation frequency fl.
[0058] In the example of [Fig.1], the modulated analog calibration signal ANA_cal,mod is formed from an analog calibration signal ANA_cal,b, resulting from the digital-to-analog conversion of an incoming digital calibration signal NUM_cal,ent entering the analog reception chain 25.
[0059] The first calibration unit 40 comprises a digital-to-analog conversion subunit 45 and a modulation subunit 50.
[0060] The digital-to-analog conversion subunit 45 is configured to receive the incoming digital calibration signal NUM_cal,ent from the digital processing module 30 which will be described later, and to provide a raw analog calibration signal ANA_cal,b to the modulation subunit 50.
[0061] The modulation subunit 50 is configured to provide the modulated analog calibration signal ANA_cal,mod, by modulation of the raw analog calibration signal ANA_cal,b by the first modulation frequency fl.
[0062] The modulation subunit 50 includes in the example a first local oscillator 55, configured to generate a sinusoidal signal carrying the raw analog calibration signal ANA_cal,b, said carrier signal having the frequency of the first modulation frequency fl.
[0063] The first modulation frequency fl is preferably within the bandwidth of the receiver 10.
[0064] The first modulation frequency fl is preferably close to the carrier frequency f_port, advantageously equal to the carrier frequency f_port.
[0065] The carrier frequency f_port is itself preferably close to the center frequency of the receiver's bandwidth.
[0066] In the example of [Fig.1], the sum of the modulated analog calibration signal ANA_cal,mod and the main analog signal ANA_p is formed at the output of an adder 57 to then constitute an incoming analog signal ANA_ent supplied at the input of the analog-to-digital converter 35. This embodiment is called the reference configuration in the following.
[0067] The reference configuration allows the calibration of the receiver 10 to be carried out simultaneously with the reception of the main analog signal ANA_p.
[0068] This arrangement makes it possible to increase the operational availability of the receiver 10 and to improve the quality of the main outgoing digital signal NUM_p in real time.
[0069] Alternatively, the calibration of the receiver 10 and the reception of the main analog signal ANA_p can be performed successively and possibly iteratively. In this embodiment, hereafter referred to as the alternative configuration, for the calibration of the receiver 10, the incoming analog signal ANA_ent can be formed solely from the modulated analog calibration signal ANA_cal,mod. Those skilled in the art will readily adapt the following description to this alternative.
[0070] Optionally, as shown in [Fig. 1], the analog receiving chain 25 includes a pre-amplification stage 60 comprising at least one low-noise amplifier device, placed downstream of the main input 20 and upstream of the analog-to-digital converter 35.
[0071] The pre-amplification stage 60 is configured to pre-amplify the main analog signal ANA_p before its summation with the modulated analog calibration signal ANA_cal,mod in the adder 57.
[0072] Optionally, as shown in [Fig. 1], the analog receiving chain 25 includes an intermediate frequency conversion stage 65 downstream of the main input 20 and, where applicable, downstream of the preamplification stage 60, and upstream of the analog-to-digital converter 35.
[0073] The intermediate frequency conversion stage 65 is configured to provide the analog-to-digital converter 35 with an incoming analog signal ANA_ent modulated by an intermediate frequency signal f_int lower than the carrier frequency f_port.
[0074] To this end, the intermediate frequency conversion stage 65 includes, for example, bandpass filters 65a and 65b, respectively upstream and downstream of a multiplier 65c. The multiplier 65c is configured to receive a sinusoidal signal generated by a second local oscillator 65d, configured to generate an intermediate carrier signal of frequency f_mod equal to the difference between the frequency f_port and an intermediate frequency f_int.
[0075] The intermediate frequency f_int is for example between 10 and 20 MHz, in particular around 15 MHz.
[0076] The analog-to-digital converter 35 is configured to convert the analog signal coming out of the adder 57 and, where applicable, of the intermediate frequency conversion stage 65, into an intermediate digital signal NUM_int received on an input 70 of the digital processing module 30.
[0077] The digital processing module 30 includes a digital signal processing unit 75, located downstream of the analog-to-digital converter 35 and configured to provide the main outgoing digital signal NUM_p on the main output 15, and a second calibration unit 80 upstream of the first calibration unit 40.
[0078] The sampling frequency of the digital processing module 30 is preferably equal to the sampling frequency Fe of the analog-to-digital converter 35.
[0079] The processing unit 75 includes the main output 15, a reconfigurable filtering stage 85 upstream of the main output 15 and, in the first embodiment shown in [Fig.1], a baseband conversion stage 90 upstream of the filtering stage 85.
[0080] The baseband 90 conversion stage is configured to demodulate the intermediate digital signal NUM_int, of intermediate frequency f_int in the example of [Fig.1], so as to bring this signal back to baseband BB and thus form a digital signal to be processed NUM_BB.
[0081] By baseband BB signal, we mean in a way known to a person skilled in the art a signal whose components are centered around a zero frequency, or in other words, a signal that is not modulated by a carrier of non-zero carrier frequency.
[0082] To this end, the baseband 90 conversion stage includes, for example, a multiplier 90a receiving the intermediate digital signal NUM_int and a sinusoidal signal of frequency equal to the intermediate frequency f_int generated by a third local oscillator 90b.
[0083] The baseband 90 conversion stage may also include one or more filters, in particular a low-pass filter with a bandwidth encompassing the baseband BB.
[0084] The filtering stage 85 is configured to receive the digital signal to be processed NUM_BB and to provide the main outgoing digital signal NUM_p on the main output 15 at least when a main analog signal ANA_p is effectively provided on the main input 20.
[0085] In the reference configuration, the calibration of the receiver 10 is implemented at the same time as the reception of the main analog signal ANA_p. In the reference example, the digital signal to be processed NUM_BB, when received by the filtering stage 85, is therefore formed in response to both the main analog signal ANA_p and the incoming digital calibration signal NUM_cal_ent.
[0086] In the alternative configuration, in which the calibration of the receiver 10 and the reception of the main analog signal ANA_p are carried out in a sequential, the digital signal to be processed NUM_BB received by the filtering stage 85 is formed for reception in response only to the main analog signal ANA_p.
[0087] The filtering stage 85 includes at least one finite impulse response equalization filter (or equivalently FIR filter).
[0088] The equalization filter is, in a known manner, made up of a delay line containing the stored received signal and complex multipliers each configured to multiply a received input signal by a respective complex coefficient Ci and an adder.
[0089] The transfer function of the equalization filter is chosen to be non-symmetric, with a non-zero group propagation time.
[0090] The number of complex Ci coefficients of the equalization filter determines the frequency resolution of this filter.
[0091] The number of coefficients Ci is chosen in particular according to the granularity of the defects of the analog transfer function.
[0092] For example, for a sampling frequency Fe of the analog-to-digital converter 35 equal to 50 MHz, the number of coefficients Ci can be at least equal to 51. The frequency resolution is in this case equal to 1 MHz if the filter is characterized by 51 coefficients Ci.
[0093] The filtering stage 85, and in particular the equalization filter, is reconfigurable.
[0094] For the reconfiguration of the filtering stage 85, the processing unit 75 is configured to receive at least one coefficient Ci of the filtering stage 85 from the second calibration unit 80 and to implement this coefficient for the reconfiguration of the filtering stage 85.
[0095] The processing unit 75 optionally includes a further processing stage 95 between the filtering stage 85 and the main output 15.
[0096] The subsequent processing of the subsequent processing stage 95 includes, by way of non-limiting example, an anti-interference processing and / or additional demodulation configured to isolate one or more useful signals, so as to form the main output digital signal NUM_p.
[0097] The processing unit 75 is configured to provide the second calibration unit 80 with the digital signal to be processed NUM_BB formed at least in response to the incoming digital calibration signal NUM_cal,ent, and therefore to the modulated analog calibration signal ANA_cal,mod.
[0098] In the reference configuration, the digital signal to be processed NUM_BB received by the second calibration unit 80 is obtained in response to the injection of the incoming digital calibration signal NUM_cal,ent into the first calibration unit 40 and in response to the main analog signal ANA_p.
[0099] In the alternative configuration, in which calibration and reception of the main analog signal are performed sequentially, the digital signal to be processed, NUM_BB, received by the second calibration unit 80, is formed solely in response to the injection of the incoming digital calibration signal, NUM_cal,ent, into the analog reception chain 25. It is therefore understood that in this alternative embodiment, the digital signal to be processed, NUM_BB, is not formed in response to the same signals for calibrating the receiver 10 and for receiving the main analog signal, ANA_p. In this case, the receiver 10 may include a switching device from a calibration mode to a reception mode.
[0100] The second calibration unit 80 comprises a sequence generation subunit 100 and a calculation subunit 105.
[0101] The sequence generation subunit 100 is configured to generate a pseudo-random sequence S of discrete states with values in a set of discrete values, form the incoming digital calibration signal NUM_cal,ent from the sequence S and provide the incoming digital calibration signal NUM_cal,ent to the first calibration unit 40.
[0102] The period of the pseudo-random sequence S, that is to say the duration separating two successive discrete states and the corresponding frequency are respectively called sequence period T'e and sequence frequency F'e.
[0103] The sequence frequency F'e is equal to a submultiple (or equivalently, to an integer divisor) of the sampling frequency Fe of the analog-to-digital converter 35.
[0104] The sequence generation subunit 100 includes for example a binary N-state shift register with values in the set {0,1} generating a periodic pseudo-random sequence S consisting of 0s and 1s, at the sequence frequency F'e.
[0105] The sequence generation subunit 100 may in this case include a device for converting the sequence S generated by the shift register into a physical signal whose value is equal to a predetermined amplitude Amax if the binary state is 0 and equal to the opposite of the amplitude Amax if the binary state is -1.
[0106] This arrangement allows the generation of a NUM_cal,ent signal without a DC component.
[0107] The calculation subunit 105 is configured to: * determine a correlation function R(NUM_BB,S) between the digital signal to be processed NUM_BB supplied to the calculation subunit 105 downstream of the digital-to-analog converter 35 in response to at least said modulated analog calibration signal ANA_cal,mod and the pseudo-random sequence S, * determine a transfer function H of the conversion chain 12 from the correlation function R(NUM_BB,S), * determine at least one coefficient Ci of the equalization filter of the filtering stage 85 based on the transfer function (H), and * transmit at least one coefficient Ci to the processing unit 75 for the reconfiguration of the filtering stage 85.
[0108] The computing subunit 105 is implemented at least partially in the form of software executable by a processor and stored in memory. Alternatively, the computing subunit 105 is implemented, for example, in the form of a programmable circuit of the FPGA (Field Programmable Gate Array) type and / or of the ASIC (Application-Specific Integrated Circuit) type.
[0109] We will now describe a calibration method 200 for the conversion chain 12 of the analog signal receiver 10 with reference to [Fig.2].
[0110] The process comprises: a) the generation 205 of the pseudo-random sequence S of states with values in a set of discrete values; b) the formation 210 of the modulated analog calibration signal ANA_cal,mod from the pseudo-random sequence S; c) the injection 215 of the modulated analog calibration signal ANA_cal,mod into the conversion chain 12; d) the 220 frequency demodulation of the intermediate digital signal NUM_int received on the input 70 of the conversion chain 12 in response at least to the modulated analog calibration signal ANA_cal,mod, to form the digital signal to be processed NUM_BB; e) the determination 225 of the correlation function R(NUM_BB,S) between the digital signal to be processed NUM_BB and the pseudo-random sequence S; f) the determination 230 of the transfer function H of the conversion chain 12 from the correlation function R(NUM_BB,S); and (g) if necessary, the determination 235 of at least one parameter of a processing step to be applied by means of the filtering stage 85 downstream of the main output 15 on the basis of the transfer function H.
[0111] The generation 205 of the pseudo-random sequence S is carried out by means of the sequence generation subunit 100.
[0112] A first example of a pseudo-random sequence of sequence frequency F'e is shown in Figure 4a.
[0113] Optionally, it is possible to slow down the frequency of the pseudo-random sequence S and the modulated analog calibration signal ANA_cal,mod, for example by generating only one pulse of duration 1 / F'e per time interval equal to L / F'e and replacing the missing values in the remainder of this time interval with the state 0, as shown in Figure 4b.
[0114] This arrangement makes it possible to reduce the disturbance of the main analog signal ANA_p by the modulated analog calibration signal ANA_cal,mod.
[0115] To this end, it is possible to generate the pseudo-random sequence S with a sequence period T'e equal to L / F'e, by replacing the missing values in the remainder of the time interval L / F'e with zero.
[0116] Optionally, it is possible to reduce the duration of the analog pulses generated at each sequence period T'e in the digital-to-analog conversion subunit 45 by a factor K greater than 1, the sequence period T'e remaining unchanged, as shown in Figure 4c.
[0117] This arrangement makes it possible to obtain a raw analog calibration signal ANA_cal,b with a spectrum that is as flat as possible in the Shannon band. Indeed, the power spectral density (PSD) of this signal as a function of frequency f is of the form: [0H8] DSP(f) = ~K~ slnC \KFÏ )
[0119] DSPo denoting in equation 1 a constant, sine denoting the cardinal sine function and F'e denoting the sequence frequency.
[0120] The formation 210 of the analog calibration signal includes in the example shown in [Fig.1] the supply by the sequence generation subunit 100 of the incoming digital calibration signal NUM_cal,ent formed from the sequence S to the digital-to-analog conversion subunit 45, to form the raw analog calibration signal ANA_cal,b.
[0121] The training step 210 also includes frequency modulation of the raw analog calibration signal ANA_cal,b by means of the modulation subunit 50 to form the modulated analog calibration signal ANA_cal, mod, modulated at the first frequency fl.
[0122] Alternatively, the transposition of the incoming digital calibration signal NUM_cal,ent into a frequency carrier with a frequency equal to the first modulation frequency fl can be carried out before the analog-to-digital conversion.
[0123] The injection 215 of the modulated analog calibration signal ANA_cal,mod is carried out upstream of the analog-to-digital converter 35.
[0124] In the example of [Fig.1], the modulated analog calibration signal ANA_cal,mod is introduced downstream of the pre-amplification stage 60 and upstream of the intermediate frequency conversion stage 65.
[0125] Consequently, the modulated analog calibration signal ANA_cal is subjected to at least some, advantageously all, of the conversion chain 12 steps which are likely to introduce distortions into the main analog signal ANA_p, in particular the analog filters.
[0126] In the reference configuration, the modulated analog calibration signal ANA_cal,mod is summed with the main analog signal ANA_p by means of the adder 57.
[0127] In this case, the calibration process 200 may include an attenuation step (not shown) before the injection 215. The attenuation step makes it possible to limit any degradation of the performance of the receiver 10 due to the injection 215 for the calibration of the receiver 10 simultaneously with the reception of signals of interest.
[0128] By way of illustrative example, if the receiver 10 receives an analog signal ANA_p from a satellite whose power is on the order of -130 dBm at the output of the antenna 22, possibly increased by 30 dB at the output of the pre-amplification stage 60, and if the power of the thermal noise in the digitized RF band is on the order of -97 dBm at the output of the antenna 22, increased by 30 dB at the output of the pre-amplification stage 60, it is possible to inject at step 215 an analog calibration signal modulated ANA_cal,mod of power of the order of -90 dBm spread in the Shannon band corresponding to a sampling frequency Fe equal to 50 MHz. In this case, the signal-to-noise ratio is equal to -90 dBm - (-97+30) dBm + 10.1ogl0(50 MHz)) or 54 dBHz after correlation and integration over a few seconds.Such a ratio is amply sufficient to estimate with satisfactory accuracy the correlation function R(NUM_BB,S) as will be described later with an integration time on the order of a few seconds.
[0129] It is important that the use of the pseudo-random sequence S for the formation of the modulated analog calibration signal ANA_cal,mod allows this signal to be spread over the entire Shannon frequency band [-Fe / 2,+Fe / 2] at the analog-to-digital converter 35. The modulated analog calibration signal ANA_cal,mod therefore has a relatively flat spectrum in this frequency band.
[0130] This arrangement minimizes the impact of the modulated analog calibration signal ANA_cal,mod on the reception of the main analog signal ANA_p of interest while obtaining accurate calibration of the receiver 10.
[0131] In response to at least the ANA_cal,mod signal, the intermediate digital signal NUM_int is received at input 70 of the digital processing module 30 and is demodulated during the demodulation step 220 in the baseband conversion stage 90, so as to form the digital signal to be processed NUM_BB used for the calibration process. The computing subunit 105 determines in the determination step 225 the correlation function R(NUM_BB,S) between the digital signal to be processed NUM_BB, generated at least from the pseudo-random sequence S, and the pseudo-random sequence S itself.
[0132] For determination 225, the pseudo-random sequence S is for example stored in a delay line, a respective correlation point being calculated by each of the correlators of a bank of correlators 105a of the calculation subunit 105. [Fig.5] represents such an embodiment.
[0133] Typically, the k-th point R_k of the correlation function R(CAL_f,S) is obtained by the discrete-time integral of equation 2:
[0134] Rk = Y?2 CAL k).T'e (2)
[0135] The number L of correlators can be chosen according to the desired frequency resolution for the subsequent determination 230 of the transfer function H.
[0136] By way of example, if the sampling frequency Fe is equal to 50 MHz, if the number L of correlators is greater than or equal to 100, the frequency resolution is equal to or better than 0.5 MHz.
[0137] The number of correlation points L is advantageously equal to or greater than the number of coefficients Ci of the digital equalization filter.
[0138] In a particular embodiment, the determination 225 may include the sequential calculation of each of the correlation points in turn. In this embodiment, the correlators share a common adder. L successive calculation cycles are then required to traverse all L correlators, each of the L cycles being performed by a respective correlator.
[0139] This provision makes it possible to limit the material cost of the determination 225, since only one adder can be shared between the L correlators in this case.
[0140] This amounts to using 1 sample out of L of the pseudo-random sequence S.
[0141] In this embodiment, it is possible to generate the pseudo-random sequence S with a sequence frequency F'e / L.
[0142] The integration time tn2-tni can be chosen according to the observed or desired signal-to-noise ratio. In the example described above, for which the signal-to-noise ratio is equal to 54 dBHz, an integration time of approximately 1 second makes it possible to obtain an accuracy of 2 per thousand to one standard deviation.
[0143] The calculation subunit 105 determines at step 230 the transfer function H of the conversion chain 12.
[0144] Due to the autocorrelation properties of the pseudo-random sequence S, the correlation function R(NUM_BB,S) is equal to the impulse response of the portion of the conversion chain 12 traversed by the modulated analog calibration signal ANA_cal,mod.
[0145] The transfer function H is the Fourier transform of this impulse response, therefore of the correlation function R(NUM_BB,S).
[0146] The computing subunit 105 can implement the calculation of a Fast Fourier Transform (FFT). This calculation can notably be performed on a number of points 2P, according to the Cooley-Tukey method, also known as the "butterfly" method, by supplementing the correlation points determined in the preceding determination step 225 with zeros to obtain the necessary 2P points.
[0147] In this case, the number of points 2P is chosen to be greater than the number of correlation points in order to obtain satisfactory accuracy in the calculation of the transfer function H. Typically, 2P can be chosen to be equal to 512 or even 1024, corresponding respectively to P = 9 and P = 10.
[0148] The determination 230 can be performed in software as a background task. It is then not limiting in terms of hardware cost or workload.
[0149] Once the transfer function H has been determined, the calculation subunit 105 can proceed to the determination 235 of at least one parameter of a processing step to be applied by means of the filtering stage 85.
[0150] The determination 235 includes the calculation of at least one coefficient Ci of the equalization filter.
[0151] This calculation can be based on the principles described in the following paragraphs.
[0152] The global transfer function H_tot of the conversion chain 12 taking into account the equalization filter, whose transfer function is denoted H_FIR, satisfies equation 3:
[0153] H tot = H*H FIR (3)
[0154] The transfer function of the equalization filter is related to the coefficients Ci of this filter according to equation 4:
[0155] Hp[R = .exp( -
[0156] The coefficients Ci can be chosen so that the overall transfer function gets as close as possible to an ideal transfer function H_id.
[0157] The ideal transfer function is for example of the form H_id( f ) = 1 in the baseband BB and zero otherwise.
[0158] For determination 235, the receiver bandwidth can then be sampled with a step q, for example equal to Fe / 2P, so as to select N frequencies f_i in the receiver baseband.
[0159] We then form a Hobs observation matrix comprising N rows H_obs(i) such that:
[0160] pp Jî\= [e+2jfPMTe e+2jf^(Ml).Te _ e-2j fp XM-D-Te e-2jfpMTe | ülf]
[0161]
[0162]
[0163]
[0164]
[0165]
[0166]
[0167]
[0168]
[0169]
[0170] With TT = 11 obs 1(6) We also form an ideal vector Y_id whose coefficients are the N values H_id(f_i). 'MA)b7) Then we look for the equalization vector C of equation 8 below, formed by the 2M+1 coefficients Ci allowing minimization of a predefined distance between the vector ideal Y_id and the product H_obs.C of the observation matrix H_obs and the vector equalization C. (8) In particular, the equalization vector C can be determined by the least squares method, that is, so that the magnitude II y _ u C II2 is minimal. The subsequent reconfiguration of the filtering stage 85, and in particular of the equalization filter, based on the equalization vector C thus determined makes it possible to correct the distortions induced in the main analog signal ANA_p by the portion of the conversion chain 12 traversed by this signal upstream of the filtering stage 85. The distortions of the useful signals received, due to the defects of the analog part of the conversion chain 12, are thus corrected in the digital signal NUM_p. In an alternative or complementary embodiment, determination 235 includes the determination of one or more digital treatments other than filtering. Thanks to the use of the pseudo-random sequence S, the modulated analog calibration signal ANA_cal,mod is broadband and of relatively flat spectrum, the duration of the pulses generated at the time of the analog-to-digital conversion allowing the flatness of this spectrum to be adjusted to a predetermined specification. The calibration of the conversion chain 12 can then be carried out with the injection of this single modulated analog signal ANA_cal,mod, with a power low enough not to degrade the signal-to-noise ratio, in parallel with the reception of main analog signals ANA_p of interest.
[0171] The calibration method 200 according to the invention therefore allows for precise calibration, easy to implement from a hardware point of view, and does not limit the availability of the receiver 10 for receiving signals of interest.
[0172] A second embodiment of the receiver 10 is shown in [Fig.3].
[0173] This second embodiment differs from the first embodiment in that the The main analog signal ANA_p is brought back to baseband before analog-to-digital conversion in the analog-to-digital converter 35.
[0174] In this case, the baseband 90 conversion stage is placed upstream of the analog-to-digital converter 35. The baseband 90 conversion stage then comprises a 90r in-phase channel and a 90im in-quadrature channel, as well as a respective analog-to-digital converter (in-phase, 35r and in-quadrature 35im) and a respective equalization filter (in-phase, 85r and in-quadrature 85hn).
[0175] Each channel has its own transfer function and is therefore calibrated independently of the other channels, on the principle described for the first embodiment.
[0176] The indices im and r denote the elements and signals specific to each channel on the [Fig.3],
[0177] In this embodiment, the two channels 90r and 90hn of the baseband conversion stage 90 share, for example, an upstream bandpass filter 90a and a third local oscillator 90d downstream of the upstream bandpass filter 90a.
[0178] The local oscillator 90d is configured to generate an intermediate carrier signal of frequency f_mod equal to the difference of the frequency f_port and an intermediate frequency f_int, of which a first component is supplied to a first multiplier 90r,c of the channel in phase 90r and a second component, in quadrature with the first component, is supplied to a second multiplier 90hn,c of the channel in quadrature 90im.
[0179] The baseband conversion stage 90 further includes a first downstream bandpass filter 90r,b downstream of the first multiplier 90r,c, and a second downstream bandpass filter 90hn,b downstream of the second multiplier 90im,c.
[0180] In this case, the intermediate frequency f_int is preferably low, in particular zero. In the example of [Fig.3], for which the analog receiving chain 25 does not include an intermediate frequency conversion stage 65 before the baseband conversion stage 90, the intermediate frequency f_int is preferably zero.
[0181] In a third embodiment not shown, the receiver 10 includes several conversion chains 12. This is particularly the case if the receiver 10 includes several antennas 22 or an antenna 22 comprising several antenna elements, such as a CRPA antenna.
[0182] In this case, each of the conversion chains 12 can be configured according to one of the embodiments described above.
[0183] In a particular embodiment, the sequence generation subunit 100 is common to several of said conversion chains 12, for example to all of these conversion chains 12.
Claims
Demands
1. A calibration method (200) for a conversion chain (12) of an analog signal receiver (10), the conversion chain (12) being configured to provide on a main output (15) a main digital output signal (NUM_p) from a main analog signal (ANA_p) received on a main input (20), the calibration method (200) comprising: a) the generation (205) of a pseudo-random sequence (S) of states with values in a set of discrete values; b) the formation (210) of a frequency-modulated analog calibration signal (ANA_cal,mod) from the pseudo-random sequence (S); c) the injection (215) of the modulated analog calibration signal (ANA_cal,mod) into the conversion chain (12);d) the frequency demodulation (220) of an intermediate digital signal (NUM_int) received at an input (70) of the conversion chain (12) in response to at least the modulated analog calibration signal (ANA_cal,mod), to form a digital signal to be processed (NUM_BB); e) the determination (225) of a correlation function R(NUM_BB,S) between the digital signal to be processed (NUM,BB) and the pseudo-random sequence (S); f) the determination (230) of a transfer function (H) of the conversion chain from the correlation function R(NUM_BB,S); and g) if necessary, the determination (235) of at least one parameter of a processing step to be applied downstream of the main output (15) on the basis of the transfer function (H).
2. Calibration method (200) according to claim 1, wherein said processing step comprises a filtering step by an equalization filter, the determination of the processing step comprising the determination of at least one coefficient (Ci) of the equalization filter.
3. Calibration method (200) according to claim 2, wherein the determination of at least one coefficient (Ci) of the equalization filter comprises the calculation of a plurality of coefficients (Ci) of the equalization filter from the transfer function (H), this calculation including minimizing a distance between the transfer function (H) and a predetermined ideal transfer function (H_th).
4. Calibration method (200) according to any one of the preceding claims, wherein a first modulation frequency (fl) by which the analog calibration signal (ANA_cal,mod) is modulated is included in a receiver bandwidth (10).
5. Calibration method (200) according to any one of the preceding claims, wherein a characteristic time (T'e) separating two successive states of the pseudo-random sequence (S) is chosen as a function of a sampling frequency (Fe) of the conversion chain (12).
6. Calibration method (200) according to any one of the preceding claims, wherein the determination of the transfer function (H) includes the calculation of a Fourier transform of the correlation function (R(CAL_f,S)).
7. Calibration method (200) according to any one of the preceding claims, wherein the determination of the correlation function (R(CAL_f,S)) includes the subsampling of the pseudo-random sequence (S) and of the digital signal to be processed (NUM_BB).
8. Calibration method (200) according to any one of the preceding claims, wherein the calibration method (200) is implemented simultaneously with the reception of the main analog signal (ANA_p).
9. Analog signal receiver (10) comprising at least one conversion chain (12) configured to provide on a main output (15) a main digital output signal (NUM_p) from a main analog signal (ANA_p) received on a main input (20), configured to implement the calibration method (200) according to any one of the preceding claims.
10. Receiver (10) according to claim 9, wherein the conversion chain (12) comprises: i) an analog receiver chain (25) comprising: * the main input (20), * an analog-to-digital converter (35) downstream of the main input (20), and
11. * a first calibration unit (40), comprising a digital-to-analog conversion subunit (45) and a modulation subunit (50), the first calibration unit (40) being configured to inject a modulated analog calibration signal (ANA_cal,mod), modulated by a first modulation frequency (fl), into the conversion chain (12) upstream of the analog-to-digital converter (35), ii) a digital processing module (30) comprising: * a digital signal processing unit (75) downstream of the analog-to-digital converter (35), the processing unit (75) comprising the main output (15) and a filtering stage (85) comprising at least one reconfigurable equalization filter upstream of the main output (15), and * a second calibration unit (80) comprising: - a sequence generation subunit (100), configured to generate a pseudo-random sequence (S) of discrete states with values in a set of discrete values and to provide the first calibration unit (40) with an incoming digital calibration signal (NUM_cal,ent) formed from said pseudo-random sequence (S) for the formation of the modulated analog calibration signal (ANA_cal,mod), and - a computing subunit (105), configured for: a) determine a correlation function (R(NUM_BB,S)) between a digital signal to be processed (NUM_BB) received from the processing unit (75) and the pseudo-random sequence (S), the digital signal to be processed (NUM_BB) being supplied to the calculation subunit (105) downstream of the digital-to-analog converter (35) in response to at least said modulated analog calibration signal (ANA_cal,mod), b) determine a transfer function (H) of the conversion chain (12) from the correlation function (R(NUM_BB,S)), c) determine at least one coefficient (Ci) of the equalization filter of the filtering stage (85) on the basis of the transfer function (H), and d) transmit at least one coefficient (Ci) to the processing unit (75) for reconfiguration of the filtering stage (85). Receiver according to claim 10, comprising a controlled radiation pattern (CRPA) antenna (22) upstream of a plurality of respective main inputs (20) of a plurality of conversion chains (12), the sequence generation subunit (100) being common to several of said conversion chains (12).
Citation Information
Patent Citations
Satellite system calibration in active operational channels
US10284308B1
Calibration system, antenna system and calibration method
US20210281335A1