Method and device for equalizing a transmission channel in a communication system.

Abstract

The invention relates to a method for equalizing an OFDM transmission channel on which a data frame is received from a transmitter, said frame comprising a plurality of binary words, a word being associated with a particular sequence of snippets obtained by cyclic shifting a root sequence, the method comprising steps of receiving (300) a first and at least a second OFDM symbol carrying respectively a first and a second sequence of snippets of said frame, of determining (301) a first Impulse Response of the channel by calculating a correlation between the first OFDM symbol received and the IDFT of the root sequence, and at least a second Impulse Response of the channel by calculating a correlation between the second OFDM symbol received and the IDFT of the root sequence,calculation (302) of a Discrete Fourier Transform of the first and second Impulse Response to obtain respectively a first estimation vector of channel H0 and a second estimation vector of channel H1, equalization (303) of the second OFDM symbol from the first vector H0, and equalization of the first symbol from the second vector H1, and demodulation (304) of the frame from the equalized symbols. Figure 3.

Description

Title of the invention: Method and device for equalizing a transmission channel in a communication system. Technical field

[0001] The invention belongs to the field of telecommunications and relates in particular to a method and a device for equalizing a transmission channel in an OFDM communication system. PREVIOUS ART

[0002] Low Power Wide Area Network (LPWAN) long-range communication networks are still a major subject of research and innovation today, driven by the deployment of satellite loT (Internet Of Things) networks, but also by the evolution of cellular loT markets with standards such as LTE-M, also called eMTC (enhanced Machine Type Communication), NB-IoT (Narrow Band loT) or LoRaWAN (long-range Wide Area Network).

[0003] Such communications are generally subject to the same constraints: a need to sporadically communicate small amounts of data over long distances, a very low link budget, and limited computing capabilities. Added to this is often the constraint of a finite energy source, as most of these objects operate on batteries.

[0004] OFDM (Orthogonal Frequency-Division Multiplexing) is a method of coding digital signals by orthogonal frequency division in the form of multiple subcarriers in which the signal to be transmitted is distributed over a large number of subcarriers. To overcome attenuation phenomena that may be due to a multipath channel, a guard interval is generally introduced between symbols. In the context of CP-OFDM modulation (Cyllic Prefix Orthogonal Frequency Division Multiplex), the guard interval is used to insert a cyclic prefix (CP) consisting of a copy of the end of a block of information to be transmitted at the beginning of the frame. The prefix thus acts as a buffer in the case of multipath transmission, which makes it possible to avoid inter-symbol interference (ISI).

[0005] A multipath channel can cause a fast fading phenomenon when the coherence bandwidth is smaller than the total bandwidth of the OFDM signal. As a result, the subcarriers are affected by a different channel. The effect of the channel on each subcarrier is modeled by a multiplication complex between the data symbol and a channel coefficient.

[0006] In order to take into account and limit the effect of these disturbances, it is necessary to estimate the channel and to equalize it. Conventionally, the channel estimation is carried out in frequency, from pilot symbols inserted by the transmitter. The transmission of pilot symbols known to the receiver on dedicated subcarriers allows the receiver to carry out an estimation of the channel in frequency at these locations. The channel on the data subcarriers is then estimated by interpolation. The receiver can then carry out a single-path (or one-tap) equalization on the basis of the channel estimation carried out.

[0007] Unfortunately, pilot symbols contribute to energy expenditure even though they do not carry useful data.

[0008] CCSK modulation (for Cyclic Code-Shift Keying in English) allows the transmission of pilot signals without additional cost. This modulation method proposes to modulate the data to be transmitted by the cyclic rotation / shift of a root sequence of complex symbols. The root sequence is such that its shifted versions are orthogonal to each other, that is to say that it offers a good autocorrelation function. Thus, each binary word to be transmitted is associated with a particular cyclic shift of the root sequence. Since the root sequence is known to the receiver, the transmitted sequences can serve as pilots for synchronization and channel estimation.

[0009] For example, a root sequence composed of 4 complex symbols [a; b; c; d] makes it possible to modulate 2-bit binary words. The word '00' is for example associated with a zero offset corresponding to the sequence [a; b; c; d], the word '01' with a one-position offset corresponding to the sequence [d; a; b; c], the word '10' with a two-position offset corresponding to the sequence [c; d; a; b] and the word '11' with a 3-position offset corresponding to the sequence [b; c; d; a].

[0010] At the receiver, demodulation is performed by cross-correlation with the root sequence. The location of the correlation maximum indicates the information modulated via inverse mapping. For example, to demodulate the sequence [d; a; b; c], the result of the cross-correlation with the root sequence [a; b; c; d] provides the normalized result [0; 1; 0; 0]. The correlation maximum at this index corresponds to the binary word '01'.

[0011] By using for CCSK modulation a root sequence which has a perfect autocorrelation function and whose inverse Fourier transform (IDFT, for Inverse Discrete Fourier Transform in English) also has good autocorrelation properties, it is possible to obtain an estimate of the channel impulse response (or CIR, for Channel Impulse Response) by correlating each OFDM symbol received with the IDFT of the root sequence. Indeed, the length of the cyclic prefix (or CP for Cyclic Prefix) OFDM being by construction of the telecommunications system always greater than the length of the CIR, each subcarrier undergoes a flat channel in frequency, corresponding to the Discrete Fourrier Transform (or DFT, for Discrete Fourrier Transform) of the CIR.

[0012] When a CCSK sequence of size N is transmitted in an OFDM symbol, a CIR of size N is then obtained, having undergone a cyclic shift linked to the cyclic shift of the CCSK sequence transmitted in the OFDM symbol.

[0013] Unfortunately, the CIR also undergoes additive white Gaussian noise, and it is not always possible to differentiate a channel coefficient from a noise coefficient, and therefore to determine the shift undergone by the CIR.

[0014] However, the DFT of an offset CIR cannot be directly used to equalize the subcarriers of the OFDM symbol in question. Indeed, the offset it has undergone is linked to the offset of the transmitted CCSK sequence. Consequently, if we attempt to frequency equalize the received sequence with such a channel estimation, the result of the correlation will always give a peak at index zero, because the modulated information will have been removed by the equalization.

[0015] Thus, there is a need for a method that makes it possible to improve the equalization of an OFDM transmission channel using CCSK modulation, without using dedicated pilot symbols. Summary of the invention

[0016] To this end, a method is proposed for equalizing an OFDM transmission channel on which a data frame is received from a transmitter, said frame comprising a plurality of binary words, a word being associated with a particular sequence of chips obtained by cyclic shifting of a root sequence, the method comprising the following steps: - Reception of a first and at least one second OFDM symbol respectively carrying a first and a second sequence of chips of said frame, - Determination of a first Impulse Response of the channel by calculating a correlation between the first OFDM symbol received and 1TDFT of the root sequence, and of at least a second Impulse Response of the channel by calculating a correlation between the second OFDM symbol received and 1TDFT of the root sequence, - Calculation of a Discrete Fourier Transform of the first and second Impulse Responses to obtain respectively a first estimation vector of the H0 channel and a second estimation vector of the Hl channel, - Equalization of the second OFDM symbol from the first HO vector, - Equalization of the first symbol from the second vector Hl, and - Demodulation of the frame from the equalized symbols.

[0017] It is thus proposed to equalize the first OFDM symbol from the CIR calculated from the second symbol and to equalize the second symbol from the CIR calculated from the first OFDM symbol. Indeed, each OFDM symbol comprising a CCSK sequence can be used to estimate the channel.

[0018] The method also makes it possible to obtain information on the relative cyclical offsets between sequences included in the OFDM symbols of the frame.

[0019] Indeed, let us assume that the sequence received in the first symbol has an index offset Po relative to the root sequence, and that the sequence received in the second symbol has an index offset Pb. The step of calculating a DTF makes it possible to obtain a channel estimate Ho for the first OFDM symbol, and Hi for the second OFDM symbol. Assuming that the channel evolves slowly over time, the channels undergone by successive OFDM symbols are strongly correlated on each subcarrier. By equalizing the second OFDM symbol with Ho and calculating a correlation maximum by CCSK demodulation, the correlation peak will be located at the index equal to Pi-P0 modulo N. We therefore obtain information on the relative offset between the sequence included in the second symbol and that included in the first symbol. Conversely, by equalizing the first zero symbol with Hb the correlation maximum will be located at P0-Pi modulo N.

[0020] The information obtained concerning the relative offsets between the sequences can be used to perform offset demodulation in order to find the transmitted symbols.

[0021] Thus, each OFDM symbol of the frame can be used to equalize all the other OFDM symbols of the frame, without it being necessary to transmit dedicated pilot symbols. In addition, this equalization makes it possible to obtain information on the relative offsets of the CCSK sequences allowing its demodulation.

[0022] According to a particular embodiment, the equalization steps comprise a term-by-term multiplication of the OFDM symbol by the complex conjugate of the corresponding obtained vector.

[0023] Let Hk be the vector of N elements resulting from the DFT of the CIR of the OFDM symbol k. This vector is used to equalize another OFDM symbol of index m. This equalization is done by a term-by-term multiplication, between the complex symbols of the received OFDM symbol m and the complex conjugate of the vector Hk. Concretely, by designating by hlk the ith element of the vector Hk, that is the complex coefficient of the channel affecting the sub-carrier of index i, and by designating by y' the complex symbol received on the sub- carrier of index i of the OFDM symbol of index m, the equalization operation is carried out by the calculation f . / ri \ * = y;„ ■ [nk)

[0024] This equalization method is similar to that of Maximum Ratio Combining (MRC). Unlike a more common equalization method such as Zero Forcing (ZF), according to which the symbol terms are divided by H, MRC makes it possible to take full advantage of the frequency diversity of the channel when calculating the cross-correlation with the root sequence. Such an arrangement makes it possible to significantly improve the equalization compared to the prior art according to which the symbol terms are divided by H.

[0025] According to a particular embodiment, all of the OFDM symbols of the frame are equalized from an estimation vector determined for a particular OFDM symbol comprising a CCSK sequence, called the reference sequence.

[0026] Thus, the entire frame is equalized with a CIR determined from a particular OFDM symbol. Such an arrangement makes it possible not only to equalize all the other symbols of the frame but also to obtain an estimate of all the relative offsets between the transmitted sequences.

[0027] Let us take for example a frame of three OFDM symbols Po, Pi and P2 each containing a CCSK sequence from which three CIRs Ho, Hi, and H2 have been respectively determined. By equalizing the frame with Ho, the correlation maxima are located at the indices Pi-P0 modulo N and P2-P0 modulo N. By arbitrarily taking a sequence of the frame as a reference, we can thus know the offset of each sequence with respect to this reference sequence. These estimates are respectively Pi-P0 and P2-P0. The method thus makes it possible to carry out an offset demodulation in which all the sequences have been offset by a value Po.

[0028] By applying this principle to all the symbols of the frame, we obtain a matrix such as the following matrix in which a vector Pm / n corresponds to the symbol m equalized with the CIR determined for the symbol n.

[0029] For example, by arbitrarily taking the sequence P2 as a reference, the last line of such a matrix makes it possible, for example, to know the relative offsets of the sequences Po and Pi with respect to the sequence P2.

[0030] The method thus allows for offset demodulation of the entire frame, with channel estimation and equalization without a pilot sequence. All sequences, including the reference sequence, carry useful information.

[0031] Such a matrix also makes it possible to develop different demodulation algorithms, in particular, an approach based on a factor graph.

[0032] According to a particular embodiment, the reference sequence is included in an OFDM symbol transmitted in the second third of the frame.

[0033] The OFDM symbol comprising the sequence used as a reference for determining the relative offsets between the sequences of the frame is selected in the middle of the frame. It is preferable to choose a central OFDM symbol in the frame to reduce the impact of the temporal decorrelation of the channel.

[0034] According to a particular embodiment, it is proposed to select a first reference sequence in an OFDM symbol located in a first third of the frame and a second reference sequence included in an OFDM symbol of the last third of the frame, the first and second sequences being further selected such that they are assigned the same offset on transmission.

[0035] It is thus proposed to consider at least two reference sequences in the received frame in order to capture the temporal evolution of the channel and improve the interpolation step when the propagation channel evolves over time and the channels become decorrelated during the transmission of the same frame. The selected reference sequences must have the same cyclic shift at transmission relative to the same root sequence, so that the relative shifts estimated from the different reference sequences are identical, apart from the noise.

[0036] The channel estimates obtained from the different reference sequences can be used to equalize the symbols in the respective neighborhood of these symbols, equalize the frame with an average of the two estimates or even equalize the frame from an interpolation calculated from the two estimates.

[0037] The second reference sequence may be a repetition of a first sequence selected in the frame which is inserted intentionally by the transmitter to allow the receiver to take into account an evolution of the channel. Whereas the prior art in which R pilot symbols not carrying useful data are inserted in a frame, the method only requires Rl symbols not carrying useful data. The method thus makes it possible to gain in spectral efficiency compared to the prior art.

[0038] According to a particular embodiment, the method comprises a step of obtaining an indication relating to the location in the frame of the OFDM symbols containing respectively the first and the second reference sequence.

[0039] The indication allows the receiver to identify the OFDM symbols comprising a reference sequence. The indication is for example transmitted in advance by the transmitter in a suitable message, or preconfigured in the receiver in accordance with a standard.

[0040] Alternatively, the method may comprise a prior step during which the receiver transmits to the transmitter a message comprising an indication relating to one or more locations at which it wishes to see the transmitter insert a reference sequence.

[0041] According to another aspect, the invention relates to a device for equalizing an OFDM transmission channel on which a data frame is received from a transmitter, said frame comprising a plurality of binary words, a word being associated with a particular sequence of chips obtained by cyclic shifting of a root sequence, the device comprising a radiofrequency receiver and a processor coupled to a memory in which are recorded program instructions adapted to implement the steps of an equalization method comprising the following steps: - Reception of a first and at least one second OFDM symbol respectively carrying a first and a second sequence of chips of said frame, - Determination of a first Impulse Response of the channel by calculating a correlation between the first OFDM symbol received and the IDFT of the root sequence, and of at least a second Impulse Response of the channel by calculating a correlation between the second OFDM symbol received and the IDFT of the root sequence, - Calculation of a Discrete Fourier Transform of the first and second Impulse Responses to obtain respectively a first estimation vector of the H0 channel and a second estimation vector of the Hl channel, - Equalization of the second OFDM symbol from the first vector, - Equalization of the first symbol from the second vector, and - Demodulation of the frame from the equalized symbols.

[0042] The invention also relates to a communication terminal comprising an equalization device as described above.

[0043] In a particular embodiment, the different steps of the equalization method are determined by computer program instructions.

[0044] Consequently, the invention also relates to a computer program comprising instructions adapted to the implementation of the steps of an equalization method as described above, when the program is executed by a processor.

[0045] This program may use any programming language, and be in the form of source code, object code, or code intermediate between source code and object code, such as in a partially compiled form, or in any other desirable form.

[0046] The invention also relates to a computer-readable information medium on which is recorded a computer program comprising instructions for executing the steps of an equalization method as described above.

[0047] The information medium may be any entity or device capable of storing the program. For example, the medium may comprise a storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, a flash memory, or a magnetic recording means, such as a hard disk.

[0048] 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 may in particular be downloaded from an Internet-type network.

[0049] 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.

[0050] The various embodiments or features mentioned above may be added independently or in combination with each other, to the steps of the equalization method.

[0051] The devices, terminals, programs and information media have advantages similar to those of the method to which they correspond. Brief description of the figures

[0052] Other characteristics and advantages will appear on reading a preferred embodiment described with reference to the appended drawings among which: - [Fig.l] represents an environment suitable for implementing the equalization method according to a particular embodiment, - [Fig.2] represents the main stages of a transmission process, according to a particular embodiment, - [Fig.3] represents the main stages of an equalization process, according to a particular embodiment, - [Fig.4] illustrates the architecture of a device suitable for implementing an equalization method according to a particular embodiment. Detailed description

[0053] In the following description, embodiments are described on the basis of non-limiting examples making it possible to explain the concepts on which the invention is based. In particular, although the examples and the terminology used may refer to certain well-known technologies or standards, these references are not limiting and other technologies may be adapted to implement the concepts of the invention. For example, the CCSK or OFDM technologies referred to can be replaced by different modulation techniques based respectively on circular shifts of complex value sequences or multi-carrier transmissions without the need to modify the invention.

[0054] [Fig. 1] represents an environment 100 adapted to implement the equalization method according to a particular embodiment.

[0055] The environment 100 comprises wireless equipment 102, for example a connected object such as a temperature sensor. Of course, different types of equipment 102 can be envisaged, such as sensors or connected objects, M2M (Machine To Machine) communication devices or terminals, such as tablets or mobile phones for example.

[0056] In the example of [Fig.l], the equipment 102 exchanges data with an application server 103 of a communication network 104 via a gateway 105. The gateway 105 is for example a radio unit of a cellular access network, such as a 3G, 4G, 5G, 6G, WiFi® or even WiMax® network, a home gateway or any type of device suitable for communicating by radio frequency with the sensor 102. In the environment 100 taken as an example, the gateway 105 is a base station, an eNodeB or a gNodeB, and communicates with the communication network 104 via a connection 107 which may be wired, optical or wireless.

[0057] In the present description, particular attention is paid to the communication channel 106 used by the equipment 102 to communicate with the gateway 105. The communication channel is a radiofrequency channel using multi-carrier modulation of the CP-OFDM type.

[0058] More precisely, when the equipment 102 wishes to transmit a data frame to the application server 103, the binary words composing the frame are modulated according to a CCSK modulation scheme before being transmitted to the gateway 105 on a plurality of OFDM subcarriers. The gateway demodulates the received binary words to reconstitute the transmitted frame before retransmitting the data to the server 103 through a connection 107 in accordance with a communication protocol which may be distinct from that used between the equipment 102 and the gateway 105.

[0059] A particular embodiment will now be described with reference to Figures 2 and 3.

[0060] During a first step 200, the equipment 102 prepares a first data frame to be transmitted to the gateway 105. Here, the term "data frame" or "packet" is understood to mean a transmission unit comprising one or more binary words with which an integrity control mechanism is associated, for example a cyclic redundancy check (or CRC, for Cyclic Redundancy Code in English). Such a mechanism is known per se and makes it possible to detect transmission errors. For example, For example, the data frame prepared by the temperature sensor 102 may include a field comprising information relating to a temperature reading and a field comprising a CRC.

[0061] The binary words composing the prepared frame are then modulated during a step 201. This modulation step consists of encoding the data to allow their transmission on the communication channel. It is proposed here to use a CCSK modulation according to which each binary word is associated with a particular version of a root sequence obtained by a particular cyclic shift. Thus, to code a binary word of p bits, the root sequence must comprise at least 2p elements called chips.

[0062] The root sequence is preferably a sequence having a perfect autocorrelation function, i.e., a single correlation peak for zero offset. The root sequence is further configured so that its IDFT has a good autocorrelation function. In a particular embodiment, the root sequence is a Zadoff-Chu (ZC) sequence. ZC sequences indeed have such properties.

[0063] The sequences encoding the binary words of the frame are presented as input to a serial-to-parallel (S / P) converter in step 202. The serial-to-parallel converter is configured to produce a vector of size N for each sequence presented as input. Thus, for a CCSK sequence presented as input to the S / P converter, a vector of size N comprising the N elements is obtained.

[0064] During a step 203, a Discrete Inverse Fourier Transform (IDFT) is then applied to the vector generated by the serial / parallel converter to obtain a vector of N samples C[ / l], with il E [0; N - 1],

[0065] The vector thus obtained is then serialized in step 204 by a parallel / serial (P / S) converter, so that, during a step 205, each element of the vector is transmitted on a distinct OFDM subcarrier of the same OFDM symbol, after adding a cyclic prefix.

[0066] In a particular embodiment, the S / P converter is configured to produce a vector of size K = q*N when it is successively presented with q CCSK sequences of size N denoted Cp. with ie [0; q - 1],

[0067] More precisely, the S / P converter is such that the q sequences of N elements coding the words to be transmitted are interleaved periodically and regularly in a vector of size K = q * N in which the elements C / > [n ] of the q sequences Cp. associated with the q words to be transmitted are arranged.

[0068] It is thus proposed to intercalate the elements of the q sequences in a vector of size K=q*N according to a periodic interlacing law according to which an element Cp [n ] of a sequence with n E [0; N - 1] and ie [0; q - 1] occupies the rank r - i*K + n in the interlaced vector.

[0069] The equalization process will now be described with reference to [Fig.3].

[0070] During a first step 300, at least a first and a second OFDM symbol transmitted by the equipment 102 are received successively by the gateway 105. The OFDM symbols comprise at least one CCSK sequence of N chips coding for a binary word and are received on N subcarriers.

[0071] During a step 301, a first Channel Impulse Response (CIR) is determined by calculating a correlation between the first OFDM symbol y1 received and the IDFT of the root sequence used by the equipment 102 to modulate the transmitted binary words. A second Channel Impulse Response (CIR) is also determined by calculating a correlation between the second OFDM symbol y2 received and the IDFT of the root sequence.

[0072] The first and second Impulse Responses obtained are of size N and each have an offset identical to the offset of the IDFT of the sequence from which it was calculated. This offset is notably linked to the offset of the sequence at transmission, but an additional offset may be caused by a frequency aliasing phenomenon. Thus, for example, by noting [ h0, h^ ..., hK_2, ] the CIR of a sequence having a zero offset (i.e. the CIR of a sequence corresponding to the root sequence), the CIR obtained for a sequence shifted cyclically by an index 1 can be of the form [^Æ-t' h^2]■

[0073] Consider for example a sequence S and the IDFT of this sequence I(S). If the sequence S has an offset P, the sequence I(S) has an offset related to P, not necessarily equal to P, noted I(P). The CIR then has the same offset as I(P).

[0074] Since the obtained CIRs may be noisy, it is not always possible to differentiate a channel coefficient from a noise coefficient, and therefore to determine the shift undergone by a CIR. In addition, a shifted CIR can be multiplied by a constant complex coefficient (this is the case for Zadoff-Chu sequences).

[0075] In a particular embodiment, the method comprises a step 302 of calculating a Discrete Fourier Transform (DFT) of the first CIR and the second CIR obtained for the channel, in order to obtain a first estimate and a second estimate of the channel in frequency on the set of subcarriers. A first vector Ho is thus obtained corresponding to an estimate of the channel from the first OFDM symbol and a second vector Hi corresponding to an estimate of the channel from the second OFDM symbol received.

[0076] In step 303, an equalization of the second OFDM symbol received is carried out from the estimation vector Ho of the channel obtained from the first OFDM symbol received and equalizing the first received OFDM symbol from the channel estimation vector Hi obtained for the second received OFDM symbol.

[0077] In a particular embodiment, the equalization is carried out by term-by-term multiplication by the complex conjugate of the vector obtained.

[0078] According to a particular embodiment, an estimation vector H; is determined for each OFDM symbol Si of the received frame, and used to equalize all of the OFDM symbols making up the data frame. These multiple equalizations can be used as input for demodulation algorithms.

[0079] By taking as a reference the CCSK sequence received in the first symbol and considering that this reference sequence has an index offset Po relative to the root sequence, and that the sequence received in the second symbol has an index offset Pb, we can obtain information on the relative offset between the sequence of the second symbol and that of the first symbol. Indeed, assuming that the channel evolves slowly over time, the channels undergone by successive OFDM symbols are strongly correlated on each subcarrier. By equalizing the second OFDM symbol with Ho and calculating a correlation maximum via CCSK demodulation, the correlation peak will be located at the index equal to Pi-P0 modulo N. Conversely, by equalizing the first zero symbol with Hb, the correlation maximum will be located at Po -Pi modulo N

[0080] In step 304, demodulation is performed from the equalized symbols and the relative offsets between the sequences. The demodulation is performed by cross-correlation with the root sequence.

[0081] Since the relative offsets between each of the sequences transmitted in the frame are known, it is possible to perform a shifted demodulation using the integrity check data of the frame to determine the shift to be applied to find the transmitted binary words. To do this, the received CCSK sequences can be demodulated by considering a zero shift of the sequence taken as reference. Since the relative offsets are known, the binary words of the frame are obtained. An integrity check, for example a CRC, is then used to determine whether the decoded frame is correct. If this is not the case, i.e. if the integrity check fails, a new demodulation attempt is made after applying a shift to the CCSK sequences of the frame. This is done until the integrity check shows correct demodulation.

[0082] According to a particular embodiment, a second sequence is taken for reference in an OFDM symbol located in a region of the frame distinct from the region in which the first reference sequence is selected. Preferably, the first reference sequence is taken in an OFDM symbol located in the first third of the frame, and the second reference sequence is taken in an OFDM symbol located in the last third of the frame. We also take care to select reference sequences with the same offset at transmission. We then determine a first CIR from the first sequence and a second CIR from the second sequence. Since the cyclic offsets are identical at transmission for the selected sequences, we obtain two equivalent CIRs, apart from noise and the temporal evolution of the channel. The symbols of the frame are then equalized from the first CIR and the second CIR. These two equalizations can then be combined, for example by taking an average for each symbol, in order to better take into account an evolution of the channel over time.

[0083] In a particular embodiment, the gateway 105 obtains an indication allowing it to determine the location of one or more CCSK sequences to be used as a reference.

[0084] For this, according to a particular embodiment, the equalization method comprises a prior step of transmitting to the transmitting equipment 102 a message comprising an indication relating to a desired location of one or more reference sequences. For example, the gateway 105 transmits a message in which it indicates that a first reference sequence must be inserted into the first OFDM symbol of a frame and that a second reference sequence must be inserted into the last OFDM symbol of the frame. Upon receipt of such a message, the equipment 102 configures communication means so as to insert CCSK sequences whose offset is identical to the locations indicated when sending one or more frames.In practice, when the gateway 105 requests that reference sequences be inserted, for example, in the second position and in the penultimate position of a frame, the equipment 102 can copy the second OFDM symbol of the frame in the penultimate position thereof. In this way, two reference sequences of identical offset are present at the indicated locations. Whereas two dedicated pilot symbols would have been necessary with the prior art technique, only one symbol not carrying useful data is inserted with the proposed equalization method.

[0085] Other techniques may be considered for the receiver to obtain the location of the reference sequences. For example, the transmitter may transmit in a pre-configuration message the locations at which these sequences will be inserted. In another example, these locations are defined by a standard and pre-configured in the receiver.

[0086] [Fig.4] illustrates the architecture of a device 400 adapted to implement the equalization method according to a particular embodiment of the invention.

[0087] The device 400 comprises a data processing module comprising a storage space 401, for example a memory (MEM), a processing unit 402, equipped for example with a microprocessor (PROC), and controlled by a computer program (PGR) 403 whose instructions are configured to implement the equalization method as described previously in relation to [Fig.3].

[0088] At initialization, the code instructions of the computer program 403 are for example loaded into the memory 401 before being executed by the processor of the processing unit 402. The microprocessor of the processing unit 402 implements, according to the instructions of the computer program 403, the steps of the method for equalizing an OFDM transmission channel on which a data frame from a transmitter is received, said frame comprising a plurality of binary words, a word being associated with a particular sequence of chips obtained by cyclic shifting of a root sequence, and in particular the steps of: - Receiving a first and at least one second OFDM symbol respectively carrying a first and a second sequence of chips of said frame, - Determination of a first Impulse Response of the channel by calculating a correlation between the first OFDM symbol received and the IDFT of the root sequence, and of at least a second Impulse Response of the channel by calculating a correlation between the second OFDM symbol received and the IDFT of the root sequence, - Calculation of a Discrete Fourier Transform of the first and second Impulse Responses to obtain respectively a first estimation vector of the H0 channel and a second estimation vector of the Hl channel, - Equalization of the second OFDM symbol from the first vector, - Equalization of the first symbol from the second vector, and - Demodulation of the frame from the equalized symbols.

[0089] For this, in addition to the memory 401 and the processor 402, the device comprises communication means 404, for example an OFDM transducer adapted to transmit and receive signals on a plurality of orthogonal carriers. The communication means 402 are for example configured by computer program instructions to allow the reception of at least a first and a second OFDM symbol each comprising at least one CCSK sequence and transmitted by a transmitter on a plurality of subcarriers.

[0090] The device 400 also comprises a serial / parallel (S / P) converter 405 adapted to generate a vector of N chips from an OFDM symbol received on N subcarriers by the communication means 404, the N chips corresponding to the N elements of a CCSK sequence of N elements coding for a binary word.

[0091] The device 400 also comprises a module 406 for calculating an impulse response transmission channel correlation (CIR) from a CCSK sequence obtained by the S / P converter from an OFDM symbol. The module 406 is for example implemented by computer program instructions which are configured to calculate an inverse Discrete Fourier Transform (IDFT) of a root sequence used by the transmitter to modulate binary words according to a CCSK coding and to calculate a cross-correlation between the CCSK sequence received in an OFDM symbol and the IDFT calculated for the root sequence. More specifically, the program instructions are configured to calculate a first CIR from a first received OFDM symbol and a second CIR from a second received OFDM symbol.

[0092] The device 400 also comprises a channel estimator 407 configured by program instructions to calculate a Discrete Fourier Transform of the first CIR and the second CIR determined by the module 406 in order to obtain a first estimation vector H0 and a second estimation vector Hl, each comprising N coefficients.

[0093] The device also comprises an equalization module 408 adapted to carry out at least a first equalization of the OFDM symbols composing the received frame from the first estimation vector H0 and a second equalization of the OFDM symbols composing the received frame from the second estimation vector HL. For this, the equalization module comprises program instructions configured to multiply term by term each OFDM symbol of the frame by the complex conjugate of the vector H0 and to multiply term by term each OFDM symbol of the frame by the complex conjugate of the vector HL.

[0094] According to a particular embodiment, the equalization module is further configured to determine a relative offset between the CCSK sequences received from at least one reference sequence.

[0095] The device finally comprises a CCSK demodulation module 409. The module 409 is for example implemented by program instructions configured to carry out an offset demodulation from the relative offsets determined by the equalization module 408.

[0096] In a particular embodiment, the device is integrated into a connected object such as a sensor, a mobile communication terminal, a gateway or even a base station.

Claims

Claims

1. Method for equalizing an OFDM transmission channel on which a data frame is received from a transmitter, said frame comprising a plurality of binary words, one word being associated with a particular chip sequence obtained by cyclic shifting of a root sequence, the method comprising the following steps: - Reception (300) of a first and at least one second OFDM symbol respectively carrying a first and a second chip sequence of said frame, - Determination (301) of a first Impulse Response of the channel by calculating a correlation between the first OFDM symbol received and the IDFT of the root sequence, and of at least one second Impulse Response of the channel by calculating a correlation between the second OFDM symbol received and the IDFT of the root sequence,- Calculation (302) of a Discrete Fourier Transform of the first and second Impulse Responses to obtain respectively a first estimation vector of the HO channel and a second estimation vector of the Hl channel, - Equalization (303) of the second OFDM symbol from the first vector H0, and equalization of the first symbol from the second vector Hl, and - Demodulation (304) of the frame from the equalized symbols.,

2. The method of claim 1 wherein the equalizing steps comprise term-by-term multiplication of the OFDM symbol by the complex conjugate of the corresponding obtained vector.

3. Method according to any one of the preceding claims in which the set of OFDM symbols of the frame are equalized from an estimation vector determined for a particular OFDM symbol comprising a CCSK sequence, called the reference sequence.

4. The method of claim 3 wherein the reference sequence is included in an OFDM symbol transmitted in a second third of the frame.

5. A method according to claim 3 wherein a first reference sequence is selected from an OFDM symbol located in a first third of the frame and a second reference sequence is selected in an OFDM symbol of the last third of the frame, the first and second sequences being assigned the same offset on transmission.

6. Method according to claim 5 such that it further comprises a step of obtaining an indication relating to the location in the frame of the OFDM symbols respectively containing the first and the second reference sequence.

7. Device for equalizing an OFDM transmission channel on which a data frame is received from a transmitter, said frame comprising a plurality of binary words, a word being associated with a particular sequence of chips obtained by cyclic shifting of a root sequence, the device comprising a radiofrequency receiver (404) and a processor (402) coupled to a memory (401) in which are recorded program instructions (403) adapted to implement the steps of an equalization method comprising the following steps: - Reception of a first and at least one second OFDM symbol respectively carrying a first and a second sequence of chips of said frame, - Determination of a first Impulse Response of the channel by calculating a correlation between the first OFDM symbol received and the IDFT of the root sequence,and at least a second Impulse Response of the channel by calculating a correlation between the second OFDM symbol received and the IDFT of the root sequence, - Calculation of a Discrete Fourier Transform of the first and second Impulse Responses to obtain respectively a first estimation vector of the channel HO and a second estimation vector of the channel Hl, - Equalization of the second OFDM symbol from the first vector, - Equalization of the first symbol from the second vector, and - Demodulation of the frame from the equalized symbols.,

8. Communication terminal comprising an equalization device according to claim 7.

9. Computer program comprising instructions adapted to implement the steps of an equalization method according to any one of claims 1 to 6, when the program is executed by a processor.

10. A computer-readable information medium on which is recorded a computer program comprising instructions for carrying out the steps of an equalization method according to any one of claims 1 to 6.

Images