WO2005062493A1

WO2005062493A1 - Device and method for transmission channel distortion equalisation in communication systems involving time domain multiplexing using sets of complementary sequences

Info

Publication number: WO2005062493A1
Application number: PCT/ES2004/000554
Authority: WO
Inventors: Vicente Diaz Fuente; Daniel Hernanz Chiloeches; Jesus Berian Mugica; Diego Lillo Rodriguez
Original assignee: Vicente Diaz Fuente
Priority date: 2003-12-17
Filing date: 2004-12-13
Publication date: 2005-07-07
Also published as: ES2244305A1; ES2244305B1

Abstract

The invention relates to a device and method for transmission channel distortion equalisation in communication systems involving time domain multiplexing using sets of complementary sequences. The inventive method consists in transmitting a medium-identifying preamble pulse s[n] which is encoded by convoluting said signal with a set of complementary sequences (2). According to the invention, upon being received simultaneously at the receiver (11), the aforementioned sequences can be used to extract the time characteristics from the medium (13 and 14) that affect the data transmitted. Using the data thus obtained, post-equalisation (15) and pre-equalisation (16) are performed a given number of times (17) until ideally a Krönecker delta (18) is obtained. The precision of the distortion equalisation depends on the length of the sequences used for coding and on the number of times that said technique is applied to the signal received.

Description

DEVICE AND METHOD FOR THE EQUALIZATION OF THE DISTORTION OF THE TRANSMISSION CHANNEL IN COMMUNICATION SYSTEMS BY MULTIPLEXATION IN TIME THROUGH COMPLEMENTARY SEQUENCE SETS

D E S C R I P C I Ó N

OBJECT OF THE INVENTION

The present invention relates to a device and a method of equalizing the distortions produced by the transmission medium in a data transmission by means of complementary sequence coding techniques; that is, it refers to a transmitter-receiver device that obtains the equalization of distortions by transmitting a Krónecker delta.

BACKGROUND OF THE INVENTION

The communication, spectral analysis, RADAR and SONAR systems transmit a signal that arrives, reflected or not, at the receiver after passing through a transmission medium. This medium behaves like a linear filter with an impulse response in frequency H (τπ) or temporal h [n].

In order to enable the recovery process of the information broadcast, in most communication systems it is essential to eliminate the effects produced by the transmission medium on the broadcast signal s [n] '. This process is known as equalization. The frequency response can also be used to make a spectral analysis of the medium and thus obtain information

LEAF OF of its physical properties.

The channel acts as a filter and distorts the signal. To this we must add the noise, n [n], due to disturbances in the channel, thermal noise or other signals that interfere with those emitted. In conclusion, the received signal, r [n], can be modeled as: r [n] = s [n] * h [n] + n [n] (1)

where * denotes a convolution.

To eliminate the distortion introduced by the medium in the signal, a filter with an impulse response, f [n], is required, such that:

that is, that the received signal is as similar as possible to the one emitted. This is never fully accomplished because equalization does not eliminate noise, n [n], or distortion completely.

To achieve equalization as well as possible, it is necessary to know the medium a priori. In other words, it is essential to analyze the h [n] of the medium in order to counteract the distortion effects. There are two methods to achieve this goal: • Static equalizers: their properties do not change over time.

• Adaptive EQs: Adapts to temporal variations in media distortion. The main problem with the former is that they are more generic and do not solve the particular problems of each situation. Adaptive EQs respond better to variations in the environment, but their implementation is more complicated and they are very sensitive to noise.

Both for some and for others knowledge of the transmission medium is still essential. The better this can be modeled, the more precision will be achieved when restoring the emitted signal.

The ideal method for analyzing the medium is to transmit a delta and analyze what is received, that is, obtain the impulse response.

Digitally this is achieved by issuing a delta of

Krδnecker, δ [n]:

As observed, the received signal has information on the impulse response, h [n], contaminated with additive noise.

This information can be obtained by using complementary sequences as explained in the Spanish patent P200201151, "Method for optimal estimation of the transmission spectrum by simultaneous modulation of complementary sequences". However, it does not explain how to use this information to correct the distortions of the medium on the data.

HAVE TO REPLACE RULE 26 From all of the above, the need for a technique that allows using the information obtained from the medium to correct the distortion effects in the transmitted data is deduced.

The existence of any patent or utility model is not known, the characteristics of which are the object of the present invention.

DESCRIPTION OF THE INVENTION

The invention presented uses the result obtained by the method described in P200201151 to equalize the data received in a communication system based on GCM / OTDM (Golay Coding Modulation / Orthogonal Time Division Multiplexing) such as that described in the Spanish patent P200002086 of August 16, 2000, "Method, Transmitter and Receiver for Spread Spectrum Communications by Golay Complementary Sequence Modulation".

In summary, the process of obtaining the information and its subsequent treatment, based on the signal received, is described below.

First of all, a preamble is generated that will allow us to synchronize and subsequently obtain the information of the medium necessary for the correction of the received data. The preamble is structured as shown in figure 1.

Interval 1 corresponds to the synchronization interval, during which the reception system synchronizes with the transmission system. Once synchronized, interval 2 corresponds to the extraction of information

SUBSTITUTE SHEET RULE 26 distortion of the medium on the data received later during interval 3. For this, at least one impulse δ [n] encoded by a pair of complementary sequences is sent by means of convolution with at least one pair of complementary sequences (A, B ).

The simplified result of this process, in the frequency domain, is as follows: Tx (ω) = A (ω) * eos ω _c t + B (ω) * sinω _c t (4)

The effect is to modulate both sequences in quadrature at the central frequency ω _c . When transmitted to the medium H (ω), in reception we obtain:

Rx (ω) = [A (CO) * eos ω _c t + B (ω) * sinω _c t] • H (ω) + N (ω) [5)

Since the demodulation of this stage is coherent, since the system is synchronized, we extract the following phases in quadrature:

Rx i (ω)

fí (ω) * cosω _c t + N (ω) * cosco _c t Rx ₀ (co) = [A (ω) * eos ω _c t + B (ω) * sena> _r t - H (co) * sinω _c t + N {ω) * sinω _c t (5) In the demodulation process, assuming that we are already synchronized, a component appears in 2ω _c that is low-pass filtered, obtaining only the original base bands convolved with two new means, H _Q \ ω) and

H, {ω): D _lu (ω) = A (ω) -H _! (ω) + N, (ω) D _RQ {ω) = B {ω) -H _Q {ω) + N _Q (ω)

where N _υ (ω) and N, (ω) are the input noise modulated at the central frequency. Now we apply the correlation with the transmitted sequences:

D „(ω) = A (ω) - A (- ω) - H _l (ω) + N _l {ω) - A (- ω) D _RQ (ω) = B {ω) B {- ω) - H _Q (ω) + N _Q (ω) - B {- ω) Obviously, the objective of this stage would be to extract the coefficients of Hχ (ω) and H _Q (ω), and to equalize both phases independently, however, given that the data is coded so that the sum of both phases must give a Krónecker delta, we must equalize the following sum expression:

D _R (ω) = A (ω) • A (- ω) • H, (ω) + N, (ω) • A. (- ω) + B (ω) • B {- ω) • H _Q ( ω) + N _Q (ω) • B (- ω) (8) To understand the following process, we must indicate that the main property of the sequences used in this invention is that they possess an ideal autocorrelation characteristic, that is, they correspond to a delta Krónecker's perfect so they meet:

where individualesü are the individual autocorrelations of each of the M complementary sequences, of length, chosen. Particularized for the case of pairs of complementary Golay sequences (A, B):

The generation of such sequences is carried out from the so-called basic kernels known to date of 2, 10 and 26 bits (the generation rules of

HA DE TIT CI NR Golay sequences are discussed in the article entitled "Complementary Sequences" by MJE Golay, published in IRÉ Transactions on Information Theory, vol. IT-7, pp 82-87, April 1961).

Once the previous expression has been sampled at the symbol frequency (figure 2), and given the properties of the complementary sequences described, we will consider that (8) can be written again in discrete time as: d _R [n] = 2Lδ [n ] * h _r [n] + n _r [n] - (11)

where h _{τ [} [n] is a new media transfer function that distorts the transmitted data.

The distorted answer can be mathematically expressed according to the expression: y [k] = s [k] h _τ [Oj + - i] + n _τ [i] (12)

In this expression the first term represents the symbol at time k, the second term corresponds to the interference between adjacent "precursor" symbols, the second term corresponds to the interference between adjacent "postcursor" symbols and, finally, n _τ corresponds to the noise obtained at the output of the demodulation process To equalize the channel we will divide it into two stages (figure 2 and 3):

• Postcursor equalization (after the maximum of the sampled channel response).

HAVE TO REPLACE RULE 26 • Precursor equalization (before the maximum of the sampled channel response).

The postcursor stage consists of the complete elimination of the interference between adjacent symbols (ISI) that are to the "right", in time, of the symbol k. This cancellation corresponds to a "zero-forcing" linear equalizer approach, or all zeros. Here, the desired symbol is s [k] where k represents the Jth sample of an infinite sequence of symbols.

If we consider the response to the causal impulse of finite duration M, the ideal "postcursor" equalizer will correspond to:

From expression (11), it is evident that the acquired samples d _R ([n] correspond to the medium transfer function h _τ ¡[n] multiplied by a factor dependent on the length of the sequences, added a given noise, and therefore coincide with an estimator of the filter coefficients that models the transmission medium h _{T [} [n], Substituting and applying the z transform, the equalizer response corresponds to the following expression:

TIT CI N RE SHEET 26 Thus, the "postcursor" equalization process will be to apply the previous filter to the received processed signal, which will allow us to eliminate the interference between "postcursor" symbols. The filter can be implemented digitally in a simple way using any of the existing direct structures. The result of filtering the signal d _E [[n] by means of filter (14) will correspond to a signal that we will call d _E [[n] and that, ideally, will not have "postcursor" interference

Once the "postcursor" interference has been eliminated, we must eliminate the "precursor" interference. Assuming that the duration of the "precursor" is N, and that d _E [[0] coincides with the maximum of the signal (which must be 1 when applying the filter (14)). Equalization is carried out using an FIR filter. The filter coefficients, h _F [[n], are extracted from the resolution of a system of N equations with N unknowns defined by:

The value of N is always greater than or equal to the length of the sequences used. The resolution of this system of equations is direct, because below its main diagonal all the values are zero and, in addition, h _{F [} [Nl] = 1.

Since the result of both processes is not ideal due to noise, quantification effects,

HA DE TIT CI NR rounding, etc., if we apply this technique iteratively, we obtain an equalization that improves each iteration.

This technique allows to equalize any linear channel, including any distortion effect such as, effects due to multiple paths, reflections, non-ideal circuits and filters, etc. Obviously, the selection of the length of the N and M filters will depend on the length of the channel response, and on the length of the sequences used.

In conclusion, it can be affirmed that the invention described constitutes a powerful distortion equalization system in communication systems, especially oriented to GCM / OTDM techniques and their variants.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1. - Shows the preamble generated for the extraction of the distortion information:

1. In this phase the receiver synchronizes with the transmitter in phase, time and symbol, so that in the next stage the receiver knows exactly where the received equalization data begins.

2. In this phase a synchronous or coherent demodulation is carried out, due to the knowledge in the previous phase of the synchronism information, so that the coefficients of the equalizing filters are obtained.

HA DE TIT CI NR 3. In this last phase, the equalizing filters obtained in the previous phase are applied, and the data is free from Inter-Symbol Interference (ISI) and reduced noise.

Figure 2. - Shows the signal received in the ADC sampling stage d _R [n] and its subsequent sampling at the symbol frequency, from which the distortion produced by the adjacent symbols h _F [n] is obtained directly, ISI.

Figure 3. - Shows the block diagram of a system that explains a possible application of estimating the distortion of the medium and its subsequent correction using a single pair of complementary sequences.

The different parts that compose it are detailed below:

1. Digital signal to emit s [n]: to estimate the medium, the ideal is a Krónecker delta. 2. Encoder with a pair of complementary Golay sequences. 3. Signals resulting from coding I [n] and Q [n]. 4. QASK modulator. Modulates the signal I [n] in phase and the Q [n] in quadrature. 5. Signal resulting from QASK Tx [n] modulation. 6. Radio frequency modulator. 7. Antenna. 8. Antenna. 9. Radio frequency demodulator. 10. Signal resulting from radio frequency demodulation Rx [n]. 11. QASK demodulator. Results in r _τ [n] and r _Q [n].

HAVE TO REPLACE RULE 26 12. Signals resulting from demodulation QASK r _τ [n] and r _Q [n]. 13. Decoder with a pair of complementary Golay sequences and obtaining d _R [n]. 14. Sampling at symbol frequency for the extraction of the coefficients of h _F [n]. 15. "Postcursor" equalization. 16. "Forerunner" equalization. 17. Iteration, if necessary, to block 16. 18. Equalized final result.

Figure 4. - Shows the block diagram that solves the system of equations and obtains the coefficients of the pre-equalizer filter. It is made up of a basic structure formed by a register, change of sign and divisor that is repeated regularly depending on the order of the system of equations. The captured "precursor" data are serially inverted in time d _R [-n]. The coefficients of the pre-equalizer filter h _F [n] are calculated from the bottom output at each clock cycle.

PREFERRED EMBODIMENT OF THE INVENTION

A possible implementation of this technique applied to the equalization of a communication system using OTDM as a data modulation technique is detailed below.

For the sake of clarity, the implementation is shown in figure 3. This implementation is based on the application of this method to radio frequency systems.

To simplify the explanation, we have resorted to the particular case of pairs of complementary sequences Golay modulated in QASK (Quadrature Amplitude Shift Keying). The system consists of two well differentiated blocks: the transmission system and the reception system. The transmission system is responsible for:

• Convolve the input signal with each of the two sequences that make up the pair of complementary sequences of length L and generate the equalization preamble.

• Modulate in QASK the two signals resulting from the encoding with the complementary sequences.

• Modulate the quadrature signal using QASK (Quadrature Amplitude Shift Keying) or QAM (Quadrature Amplitude Modulation) to transmit it in the corresponding area of the radio spectrum. • Transmit it with an antenna.

The reception system is in charge of: • Synchronizing with the transmitter and demodulating the signal received by the antenna.

• Obtain the components r _τ [n], in phase, and r _Q [n], in quadrature, using the QASK demodulation.

• Detect the beginning of the equalization preamble and perform the decoding process using correlation sums, as presented in this document.

TIT CI N RE SHEET 26 • Obtain the coefficients of the initial equalizer filter (DISTORTION) from the sum.

Perform the "Post-upgrade" and "Pre-upgrade" steps described.

• Repeat the previous process iteratively until obtaining an equalized signal within the limits established by the system, which will correspond to a Krónecker delta in the ideal case.

The resulting filters together form the inverse filter of the distortion suffered by the data in the transmission, propagation and reception process.

H A DE TIT CI N R

Claims

1.- Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized in that the device, emitter-receiver, allows to emit signals through a physical medium that It includes the generation of a set of complementary sequences whose main property is that the sum of the autocorrelations φn, of the sequences that form the set is a Krónecker delta.

2. Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized by the first claim and also because the complementary sequences used have the following characteristics:

• Any length L.

• They are issued using any symbol width, T, with any amplitude and with any level of oversampling.

• They are issued in parallel with other sets of sequences complementary or not to the previous ones.

• They are issued simultaneously using a frequency, phase or amplitude modulation, or combinations thereof.

SHEET OF TIT CI N RE LA 26

3. Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized by the preceding claims and also because the complementary sequences are emitted and received, after propagating to through the middle, using any type of transducer or antenna.

4. Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized by the preceding claims and also because the method is based on the extraction of information obtained by means of the transmission of a Krónecker delta encoded by complementary sequences transmitted simultaneously to a physical medium that, when received in a receiver, it obtains data of the distortion produced and which are used to design the equalizer filters that compensate for said distortion.

5. Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized by the preceding claims and also because the complementary sequences are used to transmit signals to a medium with in order to obtain an impulse response h [n] or frequency response H (ω).

6.- Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized by claims I ^a and 5 ^a and further because the method encodes one or more deltas of Krónecker with identical or different amplitudes and any temporal and frequency combination in order to implement in the device.

7. Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized by the preceding claims and also because the method for generating the coding with sequences complementary to the object of implementation includes at least:

• The convolution, using any method, of the input signal with each of the complementary sequences that make up the set.

• The emission of the signals resulting from the convolution.

8. Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized by the preceding claims and also because the method for obtaining the coefficients of the temporal or frequency distortion , by decoding with complementary sequences comprises at least:

• The correlation or filtering adapted, using any method, of the signals received at the input of the decoder with each of the sequences complementary that make up the set used in the issue.

• The sum of the results of the resulting correlations to obtain the characteristics of the medium.

• Storage in a memory of said result for further processing.

9. Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized by the preceding claims and also because to equalize the received data comprises at least:

• The decoding of claim 8 ^a , and its application as coefficients of the "post-qualifier" reverse filter that corrects the distortion or interference between symbols (ISI) "postcursor" due to the symbols subsequent to the instant of the decision.

• The resolution of a system of equations and its application as coefficients of the "pre-equalizer" inverse filter that corrects distortion or interference between symbols (ISI) "precursor" due to the previous symbols at the time of the decision.

• The application iteratively, that is as many times as necessary, of the two previous sections until obtaining improvement in equalization within certain limits defined by the system.

10. Device and method of equalization of the distortion of the transmission channel in time multiplexing communication systems by means of a set of complementary sequences, characterized by claim 9 ^a and also because to solve the system of equations set forth in the previous claim and to obtain the coefficients of the "pre-equalizer" filter you need at least:

• The introduction of the recovered data, after the post-qualification phase, invested in time, to a bank of N records, depending on the order of the system of equations to solve.

• The cascading of the data records that make up the record bank and its initial zeroing.

• The use of auxiliary records to store the N-2 coefficients obtained in the successive cycles of the process, and their initial zeroing.

• The realization at each clock cycle, k, of the sum of all products corresponding to the values calculated in the previous cycle, h _E [kl], found in the auxiliary registers, by the data d _R [-k + l], which are within the data records,

SUBSTITUTE SHEET (RULE 26) changed sign, thus obtaining the value of the coefficient h _E [k].

• The storage, at each clock cycle, of the calculated results h _E [k], in the corresponding auxiliary registers, to be used in the next cycle.

• The connection of the first register, entry of the register bank, with a multiplexer that allows to control the change of sign of the signal with which the described system processes.