WO2014128176A2

WO2014128176A2 - Methods and devices for transmitting and receiving a multicarrier signal reducing the peak-to-average power ratio, and corresponding programme and signal

Info

Publication number: WO2014128176A2
Application number: PCT/EP2014/053256
Authority: WO
Inventors: Jacques Palicot
Original assignee: Supelec
Priority date: 2013-02-20
Filing date: 2014-02-19
Publication date: 2014-08-28
Also published as: WO2014128176A3; FR3002395A1

Abstract

The invention concerns a method for the multicarrier transmission of a source signal comprising N subcarriers. Said method comprises the following steps: - partial response encoding (20) of said source signal, delivering an encoded source signal comprising a part created by said partial response encoding, referred to as an empty encoded part of said encoded source signal, - generation (21) of a predetermined signal from said encoded source signal, referred to as an additional signal; - addition (22) of said encoded source signal and said additional signal to said empty encoded part, in order to obtain a signal having a peak-to-average power ratio lower than the peak-to-average power ratio of said source signal, referred to as a compact envelope signal; - modulation (23) of a carrier wave by said compact envelope signal, delivering a modulated compact envelope signal; - amplification (24) of said modulated compact envelope signal, using a power amplifier, delivering an amplified signal; - transmission (25) of said amplified signal.

Description

Methods and devices for transmitting and receiving a multicarrier signal reducing peak to average power ratio, program and signal.

1. Field of the invention

The field of the invention is that of the transmission of digital signals, in particular on multi-path transmission channels, for an "eco-radio" or "Green Radio" application aimed at reducing the consumption of radio signals. the energy used during this transm ission.

More specifically, the invention relates to the m ulti-carrier modulation techniques, in particular OFDM type ("Orthogonal Frequency Division Multiplex" in English, for "orthogonal frequency division multiplexing"). O FDM modulation is increasingly used for digital transmission, especially on multipath transmission channels. This multicarrier modulation technique notably makes it possible to overcome the inter-symbol interference generally observed when using a single carrier mod ulation on a channel with multiple paths.

Due to its intrinsic robustness to frequency selective channels, the OFDM modu lation is particularly, but not exclusively, used in wireless local area networks (WiFi or Wi MAX), digital terrestrial television (DTT) or ADSL. Asymmetric Digital Subscriber Line and H I PERLAN / 2, but also for standards such as those relating to Digital Audio Broadcasting. (DAB for "Digital Audio Broadcasting"), Digital Video Broadcasting (DVB-T for "Digital Video Broadcasting").

2. Prior Art

2.1 Disadvantages of OFDM modulation

One of the main drawbacks of OFDM to date remains that OFDM signals have a very wide band and consequently have a large amplitude variation. The envelope of the modulated signal has a strong dynamic as illustrated schematically in FIG. 1 by the signal x (t) and its envelope 11.

Moreover, the signal thus modulated, before being transmitted, must be amplified to compensate for propagation attenuations. FIG. 1 also shows the typical characteristics of an RF power amplifier, with, on the abscissa, the power P _e of the source signal at the input of the amplifier and, on the ordinate, the power P _s of the amplifier. signal at the output of the amplifier. In particular, as long as the amplifier has not reached saturation (horizontal asymptot 13), the amplification is considered to operate in linear zone LI N (oblique asymptotic point 14) up to a saturation threshold power. P _SA T- Thus, with reference again to FIG. 1, when the input source signal x (t) has a strong envelope dynamics, for the high amplitude envelope portions P ₂ , the amplification is then considered as operating in a nonlinear zone with a risk of exceeding the saturation threshold of the amplifier and a generation of nonlinear amplitude and phase distortions.

The peak-to-average power ratio (PAPR) of the transmitted signals is thus generally very high and increases with the number of subcarriers N.

It follows a spectral rise of the level of the secondary lobes (ACPR, for Adjacent Channel Power Ratio), the generation of harmonics, a creation of interferences between nonlinear symbols, a creation of interferences between sub-carriers, which leads in particular to transmission errors and degradation of the bit error rate (BER).

A large value of the PAPR leads for example to take a step back from the compression point of the power amplifier (of the order of the PAPR) to avoid nonlinear distortions.

This decline results in a reduction in the energy efficiency of the power amplifier (a few% instead of, typically, 70%) and therefore a significant increase in the consumption of the transmitter. This is a very strong constraint for the use of OFDM especially in mobile terminals, knowing that the consumption of the power amplifier can represent more than 50% of the total consumption of a terminal.

2.2 Prior art for the reduction of the PAPR

Several methods are already known for reducing the peak power of the OFDM signal, a summary of which is given, for example, in the document "A classification of methods for efficient power amplification of signals" (Y. Louet, J. Palicot, Annals of Telecommunications, vol. 63, Issue 7-8, pp 351-368, July / August 2008).

A common solution is to ensure that the operating range of the amplifier is limited to a linear amplification zone, which unfortunately limits the efficiency of the amplifier.

A second approach is based on adding a signal to decrease the PAPR. More specifically, the added signal has amplitude peaks in phase opposition with those of the signal that is to be corrected.

The clipping technique, or limiter, which consists of clipping the amplitude of the signal when it exceeds a predetermined value, is also a method of adding a signal as demonstrated in particular in the document "Clipping Formulated as a Technical Addition for OFDM Peak Power Reduction" (D. Guel, J. Palicot, IEEE Vehicular Technology Conference, VTC-Spring, Barcelona, Spain, April 26-29, 2009). Indeed, the "clipping" consists in suppressing the signal, however such a deletion is mathematically equivalent to adding a corrective signal.

However, this clipping or addition of signal is inherently non-linear and introduces a distortion of the emitted signal resulting not only in a degraded BER but also in a rise in the secondary lobes of the DSP (Power Spectral Density).

Such signal adding methods are for example combined with a third approach, commonly called "T" technique (of the English "Tone Reservation"), proposing to reserve certain sub-carriers of the OFDM multiplex. Such reserved subcarriers do not carry useful information but the additional correction signal presented above, or symbols optimized for transmission to reduce the PAPR. The optimization of these symbols can be performed using, for example, a convex optimization algorithm of the SOCP (Second Order Cone Programming) type.

The combination of the addition of signal and the TR technique is illustrated in particular in FIG. 1B.

This solution does not distort the transmitted signal, but a major disadvantage of this method lies in the fact that a number of subcarriers must be reserved to reduce the PAPR significantly. These subcarriers are not used to transmit useful information data, which leads to a reduction in spectral efficiency.

2.3 Objectives of the invention

The invention particularly aims to overcome these disadvantages of the prior art. More precisely, the present invention explicitly aims at an "eco-radio" or "Green Radio" application aimed at reducing the consumption of the energy used during this transmission.

Indeed, the steady increase in the number of mobile radio equipment as well as the progress in terms of communication technologies have had a considerable impact in the increase of C02 emissions.

Therefore, an objective of the invention according to at least one embodiment is to propose a technique making it possible to significantly reduce the PAPR in view of the techniques of the prior art while maintaining satisfactory performances in terms of distortion of the transmitted signal, spectral efficiency, TET (Rate of Error Frame) and / or BER (Bit Error Rate). Another objective of the invention, according to at least one embodiment, is to provide a technique that is simple to implement, both in transmission and reception.

3. Presentation of the invention

The invention proposes a new solution that does not have all of these disadvantages of the prior art, in the form of a multicarrier transmission method of a signal representative of a source signal comprising N sub-carriers, intended to be amplified by a power amplifier and then transmitted via a transmission channel.

According to the invention, the transmission method comprises the following steps:

partial response coding applied to each subcarrier of at least a portion of said N subcarriers of said source signal, creating for each subcarrier to which said partial response coding is applied an empty spectral space, corresponding to at least one an empty subcarrier, all of said empty subcarriers forming an empty coded part of said coded source signal,

generating a predetermined signal from said coded source signal, said additional signal;

adding to said empty coded portion of said coded source signal a spectral portion of said additional signal corresponding to said empty coded portion of said coded source signal so as to obtain a signal comprising N _t > N subcarriers and having a peak power ratio at medium power less than the average power peak power ratio of said source signal, said compact envelope signal;

modulating a carrier wave by said compact envelope signal, providing a compact envelope modulated signal;

amplifying said compact envelope modulated signal, using a power amplifier, delivering an amplified signal;

transmitting said amplified signal.

Thus, the invention is based on a new and inventive approach to the transmission of a signal. Indeed, the invention thus proposes to improve the technique of reducing the PAP, the prior art of adding the signal by means of the TR technique (of the English "Tone Reservation") shown in FIG. using partial response coding of a portion of subcarriers of the source signal to create an empty coded portion dedicated to adding the additional signal to reduce the PAPR.

In other words, for each sub-carrier to which the partial response coding is applied according to the invention, there is created an empty spectral space also called an empty subcarrier (of zero value), dedicated to the subsequent addition of a reducing signal of the PAPR. Subcarrier by subcarrier of the part of the source signal to which partial response coding is applied there is created an empty spectral space (also called an empty subcarrier) between each subcarrier of this part.

Then, during the addition step, the value of the carrier of the additional signal corresponding to this empty spectral space of the coded source signal is added (in the time domain or the frequency domain) to this empty spectral space, the sub-frequency. Compact envelope signal carrier thus formed being solely dedicated to the reduction of the PAPR.

It should be noted that the addition step according to the invention does not consist in adding the complete spectrum of the additional signal to the complete spectrum of the coded source signal. On the other hand, only the sub-carriers of the spectrum of the additional signal corresponding frequently to the empty subcarriers of the coded source signal are summed two by two. Consequently, nothing is added to the N sub-carriers of origin, called N useful subcarriers, of the source signal carrying the data to be transmitted.

Thus, the compact envelope signal according to the invention comprises according to the invention a number N _t of subcarriers greater than the source signal, ie N _t > N.

We therefore speak of "sub-carrier partial response coding" according to the invention. Such a "subcarrier" application is applied after parallel serial conversion of the source signal.

Thus, the sub-carrier partial response coding according to the invention makes it possible to reserve a spectral space for each sub-carrier to which it is applied, this spectral space corresponding to the location of sub-carriers dedicated to the addition of signal corrector according to an alternative to the TR method. In fact, the implementation of the TR technique as provided for in the DVB-T2 broadcasting standard corresponds to reserving a number P of subcarriers equal to approximately 1% of the number N of the sub-carriers for the reduction of PAPR.

On the contrary, the present invention proposes to increase the number of reserved subcarriers by means of a partial response coding systematically.

Thus, with regard to the combination of the technique of adding signal and the technique TR of the prior art, the implementation according to the invention of the combination of the sub-carrier partial response coding and the addition of signal increases the reduction of the PAPR, because it is possible to add a correction signal to a larger number of empty subcarriers.

The reduction of the PAPR according to the invention is therefore increased. In addition, the technique according to the invention has a low complexity and makes it possible to limit the decrease in spectral efficiency because the empty subcarriers are created "from scratch" without reducing the number of useful subcarriers of the signal. source.

In fact, conversely, the technique of the previously mentioned prior art of adding a signal combined with the T technique, reserves P subcarriers which leads to a reduction in the number of useful subcarriers, and consequently a decrease in spectral efficiency.

According to a particular example of the invention, the partial response coding is a duo-binary coding.

Advantageously, the method also comprises a digital filtering step so as to eliminate the part of the additional signal that interferes with said source signal.

In other words, such a filtering makes it possible to eliminate the frequency subcarriers of the compact envelope signal that does not belong to the frequency band allocated to the input source signal. For example, such a filtering corresponds to the implementation of a conventional digital filtering.

According to a first embodiment of the invention, the partial response coding delivers in the frequency domain:

N subcarriers, called N subcarriers useful for carrying the data of the source signal, and

• N subcarriers, called N empty subcarriers, associated with the empty coded part of the coded source signal to which the additional signal is added.

In other words, for each of the N useful subcarriers of the signal, the partial response coding creates an empty subcarrier. Thus, there are as many partial response coding steps as there are subcarriers in the source signal, namely N steps, each of these N steps delivering two subcarriers: the useful subcarrier from the source signal and the empty subcarrier created by the partial response coding.

Thus, the PAPR is reduced on as many empty subcarriers (N) as there are no useful subcarriers (N). Thus, the coded signal delivered at the output of the partial response coding, for example a duo-binary coding, comprises 2N subcarriers, that is to say twice as many subcarriers.

It should also be noted that the coded signal outputted from the partial response coding may comprise N _t > 2N subcarriers. For example, N _t comprises N useful subcarriers, N empty subcarriers and N _a subcarriers dedicated to other functionalities such as, for example, channel estimation, synchronization, or even rate increase.

For example, according to the 802.11a standard implemented in WiFi networks alone fifty two of sixty four subcarriers are used. The other twelve subcarriers are set to zero and dedicated to other functionalities.

Advantageously, according to a second embodiment of the invention, the partial response coding delivers in the frequency domain:

• N sub-carriers, called useful sub-carriers, intended to carry the data of the source signal, and

• M subcarriers, such that P <M≤N, said empty subcarriers, P corresponding to a predetermined number of subcarriers, said M subcarrier voids being associated with said empty coded part of the coded source signal to which said additional signal is added.

According to this approach, it is possible to apply a "selective" partial response coding. In other words for N useful subcarriers of the source signal, it is possible to create fewer empty subcarriers than useful subcarriers while having a number of empty subcarriers greater than the number P of subcarriers for example reserved according to the TR technique. Thus, M empty sub-carriers are created by partial response coding of M useful subcarriers, for example dubinary coding, and (N-M) useful subcarriers are conventionally encoded in a binary manner. The conventional binary coding of useful (N-M) subcarriers therefore does not create any associated vacuum subcarriers.

The application of a "selective" partial response coding thus makes it possible to obtain higher PAPR reduction performances than those relating to the technique of the aforementioned prior art of adding a signal combined with the TR technique, while by making it possible to limit the coding complexity, the number of empty subcarriers created by the coding being less than the number of useful subcarriers.

For example, according to this embodiment, it is possible to create by partial response coding M empty subcarriers only for the central subcarriers M of the source signal. Each of the M empty subcarriers obtained is delivered by applying a partial response coding to a subcarrier.

It should be noted that in the case where M = N, we obtain the same result as for the first embodiment, namely as many empty sub-carriers created by partial response coding as useful subcarriers of the source signal and consequently an equivalent coding complexity.

According to a first variant, such selective partial response coding can be implemented in a predetermined manner according to a fixed selection criterion known to the transmitter and the receiver.

According to a second variant, such selective partial response coding may be configurable and adjusted by the user. In this case, the user can act and vary the value of the parameter M as well as the location of the useful M subcarriers of the source signal which will be used by the partial response coder to create the empty subcarrier M intended to to carry the additional signal.

According to this second variant, it is nevertheless necessary to transmit to the receiver the value of the parameter M and the location of the useful sub-carriers M of the source signal which will be used by the partial response coder to create the empty subcarrier M intended for carry the additional signal.

According to a third embodiment of the invention, the partial response coding delivers in the frequency domain:

• M subcarriers comprising Mi subcarriers dedicated to a predetermined function and M ₂ subcarriers such that P <M ₂ , said subcarriers empty, P corresponding to a predetermined number of subcarriers, said sub M ₂ sub-carriers empty carriers being associated with said empty coded part to which said additional signal is added.

According to this third embodiment, the M subcarriers created by partial response coding comprise both sub-carriers dedicated to a predetermined function, such as for example a channel estimate, a synchronization, or the increase and subcarrier M ₂ such that P <M ₂ , said empty subcarriers associated with said empty coded part to which said additional signal is added.

According to this embodiment, it is possible, for example, for M to be greater than N. It is also possible for M _{2 to} be greater than N.

The advantage of this embodiment is that it makes it possible to create empty subcarriers to reduce the PAPR and for at least one other function with respect to the original source signal. This third embodiment of the invention therefore offers many possibilities at the cost of increased complexity because of the fact that an additional parameter Mi is taken into account whose value corresponds to the number of subcarriers dedicated to at least one other function with regard to the original source signal.

According to a first particular approach of the invention, the addition step is implemented in the time domain after application of an inverse Fourier transform of size N + M to the additional signal and application of an inverse Fourier transform of size N to source coded signal.

Indeed, at the end of the partial response coding step of each subcarrier of at least a part of said N subcarriers of said source signal, M empty subcarriers are delivered according to this embodiment. Since the value of these empty subcarriers is zero, the size of the inverse Fourier transform applied to the coded source signal is equal to N.

On the other hand, once the addition of the additional signal is carried out, the value of each of the M sub-carriers empty created by partial response coding goes from zero to the value of the corresponding frequency subcarrier of the additional signal, it is therefore necessary to implement an inverse Fourier transform of size N + M.

Optionally, the inverse Fourier transform is a fast inverse Fourier transform (IFFT).

According to a second particular approach of the invention, the addition step is implemented in the frequency domain before applying an inverse Fourier transform of size N + M to the compact envelope signal.

The advantage of this approach is to reduce the number of inverse Fourier transform operations which reduces the complexity of implementation of the invention.

The additional signal can be generated from the coded source signal from the partial response coding step by means of a multitude of generation methods.

According to a first preferred embodiment option, the step of generating the additional signal implements a reduction of the peak-to-average power ratio of said source signal by a limitation also called "clipping", said limitation consisting in clipping the amplitude of the signal when it exceeds a predetermined value.

The use of this method of "clipping" has the advantage of being simple and inexpensive to implement.

According to a second embodiment option, the step of generating the additional signal for example implements a reduction of the peak to average power ratio of the source signal by applying a geometric method,

the additional signal c (t) obtained by the geometric method being defined by the following equation in the discrete domain:

with:

x _n the source signal,

n representing the index of a sample in the discrete domain,

At a maximum amplitude of said compact envelope signal

af

θ _η = 2π, with: Af the frequency offset between said source signal and said signal additional and B the bandwidth of said source signal.

The advantages of such an additional signal are notably described and evaluated in the document "Tone reservation technique based on geometry method for orthogonal frequency division multiplexing peak-to-average power ratio reduction" (D. Guel, J. Palicot, Y. Louët , IET communication, Volume 4, Issue 17, pages 2065-2073, November 26, 2010).

Advantageously, said geometric method implements a step of weighting by a real scalar β (° ^νΐ defined by the following equation:

with S _p = [n: \ y _n \> A], in other words the set of indices n such that \ y _n \> A where y _n - x _n + βο _η

and c _n

i _n with N _F the length of a digital filter used by said filtering step and h _n an index coefficient n of said filter.

Such an "improved" variant of the geometric method makes it possible to reduce the complexity implemented according to the "classical" geometric method. A comparison of the classical geometric method and the "improved" geometrical method is presented in particular in chapter 5 of the document "Study of new techniques of reduction of the" backward compatible "crest factor for multicarrier systems" (Guel, D. PhD thesis, November 2009).

It should be noted that the additional signal can be generated in multiple alternative ways to the geometric method described above. Various techniques of signal addition are described in particular in the thesis indicated above. For example, the technique of adding a signal, defined in EP 219,333 by a decreasing function when the signal amplitude increases, can also be considered.

In one example, the modulation corresponds to a partial response modulation.

Among the partial response modulations, there may be mentioned, for example, the modulation of two dubinary quadrature channels, called QPRS modulation, of the same rate as the modulation.

MAQ4.

Advantageously, the steps of generating the additional signal, digital filtering and addition are iterated at least once.

In addition, the iterations of the additional signal generation, digital filtering and addition steps are interrupted when a variation below a predetermined threshold is reached.

For example, the threshold corresponds to a value between 0.1 and 0.5 dB, in particular 0.2 dB, of the peak to average power ratio.

Thus, the invention proposes an iterative implementation of the transmission method, which makes it possible to reduce PAP efficiently and quickly.

The invention also relates to a multicarrier transmission device of a signal representative of a source signal, intended to be amplified and then transmitted via a transmission channel. According to the invention, such a transmission device comprises:

partial response coding means applied to each subcarrier of at least a portion of said N subcarriers of the source signal, creating for each subcarrier to which said partial response coding is applied an empty spectral space, corresponding to at least one empty subcarrier, all of said empty subcarriers forming an empty coded part of said coded source signal,

means for generating a predetermined signal from the coded source signal, said additional signal;

means for adding to the empty coded part of the coded source signal a spectral portion of said additional signal corresponding to said empty coded part of said coded source signal, so as to obtain a signal comprising N _t > N subcarriers and presenting a peak power to average power ratio less than the average power peak power ratio of said source signal, said compact envelope signal;

means for modulating a carrier wave by said compact envelope signal, delivering a modulated compact envelope signal;

amplification means of said compact envelope modulated signal, with the aid of a power amplifier, delivering an amplified signal;

means for transmitting said amplified signal.

The advantages and the embodiments described with regard to the transmission method are also applicable to the transmission device according to the invention. In particular, the partial response coding means correspond to as many partial response coding modules as there is a subcarrier in the part of the source signal to which the partial response coding is applied.

The invention also relates to a signal representative of a source signal comprising N sub-carriers, intended to be amplified and then transmitted according to the transmission method described above, the signal having a peak power ratio at average power less than the peak power to power ratio. mean of a source signal from which said amplified signal was constructed and transmitted by the steps of: partial response coding applied to each subcarrier of at least a portion of said N subcarriers of said source signal, creating for each subcarrier to which said partial response coding is applied an empty spectral space, corresponding to at least one an empty subcarrier, all of said empty subcarriers forming an empty coded part,

amplification of said compact envelope modulated signal, using a power amplifier, delivering said amplified signal;

transmitting said amplified signal.

It should be noted that in the optional case of a selective partial response coding creating M empty sub-carriers, such as P <M≤N, said empty sub-carriers and according to the second variant mentioned above where such a response coding partial differential can be parameterizable and adjusted by the user, the value of the parameter M and / or the location of the useful M subcarriers of the source signal which will be used by the partial response coder / decoder to create the sub-M sub-carriers. empty carriers for carrying the additional signal are also inserted in the signal according to the invention.

The invention also relates to a method for receiving the signal described above said signal according to the invention having been obtained from a source signal comprising N subcarriers. According to the invention, the reception method comprises the following steps:

reception of the signal, said signal received,

demodulating the received signal, delivering a demodulated received signal comprising N subcarriers,

partial response decoding applied to each subcarrier of said N subcarriers of said demodulated received signal, delivering a decoded signal.

Thus, the reception method comprises reciprocal steps from those of the transmission method. More precisely, on transmission, the signal addition step, as such, makes it possible to reduce the PAPR but does not require any particular treatment on reception. AT conversely, the sub-carrier partial response coding implemented on transmission requires a reciprocal sub-carrier partial response decoding step on reception.

Thus, on reception, partial response decoding makes it possible to detect the useful subcarriers of the signal. The empty subcarriers only dedicated to the reduction of the PAP during the OFDM transmission are indeed not taken into account during the reception because of their orthogonality with the useful subcarriers.

The demodulation of the received signal implementing a Fourier transform and more particularly a Fast Fourier Transform (FFT) of size N (N being the number of subcarriers of the coded and then amplified source signal according to the invention), delivers N under in parallel to the input of the partial response decoding which is applied individually to each sub-carrier.

It should be noted that in the optional case of an implementation on transmission of a selective partial response coding creating M subcarriers, such as P <M≤N, said empty subcarriers and according to the second previously mentioned variant where such selective partial response coding may be parameterizable and adjusted by the user, the value of the parameter M and / or the location of the useful M subcarriers of the source signal which will be used by the coder / decoder partial response to create the empty subcarrier M to carry the additional signal are transmitted and used at the reception to adjust the partial response decoding step.

Advantageously, the decoding step implements a Viterbi decoding. In other words, the Viterbi algorithm is used to overcome the redundancy generated by the partial response coding used during the transmission.

Indeed, the partial response coding implemented by the transmission method creates a new coding level resulting in the distance to the decision thresholds being lower and the noise resistance lower as well, the error rate being will thus be degraded compared to the binary.

Viterbi decoding on reception therefore provides a solution to this drawback of the partial response coding implemented on transmission.

Indeed, it has for example been demonstrated in the prior art that asymmetrically and at the cost of low complexity binary and duo-binary coding achieve the same error rate.

Furthermore, on reception, it is necessary to implement a partial response demodulation, for example a demodulation of two quadrature duo-binary channels, called QPRS demodulation reciprocally QPRS modulation implemented for example during transmission. The invention also relates to a device for receiving a signal according to the invention as previously described, said signal according to the invention having been obtained from a source signal comprising N subcarriers. According to the invention, the receiving device comprises: means for receiving said signal, said signal received,

means for demodulating said received signal, delivering a demodulated received signal comprising N subcarriers,

a partial response decoder assigned to each subcarrier of said N subcarriers of said demodulated received signal, delivering a decoded signal.

Thus, the invention requires a simple modification of existing receivers based on the addition of a partial response decoder.

Advantageously, the decoder also comprises a Viterbi decoding module.

The invention also relates to a computer program product downloadable from a communication network and / or recorded on a computer readable medium and / or executable by a processor. According to the invention, said computer program product includes program code instructions for implementing the transmission method and / or the reception method described above, when executed on a computer.

4. List of figures

Other characteristics and advantages of the invention will emerge more clearly on reading the following description of a particular embodiment, given as a simple illustrative and nonlimiting example, and the appended drawings, among which:

FIGS. 1A and 1B, already com pleted in relation with the prior art, respectively illustrate the characteristics of a conventional amplifier and the respective envelopes of the source signal and the method of adding a signal by the "T technique" (FIG. of the English "Tone Reservation");

Figure 2 illustrates the main steps of a transmission method according to the invention;

FIG. 3 illustrates the power spectral density of the binary and the ubinary signals; FIG. 4 is a block diagram of an exemplary transmitting device according to a first embodiment of the invention;

Figure 5 is a block diagram of an exemplary transm ission device according to a second embodiment of the invention;

FIGS. 6A to 6C illustrate the problem of reducing the PAPR and the steps of implementing a geometric method for generating the additional signal in two variants;

FIGS. 7A to 7C respectively illustrate an example of constellations of modulations implemented according to one embodiment of the invention;

FIG. 8 illustrates the main steps of a reception method according to the invention; Figure 9 is a block diagram of an exemplary receiving device according to one embodiment of the invention.

5. Description of an embodiment of the invention

5.1 General principle

The invention is therefore based on the specific combination of the partial response coding, for example a dubinary coding, the source signal and the addition of an additional signal reducing the PAPR on the empty subcarriers created by coding duo. binary.

This avoids the introduction of noise on subcarriers of data called useful subcarriers.

According to the invention, an improvement in the reduction of the PAPR of the source signal is obtained with regard to the existing combination of the signal addition technique and the TR technique.

Indeed, in comparison with the TR technique, the partial response coding makes it possible to create "from scratch" and not to reserve empty subcarriers dedicated to the reduction of the PAPR.

In addition, with regard to the TR technique as provided for in the DVB-T2 broadcast standard of reserving a number P of subcarriers equal to about 1% of the number N of the subcarriers for the reduction of PAPR, the technique according to the invention makes it possible to obtain subcarriers created specifically for the reduction of PAPR and whose number is strictly greater than P.

This gives an increase in the reduction of the PAPR with regard to the combination of the prior art.

It should also be noted that the present invention proposes a new waveform based on a partial response coding performed for example on each sub-carrier. A different spectrum of the signal obtained according to the prior art is thus delivered according to the invention.

5.2 Detailed description and implementation of the different steps of the transmission method according to the invention.

The following are described below with reference to FIG. 2, the main steps implemented for the reduction of the PAPR according to the invention.

As illustrated in FIG. 2, the multicarrier transmission of a signal representative of a source signal x, intended to be amplified by a power amplifier and then transmitted via a transmission channel, comprises the following steps:

partial response coding applied to each subcarrier of at least a portion of said N subcarriers of the source signal x, creating for each subcarrier to which said partial response coding is applied an empty spectral space, corresponding to the at least one empty subcarrier, all of said empty subcarriers forming an empty coded portion;

generation 21 of a predetermined signal from the coded source signal, said additional signal c;

digital filtering 210 of said additional signal, delivering a filtered additional signal c; adding to the sub-carriers of a predetermined subcarrier set of said coded source signal x _db , a spectral portion of said additional signal corresponding to said empty coded part of said filtered coded source signal c, so as to obtain a signal comprising N _t > N subcarriers and having a peak power to average power ratio less than the average power peak power ratio of said source signal, said compact envelope signal x _ep ;

modulating a carrier wave by said compact envelope signal x _ep , delivering a modulated compact envelope signal x _mep ;

amplification 24 of said modulated compact envelope signal x _mep , with the aid of a power amplifier, delivering an amplified signal x _a ;

transmitting said amplified signal x _a .

A) partial response coding

More precisely, the following table gives the equations of different partial response codes (20). Part response coding includes dumb coding and QPRS duo-binary coding, the respective equations of which are given in the first and last line of the table below:

Each symbol carries a bit, it is obtained by the equivalent of a convolutional coding between two consecutive bits. It's really about creating an interference between two consecutive symbols, voluntarily created at the show, which means that the duo-binary code does not meet the Nyquist criterion. The redundancy generated by the code is not, as conventionally, an addition of redundancy bits, but is found in the signal by the addition of a third level.

The duo-binary signal for example can also be considered as a transfer function filtering: H (z) = 1 + z ^-1 .

The formatting operation is then performed by a rectangular filter g (t) of width T. With the following notations, b _k {0, 1}, i _k = 2b _k - 1 {- 1, +1}, ^c k ⁼ ife + ifc-i { ^- 2.0 + 2} and a _k = Ac _k , the signal is written:

s (t) = Σ _k a _k Tg (t - kT) where a E [-2A, 0,2A} and g (t) = ^ Rect _{[0 T]} (t).

The main advantage of duo-binary coding lies in the fact that the spectrum occupied by the duo-binary signal 32 is twice as small as the binary spectrum 31 as illustrated by FIG. 3 representing the spectral power densities of the signals. binary and duo-binaries. The spectrum occupied by the duo-binary signal 32 is in particular equivalent to a binary spectrum with a Nyquist roll-off filter equal to zero, such a Nyquist roll-off filter equal to zero being impractical in practice.

Thus, with regard to FIG. 3, particular mention is made of the creation of an empty part 33 used according to the invention to add the additional signal enabling the reduction of PAP.

The disadvantage of duo-binary coding is that a new level is created by the coding, therefore the distance to the decision thresholds will be lower and the noise resistance lower as well, the error rate will be found therefore degraded compared to the binary.

However, to overcome this drawback, it is possible to manage this redundancy generated by the code by means of a Viterbi algorithm on reception, as described hereinafter with regard to the reception method according to the invention. Then, asymptotically, thanks to this Viterbi decoder, at the cost of a low complexity, the two codings reach the same rate of error.

According to the invention, the duo-binary coding releasing half (33) of the original spectrum on each sub-carrier as illustrated in FIG. 3 is advantageously implemented in order to use the freed space (33) in the spectrum to add an additional signal (thus as many additional subcarriers as useful subcarriers according to the embodiment of Figure 3) to minimize the PAPR.

Subsequently, the invention is described using duo-binary coding in particular. It is obvious that the set of implementation details described with regard to the duo-binary coding are transposable to any other partial response coding such as those illustrated for example by the above table or other partial response codes not reproduced in this table but well known to those skilled in the art.

As indicated above, according to a first embodiment of the invention, the partial response coding and for example the duo-binary coding delivers in the frequency domain:

• N subcarriers, called N empty subcarriers, associated with the empty dubinary part to which the additional signal is added.

In other words, for each of the N useful subcarriers of the signal, the dumbinary coding creates corresponding to an empty spectral space called an empty subcarrier (33), as illustrated in FIG.

According to a second embodiment of the invention, the duo-binary coding delivers in the frequency domain:

Subcarrier M, such that P <M≤N, said empty subcarriers, P corresponding to a predetermined number of subcarriers, said empty M subcarriers being associated with said empty dubinary part at which said signal additional is added.

According to this approach, it is possible to apply a "partial" duo-binary coding. In other words for N useful subcarriers of the source signal, it is possible to create fewer empty subcarriers than useful subcarriers while having a number of empty subcarriers greater than the number P of subcarriers reserved according to the technique T.

For example, according to this embodiment, it is possible to create by dumbinary coding M empty subcarriers only for the central subcarriers M of the source signal.

In another example, the empty M subcarriers can be obtained by dubinary coding of the M subcarriers useful of the source signal that have the highest PAPR.

B) Generation

From the coded signal and for example a dubinary source signal x _db delivered by the partial response coding step corresponding in this example to a duobinary coding of the source signal x, an additional signal is generated 21 in order to reduce the PAPR of the signal to be transmitted. In other words, the addition step according to the invention does not consist in adding the complete spectrum of the additional signal to the complete spectrum of the coded source signal. On the other hand, only the sub-carriers of the spectrum of the additional signal corresponding frequently to the empty subcarriers of the coded source signal are summed two by two. As a result, nothing is added to N the original subcarriers, called useful subcarriers, of the source signal carrying the data to be transmitted. More specifically, the additional signal can be generated in multiple ways. Various techniques of signal addition are described in particular in the previously cited document: "Study of new techniques for reducing the" crest factor "with backward compatibility for multicarrier systems" (D. Guel, Thesis of Doctorate, November 2009).

Preferably, the present invention implements an additional signal generated by a "clipping" method as described in more detail below and illustrated in FIGS. 4 and 5.

The use of the "clipping" technique 430 is indeed simple and inexpensive to implement.

Another example of an additional signal that can be used in the invention in combination with the duo-binary coding, corresponds to the additional signal of patent EP 219 333. According to this example, the additional signal is defined by a decreasing function approximated by at least a segment of the line. In particular, an example of this type of signal has a decreasing function corresponding to a hyperbolic secant function.

This additional signal can be represented as a discrete symbol in the time domain such as:

c _n = f (r _n ) e ^ - x _n

with f as previously defined:

2A

f {r) = A dry Ιι {ψ) = - -

e ^l + e ¹

where: A represents the control parameter of the average output power, in other words the maximum amplitude of said compact envelope signal,

Tj represents the performance parameter representing the reduction gain of the peak to average power ratio,

r the amplitude of the source signal.

Thus, according to this particular example, the generation step 21 of the additional signal delivers a symbol C _k .

The set of symbols C _k forms the additional signal C (f) in the domain frequency corresponding to the signal c (t) in the time domain.

According to another example, the present invention implements a geometric method for generating the additional signal.

Such a geometric method is described in detail in the document "Tone reservation technique based on geometry method for orthogonal frequency division multiplexing peak-to-average power ratio reduction" (Guel D., J. Palicot, Y. Louet, IET communication, Volume 4, Issue 17, pages 2065-2073, November 26, 2010).

This document notably describes two variants of the geometric method, namely the "classical" geometrical method implemented by the transmitter illustrated in FIG. 6A and the "improved" geometrical method implemented by the transmitter illustrated in FIG. 6C.

More specifically, the geometric method aims to reduce the complex envelope of the multicarrier signal x just before the power amplifier 601 by adding 602 a complex signal to the baseband signal of the multicarrier signal.

The principle of the technique consists first in generating an "artificial signal" a which is then modulated in intermediate frequency Af to give rise to the "additional signal" c which is in principle outside the useful band of the multicarrier signal x. Finally, c is added to x so as to considerably reduce the PAP of the resulting signal y = x + c. Just after amplification, the "additional signal" is eliminated by passive analog bandpass filtering.

The multicarrier baseband signal x (t) can be decomposed in phase and in quadrature as follows:

x (t) = I (t) + jQ (t),

where / (t) is the in-phase component and Q (t) is the quadrature component.

The complex envelope of the multicarrier baseband signal x (t) is written:

r (t) = | x (t) | = ² (t) + Ç ² (t)

The decomposition of the "artificial signal" has in component in phase and component in quadrature can be written:

a (t) = I _a (t) + jQ _a (t)

The calculation principle of the "artificial signal" is as follows:

Let A be the maximum allowable amplitude (also called the amplitude threshold of the source signal or the maximum amplitude of the compact envelope signal).

(i) If r (t)> A, the "artificial signal" a (t) is generated so that:

+ a (t) \ ² = \ l (t) + l _a (t) \ ² + \ Q (t) + Q _a (t) \ ² = A ² .

(ii) Otherwise, i.e., if r (t) <A, the "artificial signal" a is equal to zero.

In the case where r (t)> A, the / _a (t) and _Q (t) can be determined geometrically by considering 0≤ T≤ Ts and (l / Q) the plane defined by Figure 6B. In the plane defined by Figure 6B, (ξ) is the circle of center O and radius A, X _c is the symmetric of X with respect to O, (ξ _α ) is the circle of center X _c and radius A , and the angles φ and a are defined such that φ = (θΊ, C), a = (ÔX, ÔA?)

Let OX be the vector associated with z _x = re ^ ^v = I + jQ, or let OM be the vector associated with z _M = Ae ^ ^{a + (p} ^ and OA the vector associated with z _A = I _a + jQ _a

Solve the equation:

= A ² consists in finding AG (l / Q) such that:

\ OX + OA \ ² = A ² (/ + l _a ) ² + (Q + Q _a ) ² = A ² Coordinate points A describe the circle (ξε).

From the relation OA = OM - OX, we deduce that z _A = z _M - z _x , and substituting z _M = ~ l} ¾

This last relation is verified by an infinity of points, each point being characterized by an appropriate choice of a G [0, 2π [.

In so-called signal addition techniques, the variation (increase or decrease) of the average power has a strong impact on the quality of the transmission.

Taking this aspect into account, an "artificial signal" with the least additional power is preferably used.

In this context, the "artificial signal" with the least additional power corresponds to the vector OA with the smallest possible module, that is to say that a = 0 which corresponds to z _x .

The analytical expressions of I _a and Q _a are obtained by taking the real part and the imaginary part of z _A :

The "additional signal" c is written as follows:

a (opt) _([) ejnAft ₌ | / (^) _{(t) +} j _Q (opt) _([) ejnAft _with j (opt) ^ _Q (opt) ^ as expressed above.

The implementation of the "classical" geometric method described above implements, on the one hand, an initialization step executed once in order to set the parameters A and Af of the geometric method and, on the other hand, a step of which includes the following substeps: (i) Calculate the time OFDM signal x _n = r _n in

(ii) j or

(iii) Calculate the resulting signal y _n = x _n + c _n .

(iv) Transpose y _n to radio frequency and amplify it. After amplification, eliminate the "additional signal" by radiofrequency filtering so as to transmit only the amplified multicarrier signal.

With regard to the "classical" geometric method, it has been determined that the reduction gain of the PAP is important when the additional signal is generated close to the bandwidth of the multicarrier signal. However, when the additional signal is generated close to the bandwidth of the multicarrier signal, that is to say, when the parameter | Af | is close to zero, the analog filter placed just after the power amplifier no longer makes sense.

Indeed, this filter has all its interest when | Af | is great because it eliminates the "additional signal" after amplification and let only propagate the amplified multicarrier signal.

As a result, the "improved" geometrical method implemented by the transmitter illustrated in FIG. 6C proposes to eliminate the analog filter that exists in the conventional geometric method and to choose Af = 0 to reduce the PAPR.

When Af = 0, the distortions generated are significant (which does not result in a sustained degradation of the BER), to remedy this, the "improved" geometric method also proposes to filter 61 (by a digital filter of "high-pass" type ) the part of the "additional signal" that interferes with the multicarrier signal.

By filtering the "additional signal", the phenomenon of "peak-regrowth" occurs, which results in a loss of a part of the useful signal which serves to reduce the PAPR and consequently a decrease in the reduction gain of the PAPR. To limit this loss of performance, in the improved geometric method a weighting 603 of the additional signal c by a real scalar β ⁽ ° ^νΐ obtained by optimization 62 is implemented before adding 602 the additional signal to the multicarrier signal x. the real scalar β ⁽ ° ^νΐ calculated, an iterative process is implemented to increase the reduction gain of the PAPR. Digital filtering 61 consists in eliminating the part of the "additional signal" which interferes with the multicarrier signal. The digital filter which is of the "high pass" type therefore consists in making the signals c and x orthogonal (since they are disjoint in the frequency domain).

The realization of the digital filter 61 is performed transversely (non-recursively). More precisely, let N _{F be} the length of the filter and h _k , 0 _{k k} N N _F the coefficients of the filter, the signal c _n at the output of the filter is written: c _n = Σ¾ ₀ ¹ c _n _ _k h _not

The digital filter 61 attenuates the frequencies below the cutoff frequency f _c + - and this, in order to keep only the high frequencies of the additional signal.

With regard to FIG. 6C, the module 62 "OPT" is used to calculate the value of β (° ρ ^ί ). In fact, to increase the reduction of the PAP, it is necessary to solve the problem of convex optimization below. :

with S _v = {: \ y _n \> A] where y _n = x _n + βο _η

and c _n ⁼ Σ / co ¹ Cn-k h _n with N _F the length of a digital filter used by said filtering step and h _n an index coefficient n of said filter.

This minimization problem is nothing more than a linear least squares problem whose solution? ( ^Opt ) is given by:

The implementation of the "improved" geometrical method described above implements on the one hand an initialization step executed once in order to fix the parameters A and Af = 0 of the geometric method and on the other hand a execution stage which includes the following sub-steps:

(i) Calculate the time OFDM signal x _n = r _n e ⁿ , initialize and set i = 0,

Calculate the artificial signal a. ⁽ 0

n using the relationship

Calculate the "filtered additional signal" c _n = Σ ^ ₀ ¹ c "_ _fe h _n (iii) Calculate by applying = - ^p "" as specified above

LnESp \ ^c n \

(iv) Update the algorithm as follows: χ _η ^ ^{ι + ν>} = x _n ^ + β ^ ° ^ν ^ c _n ^

(v) Increment i and go to step (ii) if the maximum number of iterations N _iter is not reached. Otherwise, do y _n = x _n ^ and stop the execution.

C) Addition and filtering of the additional signal

Once the additional signal is generated from the dubinary source signal x _db delivered by the dubinary coding step 20 of the source signal x, the addition of the dubinary signal and the additional signal is performed. implemented.

Two embodiments, respectively illustrated in FIGS. 4 and 5, are described below and differ in particular in their addition and additional filtering steps.

In relation to FIG. 4, a block diagram of an example of a transmission device 40 according to a first embodiment of the invention is presented in an OFDM signal context.

The set of symbols X _n forms the source signal X (f) comprising N sub-carriers in the frequency domain and corresponding to the signal x (t) in the time domain.

The transmission device 40 firstly comprises a duodinary coder 41 encoding the source signal by means of a duo-binary code as described above. Such a dubinary coder is a partial response coding "per carrier" as previously written. When such a "per carrier" application is implemented, each carrier to which partial response coding is applied is derived from a prior parallel serial conversion step of the source signal, implemented by a serial-parallel converter (400). so that each carrier to be processed can be individually coded by the partial response coding according to the invention.

There are therefore as many partial response codings implemented in parallel that there are source subcarriers to be processed.

The device further comprises a module 42 making it possible to transform the dubinary source signal in the time domain by means of an inverse Fourier transform and more particularly of a fast inverse Fourier transform of size N.

From the temporal duo-binary source signal, generation means 43 generates an additional signal c (t).

As indicated, this additional signal can be generated by means of a multitude of generation methods, among which are found in particular the "classical" and "improved" geometrical methods described above, the use of a descending hyperbolic secant function such as as described in patent EP 219 333 or preferably the use of the technique of "clipping" 430, or limiter as shown in Figure 4.

In relation to FIG. 4, the additional signal is then digitally filtered 44 by application:

of a Fourier transform 441 and more particularly of a fast Fourier transform (FFT) of size N + M, with P <M≤N, P corresponding to a predetermined number of subcarriers, for example equal to the number of sub-carriers reserved according to the technique T,

a filtering in the frequency domain 440,

of an inverse Fourier transform 442 and more particularly of a fast inverse Fourier transform of size N + M.

According to the invention, the filtering 440 is selected in order to effectively limit the noise generated by the clipping technique and thus avoid degradation of the BER.

Once this filtered additional signal c (t) is obtained, it is added to the dubinary source signal. These operations, additional signal generation, addition and filtering are iterated at least twice, until a variation of the PAPR for example less than 0.2 dB.

At the output of the addition operation 22, the compact envelope signal is obtained

Xep (t).

The compact envelope signal x _ep (t) is then used to modulate a carrier wave of frequency f _c delivering a modulated compact envelope signal x _mep (t). This modulation is a multiplication of the compact envelope signal by a carrier wave, which amounts to transposing at a higher (or lower) frequency the compact envelope signal x _mep (t).

The modulated compact envelope signal x _mep (t) is then amplified using a power amplifier 45, delivering an amplified signal x _a (t) and then emitted by a transmitting antenna 46.

According to FIG. 5, the transmission device 50 differs according to a second approach of the invention from the transmission device 40 of FIG. 4 in that the addition and filtering steps of the additional signal are modified.

Indeed, according to this second approach, the duo-binary and additional source signals are added in the frequency domain, and more precisely at the input of a inverse Fourier transform 51, and not in the time domain as shown in FIG. 4.

More precisely, the duo-binary source signal in the frequency domain is directly used to generate the additional signal.

The additional signal is also delivered by the generation means 43 in the frequency domain by applying a Fourier transform 441 and more particularly a fast Fourier transform (FFT) of size N + M, with P <M≤N , P corresponding to a predetermined number of subcarriers, for example equal to the number of sub-carriers reserved according to the T technique, and filtered (440).

The source and duo-binary signals are then added in the frequency domain, the additional signal reducing the PAPR being added to each empty subcarrier created by dubinary coding of the source signal. Thus, FIG. 5 illustrates "interleaving" of the useful subcarriers of the duo-binary source signal carrying the data symbols X ₀ to X _n (solid line) on the one hand, and empty subcarriers thereof. ci on which the correction symbols X ₀ to X _M of the additional signal have been added (in dashed line) on the other hand at the input of an inverse Fourier transform 51 and more particularly of a fast inverse inverse Fourier transform N + M.

It should be noted that during the first iteration of the iterative process, the empty subcarriers created by partial response coding, for example a dubinary coding, are set to zero.

In the case where N = M, there are in this case as many useful subcarriers of the duo-binary source signal as empty subcarriers on which the additional signal has been added and the corresponding inverse fast Fourier transform size is equal to 2N.

Thus, the approach illustrated in FIG. 5, implementing an addition in the frequency domain, makes it possible to "save" a module / an inverse Fourier transform step of size N.

D) Modulation

According to the detailed example based on a partial response coding corresponding to a duo-binary coding, the invention implements, for example, a modulation of two dubinary quadrature channels, called QPRS modulation ("Quadrature"). Partial Response Signaling ").

Specifically, this modulation carries two bits per symbol, so it must be compared to the four-phase phase change modulation QPSK (English "Quadrature Phase-Shift Keying"). The QPRS constellation as shown in FIG. 7A comprises nine points instead of the four points of the QPSK constellation.

Many partial response modulations have been proposed to modulate the coded signals by means of a partial coding whose equations are in particular indicated in the table above.

FIG. 7B illustrates in particular the constellation of 81QPR modulation. This modulation carries 4 bits per symbol. From this point of view it has to be compared to the quadrature amplitude modulation with sixteen MAQ16 states.

According to another example of modulation adapted to partial response coding, FIG. 7C represents the 81/9 QPR constellation whose performance in terms of continuity of service are given in the document "Multiresolution broadcast using partial response modulation" (J. C. Roland, K. Berberidis, International Conference on Telecommunications (ICT 2005), May 2005, Capetown, South Africa.

5.3 Detailed description and implementation of the various steps of the reception method according to the invention.

As illustrated in FIG. 8, the multicarrier reception of a signal representative of a signal amplified and transmitted according to the transmission method described above, said signal according to the invention having been obtained from a source signal comprising N under -porteuses, requires a particular treatment that includes the following steps:

reception (81) of said signal, said received signal,

demodulating (82) said received signal, delivering a demodulated received signal comprising N subcarriers,

duodinary decoding (83) applied to each subcarrier of said N subcarriers of said demodulated received signal, providing a decoded signal.

Indeed, the reception method comprises reciprocal steps from those of the transmission method described above. More precisely, the duo-binary coding per subcarrier implemented on transmission requires a reciprocal step of duodinary decoding on reception.

Thus, on reception, partial response decoding makes it possible to detect the useful subcarriers of the signal. The empty subcarriers dedicated solely to reducing the PAPR during the OFDM transmission are indeed not taken into account during reception because of their orthogonality with the useful subcarriers.

The demodulation of the received signal implementing a Fourier transform and more particularly a Fast Fourier Transform (FFT) of size N (N being the number of subcarriers of the coded and then amplified source signal according to the invention), delivers N under -porteuses in parallel at the input of the partial response decoding which is applied individually to each sub-carrier. In a way, the decoding step also implements a decoding of Viterbi 830 in order to overcome the degradation of the error rate introduced by the application of a dubinary coding on transmission.

Furthermore, on reception, it is necessary to implement a demodulation 82 of two quadrature duo-binary channels called QP S demodulation reciprocally to the QPRS modulation implemented during transmission.

FIG. 9 illustrates a reception device 900 of a signal according to the invention as described above. According to the invention, the device for receiving a signal representative of a signal amplified and transmitted according to the transmission method described above, said signal according to the invention having been obtained from a source signal comprising N under - carriers, includes:

means for receiving (91) said signal, said signal received,

demodulation means (92) of said received signal, delivering a demodulated received signal comprising N subcarriers,

a duodinary decoder (93) applied to each subcarrier of said N subcarriers of said demodulated received signal, providing a decoded signal.

Preferably, the decoder also comprises a Viterbi 930 decoding module.

Claims

REVEN DICATIONS

A method for transmitting a signal representative of a source signal comprising N sub-carriers, intended to be amplified by a power amplifier and transmitted via a transmission channel,

characterized in that said method comprises the following steps:

partial response coding (20) applied to each subcarrier of at least a portion of said N subcarriers of said source signal, creating for each subcarrier to which said partial response coding is applied an empty spectral space, corresponding to at least one empty subcarrier, all of said empty subcarriers forming an empty coded part of said coded source signal,

generating (21) a predetermined signal from said coded source signal, said additional signal;

adding (22) to said empty coded part of said coded source signal a spectral portion of said additional signal corresponding to said empty coded part of said coded source signal so as to obtain a signal comprising N _t > N subcarriers and presenting a peak power to average power ratio lower than the average power peak power ratio of said source signal, said com plete envelope signal;

modulating (23) a carrier wave by said com plete signal, providing a compact envelope modulated signal;

amplifying (24) said compact envelope modulated signal, using a power amplifier, delivering an amplified signal;

em ission (25) of said amplified signal.

2. Method according to claim 1, characterized in that said partial response coding (20) is a duo-binary coding.

The method of claim 1 or claim 2, characterized in that said method also includes a digital filter step (210) so as to eliminate that portion of said additional signal that interferes with said source signal.

4. Method according to one of claims 1 to 3, characterized in that said partial response coding (20) delivers in the frequency domain:

N subcarriers, called N subcarriers useful for carrying the data of said source signal, and

• N subcarriers, called N empty subcarriers, associated with said empty coded part of said coded source signal to which said additional signal is added.

5. Method according to one of claims 1 to 3, characterized in that said partial response coding (20) delivers: • N sub-carriers, called useful sub-carriers, intended to carry the data of said source signal, and

• M subcarriers, such that P <M≤N, said empty subcarriers, P corresponding to a predetermined number of subcarriers, said empty subcarrier M being associated with said empty coded part of said coded source signal to which said additional signal is added.

6. Method according to claim 5, characterized in that said addition step is implemented in the time domain after application of an inverse Fourier transform of size N + M to said additional signal and application of a Fourier transform. inverse of size N to source coded signal.

7. Method according to claim 5, characterized in that said adding step is implemented in the frequency domain before applying an inverse Fourier transform of size N + M to said compact envelope signal.

The method as claimed in any one of the preceding claims, characterized in that said generation of said additional signal implements a reduction of the average power peak-to-peak ratio of said source signal by limitation, said limitation being to clip the amplitude of the signal when it exceeds a predetermined value.

9. The method as claimed in claim 1, characterized in that said generation of said additional signal implements a reduction of the average peak power to power ratio of said source signal by application of a geometric method,

said additional signal c (t) obtained by said geometric method being defined by the following equation in the discrete domain:

with:

x _n said source signal,

n representing the index of a sample in the discrete domain,

At a maximum amplitude of said compact envelope signal,

af

θ _η = 2π--, with: Af the frequency offset between said source signal and said additional signal and B the bandwidth of said source signal.

10. Method according to any one of the preceding claims, characterized in that said modulation corresponds to a partial response modulation.

11. The method of claim 3, characterized in that said generating steps said additional signal, digital filtering and addition are iterated at least once.

12. Apparatus for transm ission (40) m ultiporter of a signal representative of a source signal comprising N subcarriers, intended to be amplified and transmitted via a transm ission channel, characterized in that said transmitting device ission includes:

partial response coding means (41) applied to each subcarrier of at least a portion of said N subcarriers of said source signal, creating for each subcarrier to which said partial response coding is applied a space empty spectrum, corresponding to at least one empty subcarrier, all of said empty subcarriers forming an empty coded part of said coded source signal,

means for generating (42) a predetermined signal from said coded source signal, said additional signal;

means (43) for adding, to said empty coded part of said coded source signal, a spectral portion of said additional signal corresponding to said empty coded part of said coded source signal, so as to obtain a signal comprising N _t > N subcarriers and having a peak power to average power ratio not less than the peak power to average power ratio of said source signal, said com plete envelope signal;

means for modulating (44) a carrier wave by said com plete signal, providing a compacted envelope modulated signal;

amamping means (45) of said compacted-envelope modulated signal, with the aid of a power amplifier, delivering an amplified signal;

transmitting means (46) of said amplified signal.

A signal representative of an amplified signal transmitted according to the method of transm ission of any one of claims 1 to 11, said signal having a peak power to average power ratio lower than the peak power to average power ratio of a source signal comprising N subcarriers from which said amplified signal has been constructed and transmitted by the following steps:

partial response coding applied to each subcarrier of at least a portion of said N subcarriers of said source signal, creating for each subcarrier to which said partial response coding is applied an empty spectral space corresponding to least one empty subcarrier, the set of empty subcarrier gaps forming an empty coded part of said coded source signal,

adding to said empty coded part of said coded source signal a spectral portion said additional signal corresponding to said empty coded part of said coded source signal, so as to obtain a signal comprising N _t > N subcarriers and having a peak power ratio at average power less than the average power peak power ratio of said source signal, said compact envelope signal;

amplifying said compact envelope modulated signal, using a power amplifier, delivering said amplified signal;

transmitting said amplified signal.

A method of receiving a signal corresponding to a signal according to claim 13 obtained from a source signal comprising N sub-carriers, characterized in that said receiving method comprises the following steps:

receiving (81) said signal, said received signal,

partial response decoding (83) applied to each subcarrier of said N subcarriers of said demodulated received signal, providing a decoded signal.

15. Reception method according to claim 14, characterized in that said decoding step also implements Viterbi decoding (830).

16. Apparatus (900) for receiving a signal corresponding to a signal according to claim 13 obtained from a source signal comprising N sub-carriers, characterized in that said reception device comprises:

means for receiving (91) said signal, said signal received,

demodulation means (92) of said received signal, delivering a demodulated received signal comprising N subcarriers, - a partial response decoder (93) applied to each subcarrier of said N subcarriers of said demodulated received signal, delivering a signal decoded.

17. Receiving device according to claim 16, characterized in that said decoder comprises a Viterbi decoding module (930).

18. Computer program product downloadable from a communication network and / or recorded on a computer readable medium and / or executable by a processor, characterized in that it comprises program code instructions for the implementation of the transmission method according to at least one of claims 1 to 11 or at least one of claims 14 and 15 when executed on a computer.