FR2892586A1 - HYBRID RECEIVING METHOD AND DEVICE HAVING A CHANNEL EQUALIZER AND AN INTERFERENCE CANCER - Google Patents

Abstract

This device for receiving an analog signal (r (t)) with a baseband spectrum spreading from a multipath propagation channel, the analog signal conveying symbols, comprises in series: a multi-channel channel equalizer ( C1, Ck) of the analog signal (r (t)) adapted to deliver a first estimate (d1l, d1K.) Of the symbols corresponding to each of the codes; means for regenerating, from the estimated symbols (d1l, d1K), replicas (r1l, r1L) of the analog signal corresponding to the paths; means for regenerating, from the replicas (r1l, r1L), interferences for each of the paths; and means for providing a second estimate (d2l, d2K) of the symbols by canceling the interference of the analog signal (r (t)) for each of the paths.

Description

BACKGROUND OF THE INVENTION The field of the invention is that of

  digital telecommunications. The invention finds particular application in the field of digital radio-frequency communications between a base station and a mobile terminal and in particular in applications that are in line with the evolution of third-generation mobile telephony systems known as "HSDPA" ( High Speed Downlink Packet Access) and defined by the UMTS Forum. The HSDPA principle is based on the rapid adaptation of the link to allocate the majority of the resources to users with favorable channel conditions. This standard allows QPSK and 16QAM modulations, the latter offering superior spectral efficiency. However, 16QAM type modulation is very sensitive to interference and its use requires advanced processing techniques in reception. An advanced processing technique that complies with the HSDPA standard is to use an interference canceller. Interference cancelers (RFCs) are non-linear multi-user receivers with tiered structures. Their operating principle is to regenerate interference using the symbols estimated at the output of the current stage. This interference is then removed from the received signal and the resulting signal constitutes the input of the next stage. The number of stages generally depends on the desired performance and the complexity constraints. FIG. 1 represents a parallel interference canceller 12, namely an interference canceller in which interference cancellation is performed simultaneously for all codes. For more information on this interference canceller 12, those skilled in the art may refer to K. Higuchi, A.

  Fujiwara and M. Sawahachi "Multipath interference canceller for high-speed packet transmission with adaptive modulation and coding in WCDMA forward link", IEEE Journal on Selected Areas in Communications, Vol. 20, No. 2, pages 419-432, February 2002.

  In this interference canceller 12, the first stage comprises a conventional receiver Rake 13.

  In the example described here, the structure of the parallel interference canceller 12 is composed of M stages (M> -2). The parameters in the structure of the interference canceller 12 are

  defined as: ï ,,,, (i): Estimated replica of the transmitted signal along the path (1515L) to the m (1 <m <M) stage. - ¯ ,,,, (i): Input signal of the stage m (1 <m <m) following the path 1 (1 <1 <L), it is given by: Fm, (i) = r (i) ùa ~ r (m!) J (i). J = 1 Jml This signal is obtained by subtracting from the received signal r (i) all the replicas of the signal transmitted according to the different paths of the channel with the exception of the replica corresponding to the path in question 1. a: Weighting factor between 0.5 and 1 to control the error estimation of symbols from one floor to another.

  This parameter changes for each floor in an increasing way to reach a value close to the unit at the top floor. The determination of the optimal a is a problem in itself, its value is determined generally by simulation. d: Vector estimated symbols after a tough decision with

  at = ... d r, where are the vectors of the estimated symbols corresponding to the different codes after a hard decision, K is the number of spreading codes allocated and i.T represents the matrix transposition operation.

  Each stage of the parallel interference canceller 12 consists of two blocks except the last stage 15 which consists of only one block. The first block is the symbol estimation, where the receiver has a Rake structure. The second block is that of the regeneration of the interferences or replicas of the signal transmitted according to the different paths. The interferences following each path are subtracted from the received signal and the resulting signals constitute the inputs of the next stage. The last stage 15 includes only the symbol estimation block. In a manner known to those skilled in the art, the bit error rate (BER) performance of the parallel interference canceller of FIG. 1 improves as the bit rate is increased. number of floors that make up this canceller. Unfortunately, the complexity of this canceller is proportional to this number of stages and it is not possible to determine a priori the appropriate number of stages to obtain a given performance of this canceller. In practice, the number of stages is imposed by production constraints. Moreover, it is known that a phenomenon of saturation of the performances in terms of BER is observed when the level of the interference becomes significant compared to the level of the thermal noise.

  OBJECT AND SUMMARY OF THE INVENTION The present invention relates to a method and a receiving device adapted to overcome the aforementioned drawbacks. For this purpose, the invention relates to a method for receiving a base band spread spectrum analog signal from a multipath propagation channel, the analog signal carrying symbols, which method comprises: a step multi-code equalization of the channel adapted to deliver a first estimate of the symbols corresponding to each of the codes; a step of regenerating, from the estimated symbols, replicas 35 of the analog signal corresponding to the paths; a step for regenerating, from the replicas, interference for each of the paths; and a so-called "interference cancellation step" for providing a second estimate of the symbols by canceling the interference of the analog signal for each of said paths. Correlatively, the invention also provides a device for receiving a baseband spectrum spreading analog signal from a multipath propagation channel, the analog signal conveying symbols, the receiving device having in series: a multi-code channel equalizer of the analog signal adapted to deliver a first estimate of the symbols corresponding to each of the codes; means for regenerating, from the estimated symbols, replicas of the analog signal corresponding to the paths; Means for regenerating, from the replicas, interference for each of the paths; and - means for delivering a second estimate of the symbols by canceling the interference of the analog signal for each of the paths. The invention thus aims, in a general manner, at a hybrid reception method and device consisting mainly of a linear equalizer and an interference canceller. In a known manner, channel equalizers ("Channel Chip Equalizers") are linear receivers consisting mainly of an equalizer filter followed by a correlator corresponding to the code of the user of interest and a decision device. . Unlike the conventional Rake receiver, these channel equalizers take interference into account and do not consider it as inescapable noise. Consequently, when the interference becomes predominant with respect to the level of the thermal noise, the receiving device according to the invention, which takes into account the interference, gives much better performances than those obtained by the parallel interference canceller. 12 described with reference to Figure 1, the first stage symbol estimation block is a Rake. More specifically, since the symbol estimation block of the first stage of the interference canceller 12 is constituted by a Rake, the estimation of the symbols, and therefore of the interference, is of very poor quality in the case where interference dominates over thermal noise. These estimation errors propagate along the various stages 14, 15 thus leading to performances much lower than those obtained with the receiving device according to the invention. Preferably, the equalization step is a linear equalization step that minimizes the mean squared error. In particular, this filtering step can use a method of the type "zero forcing" in English (ZF). Preferably, the reception method according to the invention comprises a step for automatically determining an optimum depth used as depth in the linear equalization step, from the impulse response of the channel, and the power of the thermal noise. Thus, in a preferred embodiment, the optimal depth of the linear equalizer is automatically determined from the channel conditions. An equalizer of variable complexity is obtained that considerably reduces the use of the resources of the mobile terminal, in particular under conditions of high thermal noise. In a preferred embodiment, the linear equalization step works at the chip rate. This feature significantly reduces the complexity of the receiving device, compared to a device implementing an equalizer filter operating at the sampling rate (fast pace). Indeed, in such equalizers, the equalization is at a fast pace so that the coefficients must be systematically calculated for each chip.

  For more information on this particular embodiment, one skilled in the art can refer to the document Kari Hooli, "Equalization in WCDMA terminais", Doctoral Thesis, Oulu University, Finland, 2003. In a preferred embodiment of this variant embodiment , the receiving device according to the invention comprises, prior to the step of linear equalization at the chip rate, a step for delivering, from the analog signal, a chip rate signal, this step being performed according to the general principle of a reception step implemented in a receiver type RAKE. Thus, depending on the channel conditions (frequency selectivity and thermal noise power), the depth of the equalizer varies so that this equalizer can be reduced to a RAKE (for a depth equal to one chip) under conditions of loud noise and / or channel selectively frequency selective. This equalization step comprises: a sub-step for delivering, from the analog signal, a chip rate signal; and a filter sub-step with finite depth, working at a chip rate, for processing the chip rate signal. Thus, according to the invention, the finite-depth filtering is performed at the chip rate, which considerably reduces the computational complexity with respect to the devices of the prior art working at the fast rate of sampling. The following notations are recalled: the notation x * denotes the complex conjugate of the scalar x; - MT designates the transpose of the matrix M; and - M "denotes the conjugated transpose of the matrix M. In this document, the following generic notations are used: P: depth of a linear equalizer, and w: radius of the depth of an equalizer with P = 2w + 1. In a preferred variant embodiment, the finite depth filtering step uses a simplified form of equalization matrix GMMSE: G, 1, fc = (HHDHDH + Q Or: F-HHDHDH = (DH) "DH is a Hermitian band Toeplitz matrix, in which: - H is a block diagonal matrix of the complex gains of the channel, - D is a matrix containing time-shifted versions of the Nyquist root discrete shaping pulse channel. the analog signal, - IQN is the identity matrix, and - the variance of the thermal noise In this particularly advantageous variant, the filtering step with finite depth does not require the knowledge of the spreading codes, namely neither the knowledge nor the of the code of spread of interest directly used by the terminal implementing the reception method, nor that of spreading codes of the network in general. Those skilled in the art will understand that this characteristic advantageously makes it possible to limit the flow of information in the communication channel, as well as the resources required for the signal processing by the terminal. In addition, the explanation of the matrix F in the form of a Hermitian band Toeplitz matrix advantageously makes it possible to simplify its calculation, the coefficients of this matrix being obtainable from its only first column. This characteristic will be developed later. Preferably, in order to calculate, in this variant, the elements of the first column of the aforementioned Hermitian band Toeplitz matrix: a convolution sequence is computed between the discrete impulse response of the channel and the discrete shaping pulse in the root of the channel; Nyquist of the analog signal; an autocorrelation sequence of the convolution sequence is calculated for the positive delays of the autocorrelation sequence; and the autocorrelation sequence is sampled at the chip rate. Preferably, to automatically determine the above-mentioned optimum depth: an equalizer filter of a predetermined maximum depth is calculated and, for all the odd intermediate depths less than this maximum depth, a finite depth intermediate equalizer filter 30 of this intermediate depth , and a relative error is calculated between the central element of the intermediate filter and the central element of the maximum depth filter; and for optimum depth is chosen the minimum intermediate depth for which the relative error is less than or equal to a predetermined error threshold.

  Those skilled in the art will understand that the calculation of each of the intermediate equalizer filters requires the inversion of a square matrix whose dimensions correspond to the intermediate depth. If we consider the computation of all the intermediate filters up to a maximum exploration depth PMAX, the number of complex multiplications required is in o (P, a ,,). In order to simplify this calculation, the coefficients of the finite-depth intermediate equalizer filters can be obtained from an intermediate equalizing matrix Gp of the form G 1 = (F + 6,, 1 ,,) where Q; is the variance of the noise, Ip the identity matrix and F, a Hermitian Toeplitz matrix that translates the effect of the discrete shaping channel and pulse into the Nyquist root of the analog signal. In this embodiment, the intermediate equalization matrix Gp can be obtained recursively by the formula ## EQU1 ##, initialized by: ## EQU1 ## with ## EQU1 ## , d ,, and p ,, = + f for p> 2 where f ,, is the first column of the matrix F ,, explained f ,, = [f'f ,,. , f ,, i ',, fp - ,, .., .I] Het .f = f ,. This recursive calculation method advantageously makes it possible to reduce the complexity of the method for automatically determining the depth of the equalizer from c (i) to 7 (i,) complex multiplications. In a first variant embodiment, the maximum depth PMAX for automatic optimal depth determination is chosen such that PMAX = 4W + 1; where W is the half bandwidth of the aforementioned Hermitian band Toeplitz matrix F. In a preferred variant, the maximum depth is chosen such that = 2iV deer ((E / NJ) / 10), aVeC PMAX = 2wMAx + 1, where W is the half bandwidth of said matrix F Toeplitz Hermitian band, Eb the energy per bit, No the monolateral spectral density of the noise, and erf (.) the error function defined by: Gp erf (x) = / ùf exp (ùy-) dv n This second variant advantageously makes it possible to avoid the

  overestimation of the depth obtained by the first variant mentioned above,

  especially under the conditions for which the Eb / No ratio is relatively low.

  This second variant, which takes into account not only the dispersion of the channel but also the power of the thermal noise, therefore considerably reduces the complexity of this automatic determination of depth.

  The above-mentioned reception device can therefore be seen as a hybrid two-stage interference channel / cancellation equalizer receiver. Very advantageously and as will be demonstrated later, the receiving device according to the invention generally requires only two stages to converge, the first stage giving a first estimate of the symbols and the second stage, based on Rake structures, refines this estimate.

  The fact that the receiving device according to the invention converges only after two stages is very important and solves a crucial problem encountered with the interference canceller 12 of FIG. 1 which is that of determining the number of stages necessary for the convergence.

  In an alternative embodiment, the method and the receiving device according to the invention implement at least one additional iteration of the regeneration and interference cancellation steps in order to deliver, from said second estimation of the symbols, at least a third estimate of said symbols.

  A hybrid channel equalizer / interference canceller receiver with more than two stages is thus obtained. This receiver makes it possible to take advantage of the performance of these two types of receivers which are complementary. Thanks to the channel equalizer, we obtain a good estimate of the symbols (respectively of the interference at the level of the first stage). This improvement of the estimate propagates through the various stages of the interference canceller, which allows to obtain much better performances than those of the interference canceller 12 of FIG. The aforementioned reception can be implemented as a program on a programmable component, for example of the DSP type (for "Digital Signal Processor"). Alternatively, the different steps of the receiving method are determined by computer program instructions. Accordingly, the invention also relates to a computer program on an information medium, this program being capable of being implemented in a reception device or more generally in a computer, this program comprising instructions adapted to the implementing the steps of a reception method as described above. This program can use any programming language, and be in the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form, or in any other form desirable shape. The invention also relates to a computer-readable information medium, comprising instructions of a computer program as mentioned above. The information carrier may be any entity or device capable of storing the program. For example, the medium may comprise storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording medium, for example a diskette (floppy disc) or a disk hard. On the other hand, the information medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means. The program according to the invention can be downloaded in particular on an Internet type network. Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the method in question.

  BRIEF DESCRIPTION OF THE DRAWINGS Other features and advantages of the present invention will emerge from the description given below, with reference to the accompanying drawings which illustrate an embodiment having no limiting character. In the figures: FIG. 1, already described, represents a known parallel interference canceller of the prior art; - Figure 2 schematically shows a multi-code channel equalizer 10 used in a receiving device according to the invention in a preferred embodiment; FIG. 3 represents a multi-code receiver of RAKE type known to those skilled in the art; FIG. 4 schematically shows a reception device according to the invention in a preferred embodiment; FIG. 5 represents, in flowchart form, the main steps of a reception method according to the invention in a preferred embodiment; FIG. 6 shows the structure of a Hermitian band Toeplitz matrix used in the calculation of an equalizing filter in a reception method according to the invention; FIG. 7 represents the structure of an intermediate equalization matrix used in the calculation of an equalizing filter in a reception method according to the invention; FIGS. 8 to 11 are figures making it possible to compare the performance of the receiver according to the invention with the receivers of FIGS. 1 and 2.

  DETAILED DESCRIPTION OF THE EMBODIMENT FIG. 2 schematically illustrates a multi-code channel equalizer used in a receiving device according to the invention. It is adapted to receive an analog signal r (t) spread spectrum baseband from a multipath communication channel. This channel equalizer 1 comprises in series a linear equalizer 10 minimizing the mean squared error, a multi-code correlator 20 and decision means 30 as to the value of the symbols conveyed by the signal. This channel equalizer comprises means 121 for automatically determining an optimum depth used as depth in the linear equalizer, from the impulse response of the channel, and the power of the thermal noise. These means 121 for determining are for example constituted by a programmable component adapted to implement the automatic determination step E220 which will be described later with reference to FIG. 5. The signal received in baseband r (t) is all first sampled at the fast pace (sampling rate) Jt = T / s, which consists of taking S samples by chip Tc time. The channel equalizer 1 comprises means 119 adapted to carry out a filtering adapted to the discrete Nyquist root shaping pulse of the sampled signal corresponding to the analog signal r (t). It also comprises, at the output of the filtering means 119 adapted to the shaping, means 112 for correcting the delays z,..., R, according to the different paths and means 113 for sampling the corrected signals at the chip rate Tc. The channel equalizer 1 comprises channel compensating means 114 adapted to multiply the chip rate signal following each path by the complex conjugate h * of the gain of the corresponding channel.

  It also comprises a summator 115 of the chip rhythm signals according to the different paths. This entire chain, up to the adder 115, constitutes means 110 for delivering, from the analog signal r (t) a chip rate signal to the finite depth equalizing filter 120.

  In an alternative embodiment, the linear equalizer operates at a fast sampling rate. For more information on this particular embodiment, the skilled person can refer to the document [Hooli] supra. The signal at the output of the filter 120 is supplied at the input of a correlator 20.

  In this correlator, a multiplier 21 multiplies the chip to chip signal by the complex conjugate of the scrambling code s * to unscramble it. The correlator 20 also comprises a correlator 111 corresponding to each of the codes of interest C * 1 to C * k. The signal at the output of each correlator 111 is provided at the input of decimation means 31 adapted to keep a sample each Q chips, which consists of performing, in analog, sampling at the symbol rate.

  At the output of each decimator 31, the signal (flexible decision of the symbols il ,, ..., to, is supplied at the input of a parallel / series multiplexer 33. The output signal of the multiplexer 33 is supplied to means 30 of decision which mainly comprise, and in known manner, a decision device 32 depending on the type of modulation used to give a hard estimate of the symbols conveyed by the signal.The person skilled in the art will understand that in the case where the depth of the equalizer filter is equal to a single chip, the channel equalizer according to the invention is simply reduced to a multi-code receiver RAKE type whose structure is given in Figure 3. Figure 4 shows a receiving device 2 according to the invention. invention in a preferred embodiment In this preferred embodiment, the receiving device 25 has two stages ET1 and ET2 The first stage ET1 mainly comprises two blocks, namely a block oc ET11 estimation of symbols and an ET12 block of regeneration of interference. The block ET11 is constituted by the multi-code channel equalizer 1 30 previously described with reference to FIG. 2. For the block E12 for regeneration of interferences, one finds practically the same operations as on the level of the transmission chain ( for example a base station). Indeed, there are successively: means 111 'for spreading the estimated symbols dA respectively by the codes C1 to Ck; an adder 115 '; a multiplier 116 'adapted to apply the scrambling code by multiplying chip to chip the output signal of the adder 115' by the scrambling code S; means 119 'adapted to effect filtering at the Nyquist root discrete shaping pulse; means 114 'adapted to weight the signal after shaping by the coefficient of the corresponding channel hi to hL. means 112 'adapted to introduce a delay ii to TL.

  At the output of this regeneration block ET12 we obtain the signals (t), ..., @, (i) which are the estimated replicas of the signal transmitted according to the different paths of the channel. In the example described here, the receiving device 2 according to the invention comprises only two stages.

  In this embodiment, the second stage ET2 is the last stage. It is composed of a single block, namely a symbol estimation block equivalent or similar to the last stage 15 of the parallel interference canceller 12 described above with reference to FIG. 1. FIG. of a reception method according to the invention in a preferred embodiment. This embodiment method may for example be implemented by the reception device described above with reference to FIG. 4. In a known manner, the matrix GMMSE for channel equalization of an MMSE type equalizer can be expressed by the following: v G. ~ sis, = (HHD "DH + a, ~ (CA2C") - 'with the notation: 30: Body of the complexes L: Number of paths of the Q channel: Spreading factor HE: Matrix diagonal by block of the complex gains of the channel DE: Matrix containing delayed versions with the delays of the channel 35 of the discrete shaping pulse (_R is the body of the real ones) c E `": Matrix of the codes (spreading and scrambling ) A e: Diagonal matrix of the amplitudes of the different codes o: Variance of the noise, N: Size of the received signal in samples .v: Number of symbols transmitted by code x: Number of spreading codes.

  According to the preceding equation, one realizes that the calculation of GMMSE requires the knowledge of all the active codes and the inversion of the matrix (CAC ") .This matrix (cA, C") having no particular structure his inversion is very expensive. In order to get rid of this matrix inversion as well as the knowledge of the active codes, we introduce the following approximation: CA-C "= IQ ,, where 1 ~, is an identity matrix (QN x QN).

  Thus, the expression of the simplified matrix of MMSE equalization used in the rest of the description is given by: G, mis /. = (H "D" DH + Q ,, I (\ i (1) where F E '' is defined by:

  This matrix F, the structure of which is shown in FIG. 6, is advantageously a Hermitian band Toeplitz matrix.The reception method according to the invention described here comprises a reception step E10. of the base band spread spectrum analog signal r (t) from a communication channel This reception step E10 is followed by a channel equalization step E20 comprising, in this preferred embodiment, four main steps E210, E220, E230 and E240 for respectively delivering a chip rate signal, calculating an optimum depth and calculating an equalization matrix, and performing a finite depth filtering operation The first step E210 of step d Equalization E20 makes it possible to deliver, from the analog signal r (t), a signal at the chip rate, This step may for example be implemented by the means 110 described above with reference to FIG. 2. The method of implementation described here comprises a second step E230 for calculating the elements of the Hermitian band Toeplitz F matrix. As described above, the matrix GMMSE is expressed using the following formulas (1) and (2): G.i111 = (H ID "DH + Q, ~ 1un J (1) FH" D "DH = ( DH) "DH (2) In equation (2), F is a Hermitian band Toeplitz matrix and its construction only requires the knowledge of its first column.The calculation of this first column can be carried out preferentially in three sub-groups. Steps E232, E234 and E236 During a first substep E232, a convolution sequence v (i) h (i) * (i) (6) where h (i) is the discrete impulse response is calculated of the channel and `P (i) is the discrete shaping pulse This first substep E232 is followed by a second substep E234 in which an autocorrelation sequence is calculated for positive delays ( n) with n> 0 of y (i) This second substep is followed by a third substep E236 in which said autocorrelation sequence R n (n) is sampled at the rate me chip, which amounts to keeping a sample each S sample, these samples giving the elements of the first column of F. The second step E230 of calculation of the matrix F is followed, in this preferred mode, by a third step E220 adapted to automatically determine the optimal popt depth that will be used later in step E240 linear equalization. This optimum depth is obtained from the impulse response of the channel h and the power a, -; thermal noise. Preferably, this optimal depth calculation step E220 comprises four successive sub-steps E222, E224, E226 and E228. During this first substep E222, a maximum depth PMAX is determined. In the preferred embodiment described herein in which the GMMSE equalization matrix has the simplified form given to equations (1) and (2), the maximum depth PMAX can be chosen such that: PMAx = 4W + 1 (3) where W is the half bandwidth of the F band Toeplitz Hermitian band. In a preferred variant, the maximum depth PMA may advantageously be chosen such that: ## EQU1 ## where PM (= 2w max + 1 and where W is the half bandwidth of the matrix F Toeplitz Hermitian band, Eb the energy per bit, No the monolateral spectral density of the noise, and erf (.) the function error defined by: erf (x) = jexp (ùy) dv 20 o This first sub step E222 for calculating the maximum depth PM, by one or other of the formulas (3) or (4), is followed by a second substep E224 during which the intermediate depth is calculated for all the intermediate depths odd p less than the maximum depth PM, an intermediate equalizer filter with finite depth of this intermediate depth P. The equalizer filter for the maximum depth PMAX is also calculated In the embodiment described here, the coefficients of this filter equalizer finite depth intermediate g, ;; ', `, are the elements of the line w + 1 of an intermediate equalization matrix Gp of the form C = (F, + o-, 2j i,) where is the variance of the noise, Ip the identity matrix and F, is a Hermitian Toeplitz matrix which translates the effect of the discrete shaping channel and pulse into the Nyquist root of said analog signal.

  The structure of such an intermediate equalization matrix Gp is given in FIG. 7. By its structure, the person skilled in the art will understand that the knowledge of F 1 is summed up by the knowledge of his first column denoted by f which is expressed in the form:

f, = [f, .f f ,,].

  In a preferred embodiment, a relation is established between two equalization matrices G 1 and G 1, of respective depths p and (p 1) with p> 2, or

  G = (F + a I, _, r-1 Let d ,, be the column vector of length (pû1) obtained from the complex conjugate of f ,, deleting the first element (ie, f) and inverting the order from the rest of the elements, that is to say: d ,, = [f ,,, ..., f] ", where [.]" is the Hermitian conjugate, for simplicity in the writing of In considering the fact that F ,, is Hermitian Toeplitz, we can write: 25 F = F, _, d "rd, f Similarly, we can write: By applying the lemma of Partitioned matrix inversion described in Steven M. Kay, Fundamentals of Statistical Signal Processing: Estimate Theory, Prentice Hall, New Jersey, 1993 to, one skilled in the art will understand that the intermediate equalization matrix Gp can therefore be obtained recursively. by the formula G ,,, _, d "G + p brbH bH, initialized by: G, = o -, + f 1 p, with b ,, _ ùp, 'G1,, d ,, and p ,, = ## EQU1 ## where F 1 is the first column of the matrix F ,, Explained f, _ {f, f ,, ..., f ,,] `, d ,, = f ,, f;, _, f;]" and f = f The second substep E224 for calculating finite depth intermediate equalizer filters is followed by a third substep E226 in which a relative error er (between the central element g, t` /, t "'LC / i) is calculated for each odd intermediate depth. (w + t) of the filter é, 1 "',, /. CL' and the central element + t) of the filter g, t4A / .C / i ~ nnix g, {IA15'F. of maximum depth PM. Specifically, 15 (gAL4LCF) (w ,,, Y +1) ùgi.hm (w 1) gAlA1.17 ~ (~ Unr, tr +1) with ti, '= 0,1, .. . (wmr -1) (5) This third substep E226 for calculating the relative errors er ("") is followed by a fourth substep E228 for determining the optimum depth popt which will be used in step E240 filtering with finite depth. For this purpose, it is preferentially chosen for optimal depth popt the minimum intermediate depth for which the relative error er 'is less than or equal to a threshold er of predetermined error. The predetermined threshold is set in advance and depends on the desired performance. This fourth substep E228 completes the step E220 for calculating the optimal depth of the filter. Step E220 for determining the depth is followed by step E240 of linear equalization of the signal delivered to the first step E210 of the equalizing step E20. Those skilled in the art will recognize that this filtering step E240 is carried out at the chip rate. It completes the equalizing step E20 of the channel according to the invention.

  This equalization step E20 is followed by a signal correlation step E30, as previously described, to descramble the signal and correlate it with the codes of interest C * 1,... C * K During this step E30, a decimation operation is also performed during which a sample is kept each Q chip. This decimation step is followed by a hard decision operation to deliver a first estimate of the symbols d ,,., D ~. This step E30 of correlation and hard decision can in particular be implemented by the correlator 20 and the decision device 30 previously described with reference to FIG. 2. The correlation step E30 is followed by a step E40 during which is regenerated from the estimated symbols replicas r "to r" to the analog signal corresponding to each of the paths. This regeneration step E40 can be implemented by the ET12 regeneration block of the first stage ET1 of the receiving device 2 according to the invention. This replica regeneration step E40 is followed by a step E50 in which interference from each of these paths is regenerated from these replicas. The interference following a path I corresponds to the summation of all the replicas according to the different paths except the replica corresponding to the path in question. The step E50 of regeneration of the interference is followed by a step E60 during which a second estimate of the symbols is delivered by canceling the interference of the analog signal for each of the paths. As described previously with reference to FIG. 1, this is obtained by subtracting from the received signal r (i) all the replicas of the signal transmitted according to the different paths of the channel with the exception of the replica corresponding to the paths in question. In a preferred embodiment, the receiving device comprises more than two stages. In this case, the additional stages are in accordance with the intermediate stages 14 previously described with reference to FIG. 1. They thus make it possible to repeat the regeneration and interference cancellation steps in order to refine the estimate at each stage.

  FIGS. 8 to 11 show results for comparing the performance of the receiver 2 according to the invention with the receiver 12 of FIG. 1 and the receiver 1 of FIG. 2. More precisely, the performances in terms of the BER as a function of the Eh / No ratio of the hybrid receptor according to the invention (hereinafter referred to as "MMSEA / MPIC") are compared with those of the receptors 12 (hereinafter referred to as "MPIC") and 1 (hereinafter referred to as "MMSEA"). . The specifications of the HSDPA standard in FDD mode as defined in 3GPP documents TS 25.11, 25.211 and 25. 213 are considered for the simulations. A number of K codes with a spreading factor Q = 16 are allocated to the same mobile terminal. Severe interference conditions are considered using 16QAM type modulation in a Vehicular-A type channel as defined by ETSI TR 101 112. In Figures 8, 9 and 10 are respectively given the performance of the different receivers for K = 5, 10 and 13 codes. For the MMSEA as well as the first stage of the MMSEA / MPIC the threshold of the relative error is set at I o '. The number of stages for the MPIC and MMSEA / MPIC is set at M = 2, 3, 4 and 10. The weighting factor has been set at 1 for all cases. The evolution of the radius of the depth of the MMSEA equalizer and the first stage of the MMSEA / MPIC is given in Figure 11. Considering the curves of the performances in Figures 8, 9 and 10, we clearly recognize the complementarity character of the performance of MMSEA and MPIC. Indeed, when the level of the thermal noise is large (which corresponds to a low ratio E ,, / Nä) and / or the number of allocated codes is relatively low the MPIC leads to better performance than the MMSEA. However, once the interference becomes significant (with a high Eh / Nä ratio and / or when the number of allocated codes increases), the MMSEA wins. The phenomenon of saturation of the MPIC performance curves is very apparent in Figures 10 and 11 approximately from Eh / N = 7 dB. The performance improvement in the case of the MPIC is not up to the number of stages implemented.

  The performances obtained using the MMSEA / MPIC hybrid receiver are much higher than those obtained with the MMSEA or the MPIC when they are used separately. Indeed, for example in the case where K = 1 o codes (Figure 9) and for a TEB equal to i o- ', the gain Eh / No is approximately 5 dB compared to the MMSEA. It is also important to note that the MMSEA / MPIC converges virtually after only two stages. Thus, the first stage gives a first estimate of the symbols, while the second stage refines this estimate. The fact that the MMSEA / MPIC converges only after two stages solves problems with the MPIC. First, there is the complexity aspect, where with a much lower complexity than that required for the MPIC, MMSEA / MPIC achieves much better performances (example of 2-stage MMSEA / MPIC and 10-stage MPIC for FIGS. , 9 and 10). The second problem is that of determining the number of stages necessary for performance convergence (in terms of BER) which is a crucial problem for the implementation of the MPIC. The third problem is that of determining the weighting factor a. Indeed, in the case of MPIC we have no information a priori for its determination. As the MMSEA / MPIC is preferably made up of only two stages and the first stage gives a good estimation of the symbols, we fix a to 1 to express our confidence in this estimate and to free ourselves from the problem of its determination. From the point of view complexity, it should also be noted that at the cost of an additional computation effort compared to the MMSEA (regeneration of interference is their cancellation according to the different paths), it is possible with the MMSEA / MPIC according to the invention to achieve a noticeable improvement in performance. Thus, the two-stage MMSEA / MPIC hybrid receiver according to the invention has a good compromise between performance and complexity.

  The preferred field of application of the invention is that of advanced receivers for 3G mobile terminals and more. The hybrid receiver according to the invention has good characteristics. It gives superior performance for acceptable complexity. It represents a good compromise between performance and complexity compared to existing approaches. It is very suitable to equip mobile terminals of categories 7, 8, 9 and 10 of HSDPA mode FDD. Indeed, these categories of terminals must support a 16QAM higher order modulation and a number of allocated codes which can be up to 10 or 15. However, the hybrid receiver according to the invention can be used for any communication system without wire using CDMA as an access technique and requiring advanced processing, eg systems heavily loaded with users.

Claims (15)

  1. A method for receiving an analog signal (r (t)) with a baseband spectrum spreading from a propagation channel mufti paths, said analog signal conveying symbols, the method comprising: a step (E240 ) multi-code equalization of said channel adapted to deliver a first estimate (c! ä, ct, A) of said symbols corresponding to each of said codes (cl, ck); a step (E40) for regenerating, from said estimated symbols (d, d, K), replicas (', r,) of said analog signal corresponding to said paths; a step (E50) for regenerating, from said replicas (r ,,), interference for each of said paths; and a step (E60) known as an "interference cancellation step" for delivering a second estimate (d 1, (,, K) of said symbols by canceling said interference of said analog signal (r (t)) for each of said trips.   The receiving method according to claim 1, characterized in that said equalizing step (E240) is a linear equalization step minimizing the mean squared error.   3. Reception method according to claim 2, characterized in that it comprises a step (E220) for automatically determining an optimum depth (popt) used as depth in said linear equalization step (E240), from the response. pulse of said channel (h), and the power () of the thermal noise.   4. Reception method according to claim 3, characterized in that said linear equalization step (E240) is working at the chip rate.   5. Reception method according to claim 4, characterized in that it comprises, prior to said step (E240) of linear equalization at the chip rate, a step (E210) for delivering, from said analog signal (r (t )), a chip rate signal, said step (E210) being performed according to the general principle of a reception step implemented in a receiver type RAKE.   6. Reception method according to any one of claims 2 to 5, characterized in that said equalization step uses a simplified form of GMMSE equalization matrix: Gu.vsi. Is a Hermitian band Toeplitz matrix, wherein: H is a block diagonal matrix of the complex gains of said channel, - D is a matrix containing delayed versions of said channel of the formatting pulse of said Nyquist root analog signal, - IQN is the identity matrix, and 15 is the variance of the thermal noise.   7. Reception method according to claim 6, characterized in that it comprises a step (E230) for calculating the coefficients of the first column of said Hermitian band Toeplitz matrix (F), during which: E232) a convolution sequence (y) between the discrete impulse response of said channel (h) and the pulse (`P) for shaping said Nyquist root analog signal; calculating (E234) an autocorrelation sequence (R + yy) of said convolution sequence (y) for positive delays of said convolution sequence; and - (236) said autocorrelation sequence (R + yy) is sampled at the chip rate. 30   8. Reception method according to any one of claims 3 to 7, characterized in that during said step (E220) for automatically determining said optimum depth (Popt) - (E224) is calculated an equalizing filter of a predetermined maximum depth (PM, x) and, for all odd intermediate depths (P) less than said predetermined maximum depth (PMAx), a finite depth intermediate equalizer filter (g (i,; ',,,) of said intermediate depth (P), and calculating (E226) a relative error (er '"') between the central element ((1v + 1)) of said filter (gä) and the central element (g; ~, ~ ;: f ~ (ti, r + 1)) of said filter (g) of maximum depth (PM,); and - selecting (E228) for said optimum depth (popt), said minimum intermediate depth for which said relative error (e ') "is less than a predetermined error threshold (er).   Receiving method according to claim 6 and 8, characterized in that said maximum depth (PMAX) is chosen (E222) such that PMA (= 4W + 1 (3) where W is the half bandwidth of said matrix ( F) Toeplitz band Hermitienne.   10. Reception method according to claim 8, characterized in that said maximum depth (PM) is chosen (E222) such that: tiv ,,,,,, = 2W erf ((Eh / No) / 10) with PMAx = 2wM, x + 1 (4) where W is the half bandwidth of said matrix (F) Hermitian band Toeplitz, Eb energy per bit, No the monolateral spectral density of the noise, and eif (.) The error function defined by: eif (x) n J = fexp (_y2) ciy 25   Receiving method according to any one of claims 8 to 10, wherein the coefficients of said finite depth intermediate equalizer filter (gt; ~, ti.1) are obtained (E224) from an equalization matrix. intermediate (Go) of the form = (F, + ai,) where a. is the variance of the noise, Ip the identity matrix and F, is an intermediate Hermitian Toeplitz matrix which translates the effect of the channel and discrete shaping pulse into Nyquist root of said analog signal, characterized in that said Intermediate equalization matrix (Gp) is obtained (E224) recursively by the formula [Gp-1 + ppbpbrv bp bH - 'Pp, initialized by: G, = 1 a- + f Gp with b ,, _, d ,, and p ,, = + f for p> 2 where f ,, is the first column of the matrix F, explicit f ,, _ f. d ,, = [f ,,, f ,, _,. ,. ] "and f = f.   12. Reception method according to any one of claims 1 to 11, characterized in that it implements at least one additional iteration of said regeneration steps (E40, E50) and interference cancellation (E60) for issuing, from said second estimate of the symbols (d ,,, d, K), at least a third estimate of said symbols (dm,).   13. Device for receiving an analog signal (r (t)) with a baseband spectrum spreading from a multipath propagation channel, said analog signal conveying symbols, said receiving device comprising in series: a multi-code channel equalizer (cl, CO of said analog signal (r (t)) adapted to deliver a first estimate (dd, dj of said symbols corresponding to each of said codes; means for regenerating, from said estimated symbols); ), replicas (r, t ;,) of said analog signal corresponding to said paths; - means for regenerating, from said replicas (r,), interferences for each of said paths; and - means for delivering a second estimating (2,, 2K) said symbols by canceling said interferences of said analog signal (r (t)) for each of said paths.   14. A computer program comprising instructions for performing the steps of the receiving method according to any one of claims 1 to 12 when said program is executed by a computer.   15. A computer-readable recording medium on which is recorded a computer program comprising instructions for carrying out the steps of the reception method according to any one of claims 1 to 12.5.