WO2008031963A1

WO2008031963A1 - Method of receiving amrf-modulated signals and reception device implementing this method

Info

Publication number: WO2008031963A1
Application number: PCT/FR2007/051862
Authority: WO
Inventors: Denis Callonnec; Rodolphe Legouable; Jean-Philippe Javaudin
Original assignee: France Telecom
Priority date: 2006-09-13
Filing date: 2007-08-31
Publication date: 2008-03-20
Also published as: FR2905814A1

Abstract

The method comprises the following steps, carried out in a base station: a global signal encompassing uplink signals from terminals is received; a plurality of demodulation pathways are supplied with the global signal received; in various demodulation pathways, various phase rotations are applied respectively, producing various demodulation frequency shifts predetermined with respect to a nominal demodulation frequency; and, for a relevant uplink signal, one of the plurality of demodulation pathways is selected according to a chosen quality criterion. This method is particularly adapted for uplink signals for mobile communications.

Description

METHOD FOR RECEIVING MODULAR AMRF SIGNALS AND RECEIVING DEVICE USING THE SAME

The present invention relates to the reception of modulated signals in a communication system implementing multiple frequency division access (FDMA), for example Orthogonal Frequency Division Multiplexing (OFDM) accesses. It should be noted at the outset that, in an FDMA communication system, the transmissions can be carried out on any type of transmission medium (radio, wired, optical, etc.). The present invention is therefore not limited to the reception of signals by radio but extends to any type of transmission medium.

More particularly, the invention relates to a method of reception by a receiving station of uplink signals from transmitting terminals, in the context of a frequency division multiple access. At a given instant, one or more terminals transmit simultaneously on several frequencies or carriers (frequency bands), which are chosen in a multiplex of frequencies or carriers assigned to the terminals. The receiving station receives a global signal grouping the uplink signals from the different terminals and distributed over several frequencies or distinct carriers, then demodulates these uplinks and restores the information they convey.

In transmission, the signals to be transmitted are transposed on the frequencies or carriers assigned to the terminals for transmission. The terminals must therefore synthesize the transmission frequencies on which the signals are transposed. Various techniques are known for synthesizing these frequencies. By way of illustrative examples, there may be mentioned a local oscillator specific to the terminal or else a phase lock device on a reference frequency external to the terminal. The accuracy of the synthesis of the transposition or transmission frequencies of a terminal is conditioned by the performance of the synthesis means implemented in the terminal. Such a known reception method is in particular implemented with a plurality of users connected by means of terminals to a base station, for example by radio signals. Several users simultaneously transmit these uplink signals to the base station which retransmits the signals once decoded, for example on another medium, in order to route them to the recipient. These uplink signals are modulated on separate transmission frequencies in the same frequency or carrier multiplex.

In order to correctly demodulate the uplink signals received by the base station, the phases of the signals transmitted by the terminals must be correctly synchronized frequency and time on a common reference. Indeed, if the signals are poorly synchronized, inter-symbol interference and inter-carrier interference occur. Furthermore, the tolerance on the temporal synchronization of the phases is inversely proportional to the inter-carrier gap, whereas the tolerance on the frequency synchronization is proportional to this inter-carrier gap. These two tolerances being totally opposite each other, it is necessary to find a compromise, or to release one of the tolerances.

It is in particular known to set up a frequency synchronization loop or PLL (for Phase Locked Loop in English) in order to keep the local oscillators of user terminals synchronized frequently with the common reference. However, it is required a tolerance of the order of 2 to 5% of the inter-carrier gap. However, for a conventional system operating at a frequency of 2 GHz, this inter-carrier gap is of the order of kHz. Consequently, this tolerance requires a precision on the frequency of the order of 10 ^{~ 7} to 10 ^{~ 8} with respect to the frequency to be transmitted. Such a system is therefore difficult to implement, especially with mobile networks whose channels are changing rapidly.

For this purpose, the invention proposes a method for receiving signals, making it possible to reduce the requirements with respect to phase tolerances. The invention proposes a method for receiving uplink signals from terminals, each uplink signal having been previously modulated by a respective terminal for frequency division multiple access. The method comprises a step of receiving a global signal gathering upstream signals. According to a present definition of the invention, the method further comprises the steps according to which:

a plurality of demodulation channels are fed with the global signal received, and

in different demodulation channels, different phase rotations are applied producing different predetermined demodulation frequency offsets with respect to a nominal demodulation frequency, respectively;

for a considered uplink signal, one of the plurality of demodulation channels is selected according to a chosen quality criterion.

Thanks to the invention, it avoids the use of a complex system of synchronization closed loop. A plurality of demodulation channels slightly offset in frequency with respect to a nominal demodulation frequency are used, so that for given terminals, having a shifted phase in transmission, one of the demodulation channels will be more "in phase" than the others and will allow a better decoding of the signal coming from this terminal. Thus, a simple way is offered to loosen the constraint on the synthesis of frequencies in the transmitting terminals.

By phase, we mean the angle value that the modulated signal takes as a function of time, which, in the case of a modulation according to an OFDMA method, depends on a data item on the frequency, of a phase datum initial and initial moment. The dependence on time is a function of the choice of the modulation waveform.

Consequently, a phase rotation can be performed by a frequency and / or time shift depending on the nature of the correction to - AT -

to bring. Differential frequency offsets are preferably made for each channel so that a statistical distribution of the shifted frequencies around the nominal frequency can be obtained. In the context of the description of the embodiment of the invention to be made later, frequency offsets are used which will correct frequency differences with respect to the reference demodulation frequency. We seek to increase the tolerance on the frequency without impacting the tolerance over time as explained above.

After the demodulation, it is possible to assign to a considered demodulation channel the signal originating from a terminal, if the quality of the signal obtained by this channel is greater than the quality of the signal obtained on the other channels. This makes it possible to affect throughout the communication this signal to the most suitable channel without having to maintain a phase control loop.

This can in particular be implemented according to the quality criterion chosen. According to one variant of the invention, the demodulation channel which has the phase closest to that of the uplink signal, in the sense of a deviation in absolute value, is selected for the uplink signal.

According to another variant, the uplink signal considered comprising at least one predetermined sequence, it is allocated a demodulation channel according to the extraction quality of said sequence. This makes it possible to check the correct phase synchronization in a simple manner. However, it is necessary to impose a modification of the signal sent by the terminals.

Advantageously, a polyphase filtering type IOTA or Nyquist type is carried out on each decoding channel. This type of filtering makes it possible to format the signal and to disconnect the spectra of the different users. This makes it possible to further reduce the stress on frequency synchronization. The invention also relates to a method for transmitting uplink signals from terminals, according to which, in transmission, each terminal transmits a respective uplink signal on at least one continuous sub-multiplex of a frequency band, and it imposes between each multiplexing the presence of a carrier of zero amplitude and in reception, the steps of the method according to the invention are carried out.

Thanks to this, the constraint on frequency synchronization is further reduced and inter-carrier interference is avoided.

The invention also relates to a reception device adapted to receive an overall signal comprising a sum of uplink signals from terminals, each uplink signal being modulated by a respective terminal according to a multiple division access on a frequency multiplex and being intended for to be demodulated on reception at a given nominal demodulation frequency. According to one definition of the invention, the device further comprises a plurality of demodulation channels adapted to receive a copy of the overall signal at input and having phase rotation means adapted to introduce a predetermined frequency offset with respect to the frequency nominal demodulation circuit, different demodulation channels being adapted to apply different respective phase rotations, and selection means for assigning a considered uplink signal to a demodulation channel according to a selected criterion.

Such a reception device can be integrated in a base station of a radio access network.

Other features and advantages of the present invention will become apparent in the following description of nonlimiting exemplary embodiments, with reference to the appended drawings, in which:

- Figure 1 is a diagram showing a communication network in which the invention is implemented;

- Figure 2 schematically shows a base station according to the invention; FIG. 3 is a flowchart showing a method according to the invention; and

- Figure 4 shows the spectrum of a signal received by the base station in the context of the implementation of the invention.

As shown in FIG. 1, a communication network in which the invention can be implemented comprises a plurality of terminals T1, T2 and T3 transmitting uplink signals to a base station 1. This base station is connected to a network 2, which routes the signal to its recipient. These devices communicate with each other by means of signals modulated here in OFDMA. This is a multicarrier modulation method, using orthogonal carriers. This orthogonality notably allows a high spectral overlap, reducing the spectral bulk. This orthogonality constraint on the carriers must be respected in the time and frequency domains. It will be noted, however, that the invention applies not only to the transmission and reception of modulated signals according to an OFDM access but also to any type of frequency division multiple access, for example SS-MC-MA.

The structure of this base station 1 according to the invention is illustrated schematically in FIG. 2. In the particular example described, this base station 1 comprises a radio reception module 3, a frequency transposition module 4. , using a local oscillator 14, an analog / digital converter 5, denoted CAN, a duplication module 6 and a plurality of demodulation channels or demodulation chains Ci to C-2k + i - It will be noted that instead of being odd as in the example described, the number of demodulation channels could be even.

The radio reception module 3 is arranged to receive frequency-division multiplexed uplink signals grouped together in a global signal, from transmitting terminals.

The frequency translation module 4, placed at the output of the radio reception module 3, is arranged to transpose the overall signal received from the signal transmission frequency at a frequency adapted to the demodulation of the signal by the base station, which will be called hereinafter "nominal demodulation frequency", denoted F _n0 m- In the particular example described, the nominal demodulation frequency of the base station is an intermediate frequency In the alternative, the module 4 could transpose to the baseband.

The CAN converter 5, placed at the output of the transposition module 4, is arranged to convert the analog signal received and transposed into F1 into a digital signal.

The duplication module 6 is arranged to duplicate the global signal received in 2k + 1 copies and to supply the 2k + 1 demodulation channels with the 2k + 1 duplicated signals respectively.

The demodulation channels C1 to C2k + i are connected in parallel to the output of the duplication module 6. Each demodulation channel comprises a phase rotation module 7 adapted to be applied to an incoming signal, in this case the global signal received. transposed at the nominal demodulation frequency, provided by the duplication module 6, a phase rotation producing a predetermined frequency offset with respect to the nominal demodulation frequency F _nom . In the particular example described here, the phase rotation is different on each of the channels. It makes it possible to intentionally introduce, on the demodulation path considered, an offset of the transposition frequency around the nominal demodulation frequency F _nom , in order to correspond to the disparity of the frequency deviations with respect to the nominal modulation frequency at the transmitting terminals. Thus, it can be indicated that for the channel Cj, j varying between 1 and 2k + 1, the transposition frequency is substantially equal to Fj = F _n0 m + (jk-1) ΔF,

where F _nom is the nominal frequency of demodulation of the base station and .DELTA.F is not a predetermined frequency offset. A fraction, for example 5% of the inter-carrier gap, may be used as the predetermined frequency offset step.

It should be noted that the lower the frequency offset step, the higher the number of demodulation channels required. In the particular example described, the phase rotation module 7 operates on a digital signal. Alternatively, it could operate on an analog signal. In this case, each demodulation channel comprises an analog digital converter, positioned for example at the output of the phase rotation module.

Each demodulation channel also comprises here a polyphase formatting filter 8 placed at the output of the phase rotation module 7. OQAM-type polyphase filtering, for example an IOTA or Nyquist filter, or else BFDM type, which makes it possible to obtain well dissociated spectra, thus limiting the risk of spectral overlap.

Next, direct Fourier transform calculation means 9, for example fast Fourier transform, are used. Then, it is possible to provide a decoder 10 by demodulation. However, depending on the type of test implemented by the test device 11, this decoder may also be downstream of the plurality of channels. Since the uplink signals have undergone emission of the respective phase rotations, the aim is to assign to a demodulation channel considered the uplink signals which have undergone the substantially closest phase rotation emission to that introduced by the demodulation channel considered.

To do this, according to a first variant, each uplink signal comprises pilot symbols whose characteristics are known to the receiving base station. In particular, these pilot symbols have a predetermined phase. In the base station, a measurement of the phase rotation on the carrier of the pilot symbols is carried out, after demodulation, for each of the demodulation channels. For a given uplink signal, a demodulation channel is allocated according to these measurements made on the phase rotations of the received and demodulated pilot symbols. In other words, the channel having, after demodulation, the smallest residual phase rotation of the pilot symbol received for a given terminal is selected. In this case, the channel selection device is developed from a transmission by the terminal of one or more continuous pilots. He is not necessary to have a decoder per channel, and the test device extracts the pilot symbols.

However, it is also possible to provide the transmission of a predetermined sequence, which may for example be repeated periodically or be transmitted once. The test device 11 extracts this sequence of the signal and, for a given upstream signal, a demodulation channel is allocated according to the extraction quality of the sequence. In this case, as in the previous one, the test device 11 is upstream of the decoding device 10.

In the case where the decoder 10 is downstream of the test device 11, as shown in FIG. 2, another quality criterion is used on the upright signal of the user once decoded. For example, one can perform a test on the signal-to-noise ratio, or on the bit error rate. For example, we can test the check bit or check-sum to determine if it is correct or wrong.

This test device is connected to a selection member which will assign a user or a terminal to a particular channel, in order to obtain for this user the best possible decoded signal. Thus, the selection device allocates a demodulation channel to users whose phase of the transmission frequency is close to that of the channel.

Finally, a transmission device 13 allows the retransmission of the signals of the users to the network 2 to route the signals to their recipients.

The process according to the invention is represented by the flowchart of FIG.

In step S300, the base station receives n uplink signals in the form of a global signal grouping these n uplink signals.

In step S301, the overall signal is transposed to the nominal demodulation frequency of the base station.

In step S302, the transposed analog global signal is converted to a digital signal.

In step S303, the overall signal is duplicated and supplied to each demodulation channel j. In step S304, the global signal injected into each demodulation channel j undergoes a phase rotation, denoted ROT _j , here different on each of the channels, having the effect of causing a transposition frequency shift around the nominal frequency of demodulation, as we saw earlier. Then, the signal is filtered, or shaped according to multiphase IOTA or Nyquist (S305) filtering methods.

Then, in step S306, a test is carried out for a rising signal considered, here the rising signal of index i, with 1 ≤ i ≤ n. This is done after the application of a direct Fourier transform to the overall signal, and possibly a decoding according to the chosen test.

The steps S303 to S306 are repeated for each of the demodulation channels of index j, in other words for any index j such that 1 ≤ j ≤ 2k + 1.

After reiterating steps S303 to S306 for each index j between 1 and 2k + 1, it is determined in step S307 for which demodulation channel j the uplink signal considered i is of better quality, that is, say with which frequency shift the signal i is best rendered.

Then, in step S308, this uplink signal i is allocated on the channel j determined in the previous step (S307), so that, when the uplink signal i is transmitted in step S309, the rising signal demodulated by this channel determined in step S307.

Finally, for an uplink signal considered i, the demodulation channel j is selected which has the phase closest to that of the uplink signal i, in the sense of a deviation in absolute value the lowest, it is demodulated on the selected demodulation channel j and the data (decoding) to be retransmitted to the recipient.

The steps S306 to S309 are repeated for each of the n uplink signals of index i of the global signal received, ie for any index i varying from 1 to n, in order to determine which of the 2k + 1 demodulation channels of index j (1 ≤ j ≤ 2k + 1) is best suited to each rising signal i considered.

In order to avoid inter-carrier interference, it is advisable to keep a null carrier between two sub-multiplexes of signals corresponding to separate users. For example, in FIG. 4, there is shown a global signal for which two users use two neighboring subplexes separated by a null subcarrier.

In the above description, the received global signal is transposed to the nominal demodulation frequency upstream of the demodulation channels. Alternatively, the frequency translation (from the frequency of the transmitted signal to the demodulation frequency) could be performed in the demodulation channels. In this case, the transposition could be done using a mixing signal assigned a predetermined frequency deviation from the nominal demodulation frequency, preferably different for each demodulation channel.

Claims

A method for receiving uplink signals from terminals (T1, T2, T3), each uplink signal having been previously modulated by a respective terminal for frequency division multiple access, the method comprising a step of receiving a global signal comprising upstream signals (S300);

the method being characterized in that it further comprises the steps of:

a plurality of demodulation channels (S301) are fed with the global signal received, and

in different demodulation channels, different phase rotations are respectively applied, producing different predetermined demodulation frequency offsets with respect to a nominal demodulation frequency;

for a considered uplink signal, one of the plurality of demodulation channels is selected according to a selected quality criterion (S306).

2. Method according to claim 1, wherein for said considered uplink signal is selected the demodulation channel which has the phase closest to that of the uplink signal, in the sense of a deviation in absolute value the lowest.

3. Method according to one of claims 1 and 2, wherein the upright signal considered comprising pilot symbols with a predetermined phase, it is allocated a demodulation channel according to measurements made on the phase rotations of the pilot symbols. received and demodulated.

4. Method according to one of the preceding claims, according to which the signal amount considered is allocated a demodulation channel according to at least one quality criterion of the amount signal once decoded from at least the quality criteria of the type of rate. binary errors and signal-to-noise ratio.

5. The method of claim 1, wherein, said uplink signal comprising at least one predetermined sequence, it is allocated a demodulation channel according to the extraction quality of said predetermined sequence.

6. Method according to any one of the preceding claims, wherein is performed on each demodulation channel a calculation of a type of Fourier transform on the overall signal.

7. The method as claimed in claim 1, in which a polyphase filtering of the IOTA type is carried out on each demodulation channel (S302).

8. Method according to any one of claims 1 to 6, wherein a Nyquist-type polyphase filtering is carried out on each demodulation channel (S302).

9. A method according to any one of the preceding claims, wherein 2k + 1 demodulation channels are used, where k is a natural integer, and wherein the phase rotation is effected by a frequency offset, the demodulation frequency of the path j, j varying between 1 and 2k + 1, being substantially equal to:

F _j = F _nom + (jk-1) ΔF,

where Fnom is the nominal frequency of reception demodulation and ΔF is a predetermined frequency offset step.

10. The method of claim 9, wherein the predetermined frequency shift step corresponds to a fraction of a gap between two successive carriers.

11. A method of transmitting uplink signals from terminals (T1, T2, T3), wherein

in transmission, each terminal transmits a respective upstream signal on at least one continuous sub-multiplex of a frequency band, and the presence of a carrier of zero amplitude is imposed between each sub-multiplex and

in reception, the steps of the reception method according to one of claims 1 to 10 are carried out.

12. Reception device adapted to receive an overall signal comprising a sum of uplink signals from terminals (T1, T2, T3), each uplink signal being modulated by a respective terminal according to a division multiple access on a frequency multiplex and being intended to be demodulated in reception at a given nominal demodulation frequency,

the device being characterized in that it further comprises

a plurality of demodulation channels (C1, C2, C3) adapted to receive at input a copy of the overall signal and comprising phase rotation means adapted to introduce a predetermined frequency offset with respect to the nominal demodulation frequency, different demodulation channels being adapted to apply different respective phase rotations, and

selection means (8) for assigning a considered uplink signal to a demodulation channel according to a chosen criterion.

Receiving station, characterized in that it comprises

the receiving device according to claim 12.