WO2017098170A1

WO2017098170A1 - Method and device for combining frames of complex symbols

Info

Publication number: WO2017098170A1
Application number: PCT/FR2016/053285
Authority: WO
Inventors: Louis Adrien DUFRENE; Jean Schwoerer
Original assignee: Orange
Priority date: 2015-12-09
Filing date: 2016-12-08
Publication date: 2017-06-15
Also published as: FR3045247A1

Abstract

The method according to the invention is intended to be implemented in a receiver of a communication system and comprises: - a step (E10) of receiving a plurality N of frames resulting from N successive emissions of a given frame of complex symbols on a multipath propagation channel; - a step (E20) of filtering each frame of a plurality M of frames obtained from the N received frames, M designating an integer less than or equal to N, said step of filtering each frame being carried out by means of a linear filter determined from an estimation of the multipath propagation channel and selected so as to concentrate an energy of the paths of an equivalent propagation channel experienced by the frame on a reduced number of paths; and - a step (E30) of combining the plurality M of filtered frames obtained at the output of the filtering step.

Description

Method and device for combining complex symbol frames

Background rinenvention

The invention relates to the general field of telecommunications.

It relates more particularly to the transmission of digital data, and in particular to a method intended to be implemented at a receiver of a radio communication system and making it possible to take advantage of a plurality of successive transmissions of the same data frame on a radio propagation channel.

In recent years, we have been developing long-range radio communication systems, such as the EC-GSM (Extended Coverage Global System Communications) system or other developments in existing mobile communication systems (eg UMTS (Universal Mobile Telecommunications System), LTE (Long Term Evolution). These systems have a preferred but non-limiting application in the context in particular of networks of connected objects or the Internet of Things (or IoT for "Internet of Things" in English) to ensure inter-object communications.

The long range of a communication system reflects its ability to withstand greater attenuation than usual, which requires increased sensitivity of the receiver. A technology conventionally used to achieve this high sensitivity is to reduce as much as possible the bandwidth of the radio signal transmitted by the communication system, sometimes at the cost of a very low bit rate. The adoption of such a technology requires, however, to rely on a radio communication system exclusively designed for very low-rate use. In the case of an existing system having to be able to provide higher data rates when propagation conditions permit, such an approach is no longer possible.

To increase the radio range of an existing communication system without making significant changes to it, it is necessary to increase the energy per transmitted symbol. The most common way to achieve this result is to repeat several times (ie N times, N designating an integer greater than 1, equal for example to 4, 16 or 32), blindly (ie without feedback from the receiver), one and the same frame of complex symbols, and to emit these N repetitions successively on a propagation channel exhibiting slow variations. We thus obtain N correlated repetitions. The theoretical gain in dB resulting from this operation is then equal to 10log (N). However, to take advantage of the repetitions and reach this theoretical gain, the receiver must be able to (re) judiciously combine the N repetitions of the frame transmitted successively on the channel.

Two principles of combination or recombination are known in the state of the art and conventionally adopted at the level of the receivers of the communication systems: a recombination of the N repetitions after demodulation: each received symbol frame (ie each repetition of the original frame) is equalized then demodulated separately by the receiver into a coded bit frame. A bit-to-bit matching between the N bit frames thus obtained is then made to take, for each bit of the final frame (that is to say the recombined frame), a decision on the value of this bit that takes into account each of the N bits from the N repetitions of the original frame. This recombination principle is also known as a Chase combination or "Chase combining" in English. It is described in more detail in D. Chase's paper entitled

"Code combining - A Maximum-Likelihood Decoding Approach for Combining an Arbitrary Number of Noisy Packets", IEEE Transactions on Communications, Vol. com-33, No. 5, May 1985;

A coherent recombination of the N repeats (via a simple sum for example) carried out immediately after sampling, that is to say directly on the complex samples

(or samples IQ for "In-phase and Quadrature" in English that is to say in phase and in quadrature) obtained at the output of the propagation channel, so as to physically accumulate the energy of each transmitted complex symbol. The frame of complex samples resulting from this type of recombination, also known as IQ accumulation or "IQ Accumulation" in English, then has a signal-to-noise ratio sufficient to allow the implementation of the usual digital processing, namely equalization , demodulation and decoding. The principle of IQ accumulation is described in more detail in the document 3GPP TSG Geran # 66, Tdoc GP-150429 entitled "EC-GSM, On the impact to legacy GSM / EDGE Base stations", 25-28 May 2015.

The principle of Chase recombination offers the advantage of a relatively simple implementation, because it is done at the bit level. It is also robust to variations in the radio propagation channel that can appear during the N successive emissions of the same frame. However, the performances obtained with this type of recombination are generally well below the theoretical gain mentioned above, because of the loss of information related to the different digital processing (equalization and demodulation) carried out upstream of the recombination.

The IQ accumulation type recombination principle offers the possibility of achieving a gain close to the theoretical gain. However, this gain is achieved at the cost of increased sensitivity to the stability of the radio propagation channel on the N repetitions, which can be problematic, especially when the receiver is in a situation of mobility.

Object and summary of the invention

The invention makes it possible to overcome, in particular, the aforementioned drawbacks of the state of the art by proposing a method of combining complex symbol frames intended to be implemented in a receiver of a communication system, said method comprising:

A step of receiving a plurality N of frames resulting from N successive transmissions of the same complex symbol frame on a multipath propagation channel; A step of filtering each frame of a plurality M of frames obtained from said plurality N of received frames, M denoting an integer less than or equal to N, this filtering step of each frame being carried out by means of a linear filter determined from an estimate of the multipath propagation channel and chosen so as to concentrate a path energy of an equivalent propagation channel undergone by the frame (ie seen by the frame after filtering) on a reduced number (in relation to the number of paths of the multipath propagation channel) and determined paths; and

A step of combining the plurality M of filtered frames obtained at the output of the filtering step.

Correlatively, the invention also relates to a receiver of a communication system comprising:

A receiver module of a plurality N of frames resulting from N successive transmissions of the same complex symbol frame on a multipath propagation channel;

A filtering module, configured to filter each frame of a plurality M of frames obtained from said plurality N of received frames, M denoting an integer less than or equal to N, this filtering module comprising a linear filter determined from an estimate of the multi-path propagation channel and selected so as to concentrate an energy of the paths of an equivalent propagation channel undergone by the frame on a reduced number of paths; and

A combination module, configured to combine the plurality M of filtered frames obtained at the output of the filtering module and to provide a combined frame.

In a particular embodiment, the method further comprises a step of equalizing the combined frame resulting from the combining step, the filtering step being a pre-filtering step adapted to the equalization step. Correlatively, the receiver comprises, in this embodiment, an equalizer configured to equalize the combined frame provided by the combination module. Such an equalization step, respectively such an equalizer, allows, in a manner known per se, to compensate for the effects of a multipath propagation channel (inter-symbol interference in particular).

Thus, the invention proposes the introduction of a new recombination point of repetitions of a complex symbol frame transmitted successively on a propagation channel. This new point is advantageously after a filtering or prefiltering step (if the method further comprises an equalization step) of M complex symbol frames obtained from the N frames received complex symbols, this filter being adapted to the channel of propagation undergone by repeated symbol frames.

The (pre-) filtering is carried out, according to the invention, on each of the M frames, and leads for each of these frames, to a concentration of the energy of the paths of the equivalent propagation channel undergone by each frame on the same a small number of paths (for example on one or two paths, preferably on the first path or paths to facilitate of the invention). It is adapted to the propagation channel (in the sense that it takes into account an estimate of the propagation channel), and thus makes it possible to mitigate the effect of its variations, if any, on the M frames. Indeed, thanks to the filtering operation carried out, homogenization of the channels undergone by the M frames (ie all the equivalent channels are similar with an energy concentrated on a reduced number of samples (preferably on the first samples), even on a single sample if only one path is preferred). In this way, it is ensured that the (re) combination of the frames realized by the invention, at the output of this filtering, is performed on correlated symbols, and is constructive.

Furthermore, the recombination proposed by the invention is carried out at the level of complex symbols having undergone only a (pre-) filtering adapted to the propagation channel, in other words before any other nonlinear digital processing likely to generate a loss of information. .

The invention therefore makes it possible, like the recombination principle "accumulation IQ", to reach a gain close to the theoretical gain in 10log (N), while greatly reducing the sensitivity of the receiver relative to the variations of the propagation channel. The recombination principle proposed by the invention thus makes it possible to achieve near-optimal performances.

Different filters and, where appropriate, equalizers may be used by the receiver within the scope of the invention.

Thus, in a particular embodiment of the invention, the linear filter used during the filtering step transforms the equivalent propagation channel undergone by the frame into a channel with a minimum phase impulse response.

Such a filter facilitates where appropriate the equalization step implemented after the combination step.

In another embodiment of the invention, the linear filter used during the filtering step is a suitable filter.

In another embodiment, the equalization step implements an RSSE (Reduced-State Sequence Estimation) equalizer, and the pre-filtering step uses a linear filter adapted to such an equalizer RSSE, known per se.

Such a matched filter, as well as such an equalization block consisting of a pre-filter and an equalizer RSSE constitute in themselves conventional filtering and equalization schemes commonly used in communication systems. The invention can therefore be implemented easily in existing digital radio communication systems whose extension is sought to extend the radio range.

Thus, a preferred field of application of the invention is that of the radio connectivity for the Internet of Things (IoT) for which the criterion of broad radio coverage (or broadband range equivalent), including in the field isolated (also called "white area") or in an unfavorable location (eg in an underground), is paramount. In one In this context, the invention applies in particular judiciously to developments in second (eg GSM), third (eg UMTS) and fourth generation (eg LTE) mobile telephony standards. The invention offers, for example, a promising alternative to the Chase-type recombination techniques and the "IQ accumulation" type currently considered in the EC-GSM evolution currently being specified by the 3GPP-GERAN standardization group.

The aforementioned examples of (pre-) filters and equalizers are however given for illustrative purposes only and other filters, pre-filters and / or equalizers may alternatively be envisaged.

In a particular embodiment, the step of combining the plurality M of filtered frames comprises a weighted sum of these M filtered frames.

Such a combination is relatively simple to implement and allows a constructive recombination of the frames transmitted successively on the propagation channel. The weighting can notably take into account the quality of the channel on the frame (repetition) considered and associate a greater weight with the channels presenting a better quality. As a variant, a weighting based on a measurement of the power of each received frame (or RSSI for "Received Signal Strength Indication" in English) may be envisaged, or according to another variant still an unweighted sum or an average may be envisaged for combine the M frames.

In a particular embodiment, N = M and the filtering step is performed on each of the N frames received.

This embodiment is particularly simple, the filtering and combining steps being applied directly to the N repetitions of the original frame successively transmitted and received by the receiver.

In another embodiment, N = k.M where k denotes an integer greater than 1, and each frame of the plurality M of frames results from a combination of k successive frames distinct from the plurality N of received frames.

In other words, in this embodiment, a first (re) combination of k repetitions of the original symbol frame is implemented according to a recombination principle "IQ accumulation" per block, then a second (re) combination is implemented. after the filtering or pre-filtering step. The integer k is an integer greater than or equal to 2. It is chosen so that on each set of k distinct successive frames combined, the multipath propagation channel can be considered constant or having slow variations.

This first recombination according to the "accumulation IQ" principle makes it possible to increase the signal-to-noise ratio over a block of k repetitions and thus to make reliable the estimation of the radio propagation channel undergone on this block. Then, the frame resulting from the accumulation IQ is (pre-) filtered by means of a filter adapted to an estimate of the equivalent propagation channel seen by the frame.

Once all the frames resulting from IQ accumulation have been filtered, they are immediately combined. This embodiment combining the two types of recombination IQ accumulation and according to the invention has several advantages.

Firstly, the estimation of the equivalent channel considered during (pre) filtering is particularly reliable and allows an optimal operation of the (pre-) filtering step and the equalization step, especially in case of variations of the propagation channel on all repetitions.

In addition, the recombination "IQ accumulation" here enjoys a quasi-optimal gain since on each block of k repetitions, the variations of the channel can be considered very slow, in other words the propagation channel is almost constant.

Moreover, the recombination after (pre-) filtering has good performances thanks to the good estimation of the equivalent propagation channel undergone by each frame resulting from the recombination "accumulation IQ", decreasing the effect of the variations of the channel.

This mixed embodiment therefore makes it possible to achieve very good performance while being more robust to variations in the propagation channel.

In a particular embodiment, the various steps of the combination method are determined by computer program instructions.

Consequently, the invention also relates to a computer or microprocessor program on an information medium, this program being capable of being implemented in a receiver or more generally in a computer, this program comprising instructions adapted to implementing the steps of a combination 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 or microprocessor-readable information medium, comprising instructions of a 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 floppy disk 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.

The invention also relates to a communication system comprising: An emitter capable of emitting successively the same complex symbol frame of a plurality N of times on a multipath propagation channel; and

A receiver according to the invention capable of processing the plurality N of frames resulting from the N successive transmissions of the transmitter.

The communication system has the same advantages, mentioned above, as the receiver and the combination method according to the invention.

It can also be envisaged, in other embodiments, that the combination method, the receiver and the communication system according to the invention present in combination all or part of the aforementioned characteristics.

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:

- Figure 1 shows, schematically, a communication system according to the invention in a particular embodiment;

FIG. 2 illustrates the hardware architecture of the receiver of the communication system of FIG. 1, in accordance with the invention;

FIG. 3 represents the different modules of the receiver of the communication system of FIG. 1, as well as the steps of the combination method according to the invention implemented by these modules, in a first variant embodiment; and

FIG. 4 represents the different modules of the receiver of the communication system of FIG. 1, as well as the steps of the combination method according to the invention implemented by these modules, in a second variant embodiment.

Detailed description of the invention

FIG. 1 represents, in its environment, a radio communication system 1 according to the invention, in a particular embodiment.

The communication system 1 comprises a transmitter 2 and a receiver 3 according to the invention, able to exchange data in the form of digital signals or more particularly complex symbol frames conveying information. The transmitter 2 and the receiver 3 are able to communicate via a telecommunications network, such as for example a mobile telecommunications network. This network can be in particular a GSM network or its evolution EC-GSM offering an extended scope. The transmitter 2 is for example a base station of the network and the receiver 3 a mobile terminal able to communicate with this base station (downlink).

However, the invention also applies uplink, each entity able to communicate on the network can be both a transmitter and a receiver compliant with the invention. However, in the remainder of the description, for the sake of simplification, we focus more precisely on the transmission function of the transmitter 2 and the reception function of the receiver 3.

Moreover, the invention also applies to other mobile telecommunications networks such as a UMTS or LTE network, etc.

The transmitter 2 and the receiver 3 are separated in a manner known per se by a propagation channel 4 (eg via an overhead channel). For the sake of simplification here, it is assumed that the transmitter 2, respectively the receiver 3, is equipped with a single transmitting antenna, respectively receiving antenna.

In known manner, the transmission of digital signals between the transmitter 2 and the receiver 3 is affected by the noise related to the imperfections of the communication system 1 and the physical nature of the components used in this system (such as, for example, the antennas transmission and reception of transmitter 2 and receiver 3). Moreover, these signals undergo deformations when they propagate between the transmitting antenna (or antennas) of the transmitter 2 and the receiving antenna (or antennas) on the propagation channel 4. As a result of the description, the concept of transmission or propagation channel 4 will encompass not only the effects of the medium on which the digital signal propagates between the transmitter 2 and the receiver 3 (eg radio channel) but also the effects induced by the transmit and receive antennas of the latter on the digital signals.

The invention has a preferred but nonlimiting application in the context of transmissions of data on a frequency-selective propagation channel (multipath or multi-path channel) and with relatively slow variations in time. In a known manner, the frequency selectivity of a propagation channel is linked to the digital signal that it is desired to transmit on this channel: it reflects the fact that this digital signal has frequency components that are attenuated differently by the propagation channel. . This phenomenon occurs when the bandwidth of the signal to be transmitted is much greater than the coherence band of the propagation channel, the coherence band of a channel being defined as the minimum bandwidth for which two attenuations of the channel are independent. . In this case, the signal propagates along several paths on the radio channel, each path being characterized by an attenuation, a delay and a phase shift of its own. The signal received at the receiver is the sum (constructive or destructive) of the signals having passed through these different paths. The compensation of the effects of the distortions introduced by the multipath channels is conventionally carried out using equalization techniques. In the remainder of the description, it is assumed that the propagation channel 4 is a multi-path channel comprising L paths, L denoting an integer greater than 1, and can be modeled by a finite impulse response filter h characterized by L complex coefficients. .

As mentioned above, the transmitter 2 and the receiver 3 are able to exchange data in the form of digital signals and more particularly frames complex symbols (or IQ symbols). The transmission chain making it possible to obtain such complex symbol frames is known to those skilled in the art and depends on the telecommunications network via which the transmitter 2 and the receiver 3 communicate. As an illustration and in a simplified manner, this chain transmission system comprises in particular the following functional modules: a source 2A of binary information (information bits), a channel coder 2B able to transform the binary information into coded bits, a modulator 2C able to transform the coded bits coming from the coder in complex symbols extracted from a modulation, a 2D module for forming complex symbol frames, and a module 2E for sending these frames on the propagation channel 4 to the receiver 3 in the form of an analog signal S No limitation is attached to the choice of coding scheme or modulation.

It is assumed here that in order to benefit from an extended range, the transmitter 2 implements a simple technique making it possible to increase the energy of the symbols received by the receiver 3. This technique consists in blind repeating a predetermined number N times of the same T frame of complex symbols, N denoting an integer greater than 1. For example, in the conventional communication systems (and in the system 1 considered here), N may be equal to 4, 16 or even 32.

The N repetitions of the frame T, designated by ΤΙ,,,,, ΤΝ are sent by the transmitter 2 to the receiver 3 successively in time. These N frames are frames of complex symbols, conveyed by the signal S. The receiver 3 is able to process this plurality N of frames ΤΙ,.,., ΤΝ resulting from N successive transmissions of the transmitter 2 according to the invention. .

In the embodiment described here, the receiver 3 has the hardware architecture of a computer. It comprises in particular a processor or a microprocessor 5, a random access memory 6, a read only memory 7, a non-volatile memory 8 and a communication module 9 integrating in particular the reception antenna 9A of the receiver 3, etc.

The read-only memory 7 of the receiver 3 constitutes a recording medium in accordance with the invention, readable by the processor 5 of the receiver 3 and on which is recorded a computer program or microprocessor PROG according to the invention, comprising instructions for performing the steps of a method of combining N complex symbol frames transmitted by the transmitter 2, according to the invention.

This computer or microprocessor program equivalently defines functional modules (and software here) configured to implement the steps of the combining method, and described in more detail now with reference to two particular embodiments of the invention. . These functional modules rely on or control the previously described hardware resources of the receiver 3.

FIG. 3 illustrates the various functional modules, defined by the computer program of the receiver 3, on which the invention is based in a first variant embodiment. Equally, the processing of N repetitions of the frame T by these different functional modules constitutes the main steps of the combination method according to the invention in this first embodiment. Thus will be designated within each module in parentheses the step or steps implemented by the module.

As mentioned above, the N repetitions T1, T2, ..., TN of the frame T are sent by the transmitter 2 to the receiver 3 successively in time in the form of an analog signal S transiting on the propagation channel 4 separating the transmitter 2 from the receiver 3.

The signal S having passed through the propagation channel 4 is designated by the signal S '. Upon receipt of this signal S 'by the communication module 9 of the receiver 3, it is sampled by a sampling module 3A known per se, including an analog-to-digital converter (ADC) (step E10). A number N of frames Τ, Τ2 ', ..., ΤΝ' of complex symbols are obtained at the output of the sampling module 3A. They represent the N repetitions of the frame T having passed through the propagation channel.

In the first embodiment described here, the N frames Τ, T2 ',..., TN' are supplied to an equalization block 3B. This equalization block 3B here comprises a pre-filter 3B-1, a combination module 3B-2 and an equalizer RSSE (for Reduced-State Sequence Estimation or estimation of reduced state sequence) 3B-3. Such an equalizer is known to those skilled in the art and described for example in the document by M. Vedat Eyuboglu et al. entitled "Reduced-State Sequence Estimation with Set Partitioning and Decision Feedback," IEEE Transactions on Communications, Vol. 16, No. 1, January 1988.

In other words, in this first embodiment, the filtering performed by the pre-filter 3B-1 is applied successively to all the N frames Τ, T2 ',..., TN' (step E20), or equivalent again on a set comprising M frames respectively corresponding to N frames Τ, T2 ',..., TN', M denoting an integer equal here to N.

The pre-filter 3B-1 is adapted to the RSSE equalizer 3B-3. It is able to transform the impulse response of the propagation channel seen by each frame into its minimum phase equivalent (ie, the z-transform of this phase-minimum impulse response has its roots only inside and on the circle unit). In the case of the multipath propagation channel 4, the pre-filter 3B-1 is therefore a linear filter making it possible to concentrate the energy of the paths of the equivalent propagation channel undergone by each of the frames Τ, T2 ', ... , TN '(ie seen by each frame at the output of the pre-filter 3B-1) on a reduced and predetermined number of paths L', L 'being an integer less than the number of paths L of the propagation channel 4. In the embodiment described here, the equivalent propagation channel resulting from the passage of the frames by the propagation channel 4 and the pre-filter 3B-1 is truncated so as to also include L paths. In other words, the energy concentrated on the chosen paths is greater than the energy of the L-L 'remaining paths of the equivalent propagation channel.

For example, the pre-filter 3B-1 is chosen such that L '= 1 or 2, and the L paths retained here are the first L paths of the equivalent propagation channel (temporally speaking). The energy of (L'-L here) last coefficients of the impulse response of the equivalent propagation channel seen by each frame (resulting in a manner known to those skilled in the art of the convolution of the impulse response of the propagation channel 4 and the response of the pre-filter 3B-1) is in this case minimal. Such a pre-filter is described for example for the RSSE equalizer in the WH Gerstacker et al., Entitled "On Prefilter Computing for Reduced-State Equalization," IEEE Transactions on Wireless Communications, Vol. 1 No. 4, October 2002 and is not repeated here. Applying such a pre-filter upstream of the RSSE equalizer maximizes the performance of the RSSE equalizer.

Alternatively, the paths selected are not necessarily the first paths of the equivalent channel.

The pre-filter 3B-1 and the equalizer RSSE 3B-2 of the equalization block 3B are determined (and updated, for example periodically) from an estimate h_est of the propagation channel 4 and its different paths. . Since the channel varies slowly, the estimate h_est varies little from one frame to another (the channels seen by each frame are correlated, ie the correlation between the channels is greater than 0.5) and a periodic update of the impulse response. pre-filter and RSSE equalizer parameters is sufficient to track possible variations of the propagation channel, the period can be chosen according to the coherence time of the channel. No limitation is attached to how the estimate h_est of the propagation channel 4 is obtained. This estimate can be calculated by the receiver 3 in a manner known per se, in particular using pilot sequences transmitted by the transmitter 2 to the receiver 3, or be provided by the transmitter 2 to the receiver 3 via a return channel, etc.

The examples of pre-filter and equalizer RSSE described in the example envisaged here are given for illustrative purposes only. Other equalizers and / or pre-filters can be envisaged in the equalization block 3B. For example, the equalization block 3B may comprise an equalizer MLSE (for Maximum Likelihood Sequence Estimation) or a equalizer DFE (for Decision Feedback Equalizer) known to those skilled in the art.

In another alternative embodiment, the equalization block 3B can be reduced to a filter 3B-1 and a combination module 3B-2 (i.e. the block 3B does not include an additional equalizer). In this case, the filter 3B-1 is for example a matched filter known per se, the impulse response of which is given by h * (- t) where h (t) denotes the temporal impulse response of the propagation channel 4 seen by each repetition of the frame T and * denotes the conjugation operator.

The (pre-) filtering step E20 of the N frames Τ,,,,, ΤΝ 'by the pre-filter 3B-1 results in N prefiltered frames Tlf', ..., TNf. These N prefiltered frames Tlf ', ..., TNf are provided at the input of the combination module 3B-2 of the equalization block 3B.

The combination module 3B-2 then combines the N prefiltered frames Tlf ',..., TNf into a combined frame Tcomb of complex symbols (step E30). In the first embodiment described here, each symbol of the combined frame Tcomb is obtained by summing the corresponding complex symbols of the N prefiltered frames Tlf ', ..., TNf. Alternatively, other combinations may be envisaged, such as, for example, a weighted sum taking into account the attenuation of the channel on each symbol, an average of the symbols of the N prefiltered frames, etc.

The combined complex symbol frame Tcomb is then supplied to the equalizer RSSE 3B-3 (step E40).

The equalized frame Tega obtained at the output of the equalizer 3B-3 is then demodulated by a demodulator 3C, and then decoded by a 3D decoder, in a manner known per se and not described in detail here. The 3C demodulator and the 3D decoder, of course, depend on the modulation and coding scheme used at the transmitter 2 and perform the inverse operation of this scheme.

Thus, in this first variant embodiment, the recombination of the N repetitions of the frame T successively sent on the propagation channel 4 by the transmitter 2 is performed on complex symbols having undergone a first filtering (pre-filter 3B-1). adapted to the propagation channel 4 which mitigates the effects of the propagation channel. Updating the coefficients of the pre-filter 3B-1, for example periodically from the estimate h_est of the channel which the receiver 3 has, makes it possible to take into account and mitigate the effect of the variations of the propagation channel. Recombination is done directly on complex symbols (samples) before any nonlinear processing. It thus allows the constructive accumulation of the energy of the N repetitions of the frame T with respect to a white noise. It results from this recombination scheme implemented according to the invention by the receiver 3 in a gain close to the optimal theoretical gain equal to 10log (N). This scheme therefore makes it possible to reach a quasi-optimal level of performance such as that achieved by an "accumulation IQ" recombination but with a vulnerability to variations in the reduced radio propagation channel.

In the first embodiment described here, the combination of the frames is performed at the output of the pre-filter 3B-1 on the N successive repetitions sent by the transmitter 2. In a second variant embodiment, the receiver operates a recombination of the energy carried by these frames in two stages: firstly at the output of the sampling module, by block of k frames, k designating an integer greater than 1, then in a second time at the output of the pre-filter 3B-1 .

FIG. 4 illustrates such a receiver (designated for the sake of clarity by 3) as well as the various functional modules defined by the computer program of the receiver 3 ', on which the invention is based in this second variant embodiment. the processing of the N repetitions of the frame T by these different functional modules constitutes the main steps of the combination method according to the invention in the second embodiment of the invention: the modules of the receiver 3 'and the steps implemented by these modules according to the second embodiment which are identical or similar to the modules of the receiver 3 and the steps implemented by the modules of the receiver 3 in the first embodiment are designated by the same references. As in the first variant embodiment, the N repetitions T1, T2,..., TN of the frame T are sent by the transmitter 2 to the receiver 3 ', successively in time, in the form of an analog signal S transiting on the propagation channel 4 separating the transmitter 2 of the receiver 3 '.

The signal S 'received by the communication module 9 of the receiver 3' is sampled by a sampling module 3A (step E10), resulting in N frames Τ, Τ2 ', ..., ΤΝ' of complex symbols.

In the second variant embodiment, a first combination module 3A 'firstly combines the N frames Τ, T2',..., TN 'by block of k frames (step E15), so as to obtain M frames T1. ", T2", ..., TM ", k designating an integer greater than 1 such that N = kM Each frame T" i is a complex symbol frame resulting here from the sum, symbol to symbol, of k frames Tj with j = 4i, 4i + 1, 4i + 2, ..., 4i + kl. In other words, the first combination module 3A 'performs an "IQ accumulation" recombination by block of k successive frames on the N frames Τ, T2', ..., TN '.

Alternatively, other combinations may be envisaged, such as, for example, a weighted sum, an average, and so on.

This first combination realized by the combination module 3A 'is intended to improve the estimation of the radio channel used by the equalization block 3B of the receiver 3', by increasing the signal-to-noise ratio of the signals used to estimate the channel radio frequency thanks to the first recombination performed on k successive frames, k being greater than or equal to 2. The integer k is chosen preferentially so that on a block of k successive frames, the propagation channel 4 remains invariant or undergoes slight variations , k can therefore be chosen as a function of the coherence time of the channel or, alternatively, be constant, but be determined to suit most of the propagation channels experienced by the transmitter 2 and the receiver 3. In this way, it is ensured that that the combination made by the combination module 3A 'is constructive which makes reliable estimation of the radio channel suffered on average by each block of k repetitions.

The M frames delivered by the first combination module 3A 'are then successively supplied to the equalizing block 3B to be filtered.

As in the first embodiment, the equalization block 3B comprises a pre-filter 3B-1, a second combination module 3B-2 and an equalizer RSSE (Reduced-State Sequence Estimation) 3B-3. In other words, in this second embodiment, the filtering performed by the pre-filter 3B-1 is successively applied to the M frames Tl ", T2 ', ..., TM" obtained from the N frames Τ, T2' , ..., TN '(step E20), and makes it possible to generate M filtered frames Tlf ", ..., TMf". The equalization block 3B is identical and operates identically to that considered in the first variant, except that the combination module 3B-2 now operates on the M frames Tlf ", T2f", ..., TMf with M <N instead of N filtered frames Tlf, T2f ', ..., NPT, and the channel estimation h_ is taken into account to determine the pre-filter 3B-1 and the equalizer 3B-3 preferentially takes into account the first combination performed by the module 3A '(for example, it takes into account an average value of the propagation channel on the k successive successive frames) and is updated for each filtered frame Tjf ", j = 1, ..., M .

The combined complex symbol frame Tcomb from module 3B-2 is then supplied to the equalizer RSSE 3B-3. The equalized frame obtained at the output of the equalizer 3B-3 is then demodulated by a demodulator 3C, and then decoded by a 3D decoder as in the first embodiment.

This second embodiment has many advantages:

The estimation of the radio channel is reliable, allowing an optimal operation of the pre-filtering and the equalization of the frames;

The combination of the "accumulation IQ" type performed by the module 3A 'here benefits from its almost optimal gain because the variations of the radio propagation channel within each block of frames are assumed to be very small;

The combination carried out immediately after pre-filtering by the module 3B-2 retains good performances thanks to a good re-estimation of the propagation channel 4 for each frame Tj ", j = 1, ..., M, so that reduce the effect of channel variations where appropriate.

This mixed solution makes it possible to combine the advantages of the two types of recombination "accumulation IQ" and according to the invention.

In the embodiment described here (first and second variants), a communication system of the GSM or EC-GSM type based on time division multiple access (TDMA) technology has been considered. " in English). However, the invention also applies to communication systems relying on other types of access technologies, for example code division multiple access or CDMA technology (for "Code Division Multiple Access"). " in English). In this type of communication system, in a known manner, a suitable filtering (eg using a Rake receiver) is performed in reception before or after despreading and the invention can be implemented by recombining the different frames. obtained directly at the output of this adapted filtering.

Claims

A method of combining frames (Τ -ΤΝ '; T1 "-TM") of complex symbols to be implemented in a receiver (3,3') of a communication system (1), said method comprising :

A step (E10) for receiving a plurality N of frames (Τ -ΤΝ ') resulting from N successive transmissions of the same frame (T) of complex symbols on a multipath propagation channel (4);

A step of filtering (E20) each frame of a plurality M of frames (Τ -ΤΝ '; T1 "-TM") obtained from said plurality N of received frames, M denoting an integer less than or equal to N , this filtering step of each frame being carried out by means of a linear filter (3B-1) determined from an estimate (h_est) of the multipath propagation channel and chosen so as to concentrate an energy of the transmission paths. an equivalent propagation channel undergone by the frame on a reduced number of paths; and

A step (E30) of combining the plurality M of filtered frames obtained at the output of the filtering step.

2. Combining method according to claim 1, further comprising an equalization step (E40) of the combined frame resulting from the combining step, the filtering step being a pre-filtering step adapted to the step d 'equalization.

3. Combination method according to claim 1 or 2, in which the linear filter (3B-1) used during the filtering step transforms the equivalent propagation channel undergone by the frame into a channel with a minimum phase impulse response. .

4. Combination method according to any one of claims 1 to 3 wherein the linear filter (3B-1) used in the filtering step is a suitable filter.

5. Combination method according to any one of claims 1 to 4 wherein N = M and the filtering step (E20) is performed on each of N received frames (Τ -ΤΝ ').

The combining method according to any one of claims 1 to 4 wherein N = kM where k is an integer greater than 1, and each frame (T1 "-TM") of the plurality M frames results from a combination (E15) of k successive frames distinct from the plurality N of received frames.

A combining method according to claim 6 wherein k is chosen so that on each set of k separate distinct frames combined, the multipath propagation channel can be considered constant or having slow variations.

The combining method according to any one of claims 1 to 7 wherein the step of combining (E30) the plurality M of filtered frames comprises a weighted sum of these M filtered frames.

9. Combination method according to claim 2 wherein the equalization step (E40) implements an equalizer (3B-3) type RSSE (Reduced Sequence State Estimation).

A computer (PROG) or microprocessor program comprising instructions for performing the steps of the combining method according to any one of claims 1 to 9 when said program is executed by a computer or a microprocessor.

A computer-readable recording medium (7) on which is recorded a computer or microprocessor program including instructions for performing the steps of the combining method according to any one of claims 1 to 9.

12. Receiver (3) of a communication system (1) comprising:

A reception module (9) of a plurality N of frames resulting from N successive transmissions of the same complex symbol frame on a multipath propagation channel;

A filtering module (3B-1), configured to filter each frame of a plurality M of frames obtained from said plurality N of received frames, M denoting an integer less than or equal to N, this filtering module comprising a linear filter determined from an estimate of the multipath propagation channel and selected so as to concentrate an energy of the paths of an equivalent propagation channel undergone by the frame on a reduced number of paths; and

A combination module (3B-2) configured to combine the plurality M of filtered frames obtained at the output of the filtering module and provides a combined frame (Tcomb).

The receiver (3) of claim 12 further comprising an equalizer (3B-3) configured to equalize the combined frame (Tcomb) provided by the combining module (3B-2).

14. Communication system (1) comprising:

A transmitter (2) capable of successively transmitting the same complex symbol frame a plurality N of times on a multipath propagation channel (4); and - A receiver (3) according to claim 12 or 13 adapted to process the plurality N of frames resulting from N successive emissions of the transmitter.