FR2858736A1 - Useful wireless signal filtering method for mobile telephone, involves initializing filter using initial coefficients, writing updated coefficients in volatile memory of RF module, and calculating updating of coefficients - Google Patents

Abstract

The method involves initializing an adaptive filter using initial coefficients. The updated coefficients are written in a volatile memory of an RF module, read and stored in a non volatile memory of a base band module. The stored coefficients are written in the volatile memory. The RF module is reactivated and useful signals are received. Updating of the coefficients is calculated, from the written coefficients. An independent claim is also included for a wireless device.

Description

i 1

  A method of filtering a radio communication signal embodying a calibration of an adaptive system and corresponding radio communication device.

  FIELD OF THE DISCLOSURE 1.1 General Field The field of the invention is that of radiocommunications. More specifically, the invention relates to a digital filtering technique of a useful radiocommunication signal.

  1.2 Particular Domain of Low Frequency Digital Receivers The invention applies in particular, but not exclusively, to the field of low frequency digital receivers, which are conventionally used in the field of radiocommunications.

  The general architecture of such digital receivers is illustrated in FIG. 1. This so-called Weaver architecture comprises an antenna 10 on which the signal is received and a low noise amplifier 11 (or LNA for Low Noise Amplifier). , intended to compensate for the attenuation suffered by the signal in the path of its radio transmission. It also includes a quadrature mixer, which converts the received signal at the RF frequency into an intermediate frequency IF (for Intermediate Frequency), by means of a local oscillator LO (for Local Oscillator) 13.

  The frequency-converted signal then passes through an AGC (Automatic Gain Controller) analog gain control device 14, an LPF (low pass filter) low pass filter 15, and finally an ADC analog-to-digital converter. (for Analog-to-Digital Converter) 16.

  The two components in quadrature I and Q of the signal feed a processor of the DSP (for Digital Signal Processor) type 17.

  For proper operation of such a digital receiver, it is necessary to obtain a certain signal-to-noise ratio. However, in digital receivers of the type LIF (for Low Intermediate Frequency or NZIF (for Near Zero Intermediate Frequency, or intermediate frequency close to zero), the IF intermediate frequency is generally close to the spacing of the channel , especially for GSM 5 technologies (Global System for Mobile Communications). Sufficient image rejection therefore requires complex analog filtering or dedicated digital signal processing.

  In fact, in view of the low-intermediate-frequency architecture illustrated in FIG. 1, the imperfect phase and amplitude adaptation of the I and Q channels constitutes a major problem, which results in insufficient rejection of the frequency band. image, and therefore in the appearance of an interfering signal harmful to a good reception of the useful signal, as explained below.

  Infinite image rejection can be theoretically obtained for low frequency conversion in which the two quadrature components I and Q of the signal are perfectly balanced. However, the analog components constituting a radiocommunication receiver have finite tolerances, which introduce an imbalance between the phases and between the amplitudes of the analog signal. As a result, the analog frequency converter mixes the wanted signal and the image signal, resulting in a finite image rejection.

  Although a careful design of the analog part of the receiver of FIG. 1 makes it possible to obtain an attenuation of the image signal of the order of 35 dB, it is however necessary to carry out an efficient compensation of the imbalance of the I and Q components. Indeed, depending on the type of radio network considered (GSM, GPRS, EDGE, etc.) and the value of the intermediate frequency, the image signal can be 40 to 90 dB more powerful than the useful signal.

  Thus, for a GSM type network for example, the antenna 10 receives a channel centered on an RF frequency (typically of the order of a few MHz), with a width of approximately 200 kHz. In the worst case specified by the GSM standard, the adjacent channels are of higher power than the RF-centered channel.

  This can be modeled mathematically by a quadrature mixer with an unbalanced local oscillator signal, which can be written as: XLo (t) = cos (oLot) - jgsin (coLot + qm) where g represents the imbalance in amplitude, q the imbalance in phase, and where o = 2zrf.

  This relation can also be expressed in the form: ## EQU1 ## where the unbalance coefficients K.sub.o and K.sub.2 are given by: (l + gj (1gejr) K = and K2 = 2 2 L The attenuation of the image obtained by analog quadrature frequency conversion is then defined as: IRRANALOG = K2 12 By defining z (t) as the baseband equivalent of the frequency band of interest, containing the wanted signal and the image signal, the signal received on the antenna 10 of the receiver of FIG. 1 is: rRF (t) = 29 {z (t) ej '} = z (t) eJ' LO '+ z * (t) e-joLot The complex analog signal after quadrature conversion at intermediate frequency and low pass filtering can be expressed as rlF (t) = Klz (t) + K2z * (t) The second term of this latter equation results from the existence of imbalances and represents the image (namely the image channel of the channel referenced 22 in FIG. 2a) which interferes with the useful signal 21, as illustrated in FIGS. 2a (spectrum of the received signal rRF (t) at the RF stage) and 2b (spectrum of the received signal rIF (t) at the intermediate frequency stage IF).

  These figures 2a and 2b illustrate the case of a low intermediate frequency receiver operating on a radio communication network of GSM, GPRS or EDGE type; the value of the IF intermediate frequency is 200 kHz, so that the interfering signal is the bi-adjacent channel 22, which can be up to 41 dB more powerful than the useful signal 21. The maximum performance that can reach the mixers Since the best-performing analogs are conventionally of the order of 35 dB, the interfering signal 22 is therefore approximately 6 dB above the useful signal 21. It is generally accepted that, in order to be able to achieve satisfactory performances during decoding of the wanted signal, it is necessary that the interfering signals superimposed on the wanted signal have a power of at least 9 dB less than that of the wanted signal.

  The image rejection obtained by the structure of the mixer is therefore not sufficient, and it is necessary to correct the defects of the receiver of FIG. 1 to cancel the interference existing between the useful and bi-adjacent channels 22.

  To do this, the low-intermediate-frequency receivers as illustrated in FIG. 1 must integrate an interference cancellation system, either in the analog domain or in the digital domain.

  2. Methods of the Prior Art and Drawbacks of these Methods There are to date several known techniques for improving image rejection in low-intermediate-frequency receivers.

  Among these various methods, some, which constitute the closest prior art of the present invention, are numerical methods which rely on adaptive cancellation of the interference.

  Thus, Valkama and Renfors in Advanced Methods for I / Q Imbalance Compensation in Communication Receivers (IEEE Transactions on Signal Processing, Vol 49, No. 10, October 2001) have proposed an adaptive interference cancellation algorithm, according to wherein the term interfering is estimated by adaptive filtering 30 of a reference signal 31, as illustrated in Figure 3. The resulting estimate 32 of the interfering signal, obtained by filtering, is then subtracted from the observation of the useful channel 33 , to deduce a good estimate of the useful signal alone.

  In order to calculate an accurate estimate 32 of the interfering signal, the reference signal 31 must be highly correlated to the interfering, and ideally, non-correlated to the useful signal 33.

  The adaptation of the coefficients of the filter is performed from the estimation error 34, which makes it possible to calculate an update of the filter coefficients.

  In a conventional radiocommunication device, of the mobile phone type, for example, these filtering coefficients are used in the part of the device operating in radiofrequency, which realizes the reception of the useful signals. Recall that such a device conventionally comprises a portion operating in radio frequency (RF), and a portion operating in baseband.

  These two parts of the device may be in the form of components or specific modules.

  Upon power-up of the RF portion, the adaptive filter is initialized using unit coefficients, which are progressively updated until the optimum convergence coefficients of the filter are reached.

  A disadvantage of this prior art technique is that the convergence time of the adaptive filter can be long, so that the filtering of useful signals received by the RF portion before the filter converges to its optimum, is of poor quality. More precisely, as long as the coefficients of the adaptive filter have not converged towards their optimum value, the image rejection made by the radiocommunication device is poor, and the useful signals received are therefore strongly disturbed by the interfering signals.

  This problem arises in particular when the radio frequency portion of the radiocommunication device is first switched on at the time of shipment from the factory. It also arises whenever the RF module, in standby mode, must be reactivated to receive new useful signals.

  3. OBJECTIVES OF THE INVENTION The object of the invention is notably to overcome these disadvantages of the prior art.

  More specifically, an object of the invention is to provide a filtering technique of a useful radiocommunication signal making it possible to improve the quality of reception of the useful signals compared to the techniques of the prior art.

  Another objective of the invention is to implement such a technique which makes it possible to increase the convergence speed of the filter coefficients of a digital adaptive filter.

  The invention also aims to provide such a technique that is simple to implement and inexpensive to implement.

  Another object of the invention is to implement such a technique which makes it possible to adapt the filtering of the useful signal to variations in temperature or operating conditions of a receiver of this signal.

  Another object of the invention is to provide such a technique which makes it possible to improve the image rejection compared to the techniques of the prior art.

  Yet another object of the invention is to implement such a technique which consumes little in terms of resources and which, in particular, does not reduce the autonomy of a radiocommunication device with respect to the techniques of the prior art.

  4. Main features of the invention These objectives, as well as others which will appear subsequently, are achieved by means of a method of filtering a useful radiocommunication signal.

  According to the invention, such a method implements a digital adaptive filter, and comprises a step of calculating coefficients of said digital adaptive filter, delivering a periodic update of said coefficients. The initialization of said calculation step takes into account a set of coefficients specific to each radiocommunication device, previously stored in a memory provided for this purpose.

  Thus, the invention is based on a completely new and inventive approach to adaptive digital filtering of a useful radio signal. Indeed, the invention proposes to initialize the step of calculating the optimum filter coefficients, not from unit coefficients as provided in the prior art, but from a set of specific coefficients, adapted to the radiocommunication device considered, which has been pre-registered. The convergence time of the adaptive filter is thus considerably increased, since the starting coefficients are much closer to the optimum filter coefficients; in this way, even the signals received immediately after powering on the device, while the filter has not yet converged to its optimum value, are received with satisfactory quality.

  This method applies both to the first power on of the radio communication device, at the factory outlet, and to each of the activations of the device after a more or less prolonged period of standby. Thus, as will be seen in the rest of the document, this set of specific coefficients may for example have been calculated on the production line of the device, and depends on its intrinsic operating characteristics, or may have been calculated during the last phase of receiving useful signals by the device, and then depends on the reception conditions at a given time and in a given area.

  It is recalled that these filtering coefficients are likely to vary as a function of the temperature, but also as a function of the frequency band in which the radiocommunication device in question operates.

  Advantageously, such a method comprises a storage step in a memory belonging to a first component and / or radio frequency (RF) processing module, performing said calculation step, and a step of reading said memory by a second component and / or baseband processing module.

  This allows the baseband portion of the radio communication device to have read access to the volatile memory of the RF module. In other words, the baseband portion reads the last coefficients calculated for the adaptive filter in the volatile memory of the RF part in which they are stored. The baseband portion can then store the read coefficients in a FLASH memory, or any other nonvolatile memory of the baseband portion. This avoids the coefficients, stored in a volatile memory of the RF part, are lost during the deactivation and power off of this signal receiving module.

  Preferably, such a method comprises a step of writing by said second component and / or baseband processing module of specific coefficients in said memory.

  So we give a write access to the baseband portion, which can write, in the volatile memory of the RF part, the coefficients it has previously stored in its Flash memory. The baseband portion therefore plays a role of temporarily storing the filter coefficients, which it then returns to the RF module.

  Advantageously, said writing step is carried out at least before receiving said useful radiocommunication signal.

  Thus, after the RF module has been powered back for receiving a new wanted signal, and before the reception of a new RF signal begins, the baseband portion is written to the transceiver the last time. filter coefficients used by the RF module. The reception conditions being of slow variation, it is therefore very likely that the update of the filter coefficients is initialized using coefficients close to the optimal convergence coefficients.

  According to an advantageous characteristic of the invention, said reading step is implemented when said filter coefficients have reached a predetermined convergence threshold and / or a predetermined number of calculation iterations has been performed.

  Advantageously, said reading step is carried out at least before switching off and / or putting in sleep mode said radio communication device implementing said method.

  This prevents the last coefficients calculated from being erased from the volatile memory of the RF part when it is switched off. On the contrary, they are provisionally saved in the baseband portion.

  Advantageously, said storage, reading and writing steps also apply to a set of specific coefficients for each frequency subband in which said radiocommunication device can operate.

  Thus, when the radiocommunication device is for example a tri-band mobile telephone, a set of specific coefficients is stored for each of the three frequency bands in which the device can operate.

  Preferably, said storage, reading and writing steps 10 also apply, together with said specific coefficients, to at least one of the information belonging to the group comprising: interrupting thresholds for calculating the filter coefficients; - useful signal strengths and received image signals; a convergence step of said calculation step.

  Advantageously, such a method comprises a step of determining specific initial coefficients, implemented during the production control of said radiocommunication device.

  As indicated previously in this document, the invention therefore applies both to the coefficients calculated during the operation of the RF module 20 and the initialization coefficients calculated at the factory, at the end of the assembly line. These coefficients are the coefficients of the adaptive filter corresponding to the typical mismatch (at room temperature) of the radio receiver.

  Preferably, said initial coefficients are determined by means of a tester sending a test signal, analyzed during a predetermined convergence time.

  Thus, while, according to the techniques of the prior art, the calibration of the adaptive filter of the digital receiver was achieved by means of two testers, respectively sending a useful signal and an image signal, which must then be mixed (which is long and expensive), the invention proposes to use a single tester and to advantageously exploit the baseband module, which is programmed for reception on the frequency of the wanted signal.

  The convergence time is about 100 ts.

  Advantageously, the frequency of said test signal is substantially equal to that of an interfering signal capable of disturbing said useful signal.

  According to an advantageous characteristic, the ratio of the powers of said useful and test signals is greater than a predetermined threshold.

  This threshold is for example between -30 dB and 0 dB. The convergence of the adaptive filter to optimal filter coefficients is indeed possible only if the interfering signal, and therefore the test signal which simulates it, is sufficiently powerful compared to the useful signal.

  More precisely, the choice of the transmit frequencies of the tester (that is to say the frequency of the interfering channel) and of listening of the receiver (that is to say the frequency of the useful channel) is such that the power received on the useful channel corresponds to the noise floor of the receiver. Thus, the transmit power needed on the tester must be at least 0 to 30 dB greater than the noise floor of the radio module.

  Preferably, the frequency of said test signal is the image frequency of that of said useful signal.

  Advantageously, said digital adaptive filtering is a filtering of said digitized useful signal after intermediate frequency transposition, and said useful signal is centered on said intermediate frequency and said test signal is centered on the opposite of said intermediate frequency.

  Sif is the frequency of the tester and IF the intermediate frequency, the obtaining of the initial specific coefficients is thus obtained, on the production line, 25 by programming the chip in baseband for a reception at the frequency f + 2IF. With a neighboring architecture, the reception can also be done at the frequency f-2IF.

  According to another characteristic of the invention, said intermediate frequency is substantially equal to the frequency width of said useful signal.

  The invention also relates to a radiocommunication device 11 comprising means for processing a useful radiocommunication signal.

  According to the invention, said processing means comprise a digital adaptive filter, and said device comprises means for calculating coefficients of said digital adaptive filter, delivering a periodic update of said coefficients. The initialization of said calculating means takes into account a set of coefficients specific to said radiocommunication device, previously stored in a memory provided for this purpose.

  Advantageously, said processing means are digital processing means implementing a Weaver type radio architecture. 5. List of Figures Other features and advantages of the invention will appear more clearly on reading the following description of a preferred embodiment, given as a simple illustrative and nonlimiting example, and the accompanying drawings. among which: - Figure 1, already commented in relation to the prior art, presents the general architecture of a low frequency intermediate frequency digital receiver; FIGS. 2A and 2B, already commented on in relation to the prior art, illustrate the spectrum of a signal received by the receiver of FIG. 1, at the RF stage on the one hand, and on the IF stage; on the other hand; FIG. 3, also described in relation to the prior art, presents the general principle of the adaptive cancellation method of the interference implemented by the present invention; FIG. 4 illustrates the baseband equivalent model of the defects of the receiver of FIG. 1 and the image rejection system implemented by the present invention; FIG. 5 is a block diagram of the various steps implemented by the adaptive filtering method of the invention comprising an initialization mechanism for calculating the filter coefficients from 30 pre-recorded specific coefficients; - Figure 6 illustrates more precisely the calibration of the adaptive filter made in the factory, at the output of the production line.

  6. DESCRIPTION OF AN EMBODIMENT OF THE INVENTION The general principle of the invention is based on the initialization of the step of calculating the coefficients of a digital adaptive filter by means of a set of coefficients specific to the invention. considered radiocommunication device, which have been pre-recorded.

  For the sake of simplification, throughout the rest of this document, a particular embodiment of the invention is described in the context of a low frequency intermediate digital receiver. Those skilled in the art will easily extend this teaching to the more general case of any digital filtering of a useful signal, whatever its frequency (RF, baseband, etc.).

  FIG. 4 shows a modeling of the defects due to the I / Q imbalance of a digital receiver of FIG. 1, and of the adaptive filtering image rejection system implemented according to the invention.

  A low intermediate frequency digital receiver according to the invention advantageously implements an adaptive interference cancellation algorithm, of the type proposed by Valkama and Renfors, mentioned above. At the intermediate frequency IF stage, the observation of the useful channel 33 is that of the channel centered on the positive intermediate frequency + IF. This channel corresponds to the combination of the useful signal 42 and the image signal, which is superimposed on it (see FIG. 2B) during the intermediate frequency transposition.

  In order to avoid having to use a test signal (or test tone), the reference signal 31 is the channel centered on the negative intermediate frequency-IF.

  As illustrated in FIG. 2B, this channel is mainly composed of the interfering signal 41 at full power (that is to say, not attenuated by the analog mixer), provided that the bi-adjacent is sufficiently more powerful than the useful signal 42. (Note that, because of the particular choice of intermediate frequency in this embodiment, namely 200 kHz, the interfering signal is the bi-adjacent of the wanted signal With another intermediate frequency value the signal interfering could for example be the signal adjacent to the useful signal).

  The interference cancellation system of the invention operates on the baseband channels. The two channels in + IF and -IF are both converted to baseband, and undergo low-pass filtering, in order to generate observation of the wanted channel 33 and the conjugate complex of the reference channel 31.

  By subtracting the interfering in-IF to the observation in + IF, we find the useful signal.

  The present invention makes it possible to increase the speed of convergence of the filter 40, by initializing the calculation of the update of its coefficients, not from unit or arbitrary coefficients, but from coefficients specific to the radiocommunication device under consideration. pre-recorded.

  This mechanism is explained in more detail in FIG.

  As explained later in connection with FIG. 6, initial coefficients specific to the radiocommunication device are calculated at the factory, on the production line of the device.

  During the first commissioning 50 of the radio communication device, the initialization 51 of the adaptive filter is carried out using the initial specific coefficients, corresponding to the receiver's typical mismatch, calculated at the factory. To do this, the baseband module writes these initial specific coefficients to a volatile memory or a register of the RF module that it has previously stored in a non-volatile memory, of the Flash type, for example.

  During a step referenced 52, the RF module proceeds to calculate updates of the filter coefficients, and receives useful signals, which are filtered from the calculated filter coefficients. These updated filter coefficients are stored in a volatile memory of the RF module, and are therefore accessible until the RF module is powered down.

  During a step referenced 53, the baseband module reads in this volatile memory the filtering coefficients stored therein, for storing them, for example in flash, in one of its nonvolatile memories, where they can be kept when disabling the RF module.

  This step referenced 53 is for example implemented before the power off of the RF module, between two phases of receiving successive signals. It can also be implemented at regular time intervals by the baseband module, or when the filter coefficients have reached a certain convergence threshold, or have undergone a number of calculation iterations.

  During this step referenced 53, the baseband module can also read, in a register of the RF module, the estimated power of the useful and interfering signals received by the RF module, the convergence pitch of the adaptive filter, as well as a threshold for triggering the adaptation of the filter. Indeed, for a good functioning of the adaptive filtering algorithm of the invention, the hypothesis mentioned above in connection with FIG. 4, according to which the -IF image is almost entirely composed of the bi-adjacent channel 22. to the useful signal 21 must be verified. However, this hypothesis is verified only if the biadjacent 22 constituting the interfering signal 41 is sufficiently powerful compared to the useful signal 21, 42.

  According to the invention, therefore, any updating of the coefficients of the adaptive filter 30 is stopped when the interfering signal 41 is no longer sufficiently powerful with respect to the useful signal 42, so as to maintain the convergence of the filter 30 towards correct values.

  The ratio of the powers of the interfering and useful signals is for example between -30 dB and 0 dB. This threshold is stored in a volatile memory of the RF module.

  The updating of the filter coefficients can also be the subject of a flexible decision, for example by playing on the convergence step of the adaptive algorithm: this step can therefore also be the subject of the reading step 53.

  At the time of reactivation of the RF module, a step 54 is carried out, during which the baseband module writes to the RF module the last filter coefficients that it has memorized during the referenced step. 53. These coefficients, which are written by the baseband module, are used to initialize the calculation of new filter coefficient updates by the RF module, which can filter the received signals from coefficients converging rapidly to optimal filtering.

  The initial calibration of the adaptive filter, made in the factory, is now presented in greater detail in connection with FIG.

  This post-production calibration method consists in injecting into the RF (or transceiver) module 60, during a reception operation of a useful signal, an interfering signal (in the present example, a bi-adjacent signal 15 to the useful signal) sufficiently powerful to be able to adapt the image rejection filter.

  For this purpose, the RF module 60 is connected to a tester 61 and to a baseband module 62. The baseband module 62 programs a reception on a determined frequency 64. while the tester 20 is configured to emit a signal at the frequency 63 image of the frequency f, in this case f + 2IF, where IF is the intermediate frequency in which the useful signal is transposed. Indeed, in the particular example set forth herein, the signal interferes with the wanted signal is the bi-adjacent signal.

  From the useful signals 64 and test 63 that it receives, the RF module 60 can proceed to the calibration of the adaptive filter, by calculating successive updates of the filter coefficients, so as to converge the filter. These initial coefficients calculated by the calculation block 65 correspond to the typical mismatch, at ambient temperature, of the RF module. They are stored, as long as the RF module 60 is powered on, in a volatile memory 66.

  At the end of this post-production calibration phase, the baseband module 62 reads (67) the coefficients stored in the volatile memory 66 of the RF module, to write them (68) in one of its memories non-volatile 69, where they remain stored as long as the RF module is idle.

  At the time of the first power-up of the RF module 60, the baseband module 62 writes the coefficients stored in the volatile memory 66 of the RF module 60, so that they can initialize the calculation 65 new updates of the filter coefficients.

  In this way, during the first commissioning of the RF module 60, the convergence of the adaptive filter is reached very rapidly, which makes it possible to increase the quality of reception of the useful signals.

Claims (17)

  1. A method of filtering a useful radiocommunication signal, characterized in that it implements a digital adaptive filter, in that it comprises a step of calculating coefficients of said digital adaptive filter, delivering a periodic update said coefficients, and in that the initialization of said calculation step takes into account a set of coefficients specific to each radiocommunication device, previously stored in a memory provided for this purpose.   2. A filtering method according to claim 1, characterized in that it comprises a storage step in a memory belonging to a first component and / or radio frequency (RF) processing module, performing said calculation step, and a step reading said memory by a second component and / or baseband processing module.   3. A filtering method according to claim 2, characterized in that it comprises a step of writing by said second component and / or baseband processing module of specific coefficients in said memory.   4. The filtering method as claimed in claim 3, characterized in that said writing step is carried out at least before receiving said useful radiocommunication signal.   5. The filtering method as claimed in claim 2, wherein said reading step is implemented when said filtering coefficients have reached a predetermined convergence threshold and / or a predetermined number of times. iterations of calculation was performed.   6. A filtering method according to any one of claims 2 to 5, characterized in that said reading step is implemented at least before switching off and / or in standby mode said device radiocommunication implementing said method.   A filtering method according to any of claims 1 to 6, characterized in that said storing, reading and writing steps also apply to a set of specific coefficients for each frequency subband in which said radiocommunication device is operable.   8. A filtering method according to any one of claims 1 to 7, characterized in that said storage, reading and writing steps also apply, together with said specific coefficients, to at least one of the information belonging to the group. including: - interruption thresholds for calculating the filter coefficients; - useful signal strengths and received image signals; a convergence step of said calculation step.   9. A filtering method according to any one of claims 1 to 8, characterized in that it comprises a step of determining specific initial coefficients, implemented during the manufacturing control of said radio communication device.   10. A filtering method according to claim 9, characterized in that said initial coefficients are determined by means of a tester sending a test signal, analyzed during a predetermined convergence time.   11. A filtering method according to claim 10, characterized in that the frequency of said test signal is substantially equal to that of an interfering signal capable of disturbing said useful signal.   12. A filtering method according to claim 11, characterized in that the ratio of the powers of said useful and test signals is greater than a predetermined threshold.   13. A filtering method according to any one of claims 10 to 12, characterized in that the frequency of said test signal is the image frequency of that of said useful signal.   14. A filtering method according to any one of claims 10 to 13, characterized in that said adaptive digital filtering is a filtering said digitized useful signal after transposition in intermediate frequency, and in that said useful signal is centered on said intermediate frequency and said test signal is centered on the opposite of said intermediate frequency.   15. A filtering method according to claim 14, characterized in that said intermediate frequency is substantially equal to the frequency width of said useful signal.   16. A radiocommunication device comprising means for processing a useful radiocommunication signal, characterized in that said processing means comprise a digital adaptive filter, and in that said device comprises means for calculating coefficients of said digital adaptive filter, providing a periodic update of said coefficients, and in that the initialization of said calculating means takes into account a set of coefficients specific to said radiocommunication device, previously stored in a memory provided for this purpose.   17. A radio communication device according to claim 16, characterized in that said processing means are digital processing means implementing a Weaver type radio architecture.