EP0629059A1 - Spread spectrum digital transmission system with low frequency pseudorandom coding of the useful information and method for spectrum spreading and compressing used in such a system - Google Patents

Abstract

System and method for combining, within a transmitter, each block of a digital signal to be transmitted with a sample originating from a pseudo-random generator operating at low frequency. The results of the various combinations are converted into orthogonal or quasi-orthogonal sequences, modulated and transmitted to the receiver. The receiver performs demodulation of the signal received and combines each sequence with a sample identical to that used for the low-frequency coding within the transmitter in order to reconstitute the various blocks. Makes it possible to carry out low-frequency spectral spreading of a signal to be transmitted. <IMAGE>

Description

The field of the invention is that of digital signal transmission modems and in particular that of spread spectrum modems. More specifically, the present invention relates to a spread spectrum transmission system between a transmitter and a receiver of digital signals in which the spread spectrum is obtained by pseudo-random coding of the useful information to be transmitted. The invention is particularly applicable in wireless telecommunications in the military field.

In the military field, a spectrum spreading operation is generally used in ECCM (Electronic Counter-CounterMeasures) and consists in multiplying the useful signal to be transmitted by a code, called code or sequence of spreading, coming from a pseudo generator - random whose frequency of the clock signal is much greater than the maximum frequency of the useful signal. The number of useful information bits transmitted per Hz is therefore very low.

FIG. 1 represents a timing diagram making it possible to understand the principle of spectrum spreading by a spreading sequence.

A useful signal SAT to be transmitted, here coded on two levels +1 and -1 following an NRZ coding, is multiplied by a cyclic spreading sequence SE, also coded on two levels. The signal resulting from the multiplication is the ST signal transmitted from the transmitter to a receiver after modulation. The transmission medium for the modulated ST signal is generally constituted by a radio link. On reception, after demodulation, the multiplication of the received signal ST with the same spreading sequence SE (same phase and same frequency) makes it possible to reconstitute the useful signal SAT.

Direct sequence spread spectrum transmission is usually used to give the transmitted signal better discretion, resistance to ECM (Electronic CounterMeasures) and resistance to selective fading.

The spread between the chip time and the bit time is defined by spreading gain, the chip time corresponding to the duration of a bit of the spreading sequence and the bit time to that of the useful signal. The higher this spreading gain, the more the transmitted signal is able to be transmitted discreetly and therefore to resist the ECM devices intended to detect it and, possibly, to jam it. An essential step of the ECM analysis consists in determining the spreading hazard of the signal picked up because this step makes it possible to penetrate the information content of the signal picked up, that is to say to reconstruct the useful signal.

The main drawback of spread spectrum by direct sequence is that the generator of the direct sequence must operate at the chip transmission frequency, ie at a frequency of the order of several MHz. It is therefore necessary to install this generator in an ASIC, which increases the hardware complexity and the cost of hardware development.

The object of the present invention is in particular to overcome this drawback.

More specifically, one of the objectives of the invention is to provide a system for transmitting a digital signal, where spectrum spreading is implemented, this system not requiring a hazard generator operating at the chip frequency. It is therefore simpler to produce and less costly, while allowing significant spreading of the spectrum of the useful signal intended to resist ECM devices.

Another objective of the invention is to provide such a system where spectrum spreading is carried out from orthogonal sequences, for example using sequences of M-sequence type (also called sequences of maximum length or of Hadamard ), well known in the field of digital signal transmission.

An additional objective is to provide a method for transmitting digital spread spectrum signals in which the spread is carried out at the bit frequency and not at the chip frequency.

• * l'émetteur comporte successivement :
  • - des moyens de codage recevant ce signal numérique et fournissant, pour chaque bloc de k bits du signal numérique, un échantillon codé prenant une valeur entière comprise dans l'intervalle [0, N-1], chaque valeur entière étant représentative des k bits du bloc dont elle est issue ;
  • - des moyens de combinaison des échantillons codés avec des échantillons issus d'un générateur d'aléas de phase pseudo-aléatoire, les moyens de combinaison fournissant un entier compris dans l'intervalle [0, M-1] pour chaque combinaison d'un échantillon codé et d'un échantillon d'aléa de phase issu du générateur d'aléas de phase, M étant supérieur à N ;
  • - des moyens de génération de signaux fournissant, pour chaque entier compris dans l'intervalle [0, M-1], une suite de q nombres entiers correspondant à cet entier, les différentes suites étant orthogonales ou quasi-orthogonales entre elles ;
  • - des moyens d'émission des suites de q nombres entiers à l'attention du récepteur, les moyens d'émission comprenant un modulateur à décalage de phase dont le nombre d'états est égal à M ;
• * le récepteur comporte successivement :
  • - des moyens de réception restituant les suites de q nombres entiers ;
  • - des moyens de traitement recevant d'une part les suites de q nombres entiers des moyens de réception et d'autre part des échantillons d'aléas de phase issus d'un générateur d'aléa de phase synchronisé avec le générateur d'aléas de phase de l'émetteur, les moyens de traitement assurant une démodulation des suites de q nombres entiers et effectuant une opération inverse de celle des moyens de combinaison pour restituer les échantillons codés ;
  • - des moyens de décodage restituant le signal numérique à partir des échantillons fournis par les moyens de traitement.

These objectives, as well as others which will appear subsequently, are achieved by means of a system for transmitting a digital signal between a transmitter and a receiver, characterized in that:

• * the transmitter successively includes:
  • coding means receiving this digital signal and supplying, for each block of k bits of the digital signal, a coded sample taking an integer value included in the interval [0, N-1], each integer value being representative of the k bits the block from which it came;
  • - means for combining the coded samples with samples from a pseudo-random phase generator, the combination means providing an integer in the interval [0, M-1] for each combination of a coded sample and a phase random sample from the phase random generator, M being greater than N;
  • - means for generating signals providing, for each integer included in the interval [0, M-1], a sequence of q integers corresponding to this integer, the different sequences being orthogonal or quasi-orthogonal to each other;
  • - Means for transmitting sequences of q whole numbers for the attention of the receiver, the transmission means comprising a phase shift modulator whose number of states is equal to M;
• * the receiver successively includes:
  • - reception means restoring the sequences of q whole numbers;
  • processing means receiving on the one hand the sequences of q whole numbers of the reception means and on the other hand samples of phase hazards originating from a phase hazard generator synchronized with the hazard generator phase of the transmitter, the processing means ensuring a demodulation of the sequences of q whole numbers and performing an inverse operation to that of the combining means for restoring the coded samples;
  • - decoding means restoring the digital signal from the samples supplied by the processing means.

The M sequences of q integers are preferably made up of Hadamard sequences.

• - la figure 1 représente un chronogramme permettant de comprendre le principe de l'étalement de spectre par une séquence d'étalement ;
• - la figure 2 est un schéma synoptique d'un émetteur du système de transmission de la présente invention ;
• - la figure 3 est un schéma synoptique d'un récepteur des signaux numériques transmis par l'émetteur de la figure 2.

Other characteristics and advantages of the invention will appear on reading the following description of a preferred embodiment, given by way of nonlimiting illustration, and of the appended drawings in which:

• - Figure 1 shows a timing diagram for understanding the principle of spectrum spreading by a spreading sequence;
• - Figure 2 is a block diagram of a transmitter of the transmission system of the present invention;
• - Figure 3 is a block diagram of a receiver of digital signals transmitted by the transmitter of Figure 2.

Figure 1 has been described above with reference to the state of the art.

Referring to FIG. 2, the digital signal to be transmitted SN is applied, here via a serial access, to coding means 21 which supply, for each block of k bits of the signal SN, a coded sample E _c taking an integer value included in the set {0, ..., N-1}, each integer value being representative of the k bits of the corresponding block. The coding means 21 can for example be constituted by a simple binary-decimal converter and the bit rate leaving the coding means is then k times lower than the bit rate entering.

The coding means 21 can optionally also perform an interleaving of the bits of the signal SN.

The coded samples E _c are applied to means 22 for combining these samples with samples E _a originating from a pseudo-random generator 23, which will subsequently be called phase random generator. In general, the combination means 22 comprise a transformation algorithm which transforms each coded sample E _c into an integer s included in the set {0, ..., M-1}, with M integer greater than N. We at :

where f is any function taking its values in {0, ..., M-1} and E _has a sample of phase randomness.

M
où ^e désigne l'addition modulo M pouvant aussi s'écrire :

The combination means 22 can for example be constituted by a simple modulo M adder, as shown and providing:

M
where ^e denotes the addition modulo M which can also be written:

This addition modulo M, apart from the fact that it can be implemented by an algorithm very simple to implement, provides optimum performance of resistance to ECM interference.

Each integer is then supplied to means 24 for generating signals providing, for each integer, a corresponding sequence SQ of q samples, each sample q being an integer. The signal generation means 24 transform each integer s into a sequence SQ, this transformation being one-to-one, that is to say that to a given integer s corresponds a single sequence SQ and vice versa.

où bs est un entier compris entre 0 et L-1.We can write :

where bs is an integer between 0 and L-1.

The signal generator can for example be constituted by a transcoding table. We will usefully refer to French patent no. 2,337,465 in the name of COMPAGNIE IBM FRANCE TM which describes sequences called CAZAC which are periodic pseudo-random sequences of complex numbers which have a periodic autocorrelation function of which only the first coefficient is nonzero and of which all complex numbers have a constant amplitude. The generation of such sequences can be generalized to obtain sequences consisting of whole numbers, these sequences being orthogonal to one another, that is to say having optimal autocorrelation properties. We can also mention the sequences of Gold which are quasi-orthogonal, like those of Kasami, or those called polyphases.

In a preferred embodiment, the means 24 generate sequences SQ which are substantially orthogonal to one another. By way of example, the signal generation means 24 can transform each integer s into a sequence SQ of q bits (samples each taking a value in {0,1}) according to table 1 below.

In this configuration, M = 8 and q = 7. Each sequence of q bits comes from circular shifts of a sequence of maximum length of length 7, excluding the first sequence always made up of zeros. These sequences have quasi-orthogonality properties, that is to say that for two different and arbitrary sequences, the sum of the EXCLUSIVE OUs of each term is equal to 4.

It is possible to generalize this principle of generating quasi-orthogonal SQ signals for any M power of 2. For this, after having determined a sequence of maximum length of period M-1 (by one of the methods well known in the field of processing digital signal), the M sequences of M-1 bits are obtained by circular shifts of the initial sequence, with the exception of the first always consisting of zeros.

Another class of usable sequences, perfectly orthogonal, is that constituted by Hadamard sequences. An example of such sequences is illustrated in table 2, for samples also constituted by bits.

The length of these sequences or sequences is 8.

The preceding description shows that each block of k bits of the signal SN has been transformed into a corresponding sequence SQ, each sequence SQ comprising a pseudo-random component. Useful information is coded in this SQ sequence and the different sequences are orthogonal or quasi-orthogonal to each other. As soon as M and q are large in front of k or in front of N, we understand that this coding operation consisted in significantly increasing the number of samples to be transmitted and that the spectrum of the useful signal SN was therefore spread out using hazards provided at low frequency.

The main advantage of the invention lies precisely in this coding which is carried out at the bit frequency and not at the chip frequency (where the spectrum spreading is carried out by direct sequence). The working frequency of the means described so far can thus be very low, of the order of 16 Kbits, to be compared with 10 Mchips in the case of spread spectrum by direct sequence.

It should be noted that the samples can take larger values, as a function of the modulation used in transmission means 25 to which the suites SQ are supplied.

These transmission means 25 supply a signal STR transmitted to the attention of the receiver. They can be of any type, analog or digital.

In the embodiment shown, the transmission means 25 are of digital type and include a phase shift modulator 28. This modulator 28 is for example of MPSK (Multiple Phase Shift Keying) type where M corresponds here to the number of values possible samples q of the sequences SQ and therefore the number of phase states of the modulated signal STR. It is for example possible to perform a BPSK modulation if the SQ sequences consist exclusively of bits, a QPSK modulation if the integers of the SQ sequences are each included in the set {0, 1, 2, 3}, and a modulation 64-PSK if the integers of the SQ suites are each included in the set {0, 1, ..., 63}. The phase shift modulator 28 can also be of the QAM type. It provides a modulated signal denoted SM.

The transmission means 25 may also include means 26 for spreading spectrum by spreading sequence. The spreading sequence SE is generated by a spreading sequence generator 27. In the embodiment shown, it is assumed that the bits of the sequences SQ take their values in {0,1} and that the chips of the sequence of spreading SE also take their values in {0,1}. Each sample b ^s _i produced by the signal generation means 24 is added modulo L to G random elements e, belonging to the set {0, 1, ..., L-1} and coming from the generator 27, where G represents the spread gain by direct sequence. The increase in bit rate caused by this processing is equal to G. In the case of a spread of spectrum by direct sequence, it is therefore the output signal of the means 26, denoted SQE, which is applied to the modulator 28.

Each sample has an SQE sequence takes its value in {0, 1, ..., L-1}. In the case where no spreading by direct sequence is implemented, G = 1 and e _s = 0, that is to say that this operator is transparent.

où g est la fonction de mapping réalisée par le modulateur 28, Ts le temps symbole et h_e(t-iTs) le filtrage émission. A titre d'exemple:

• - en modulation BPSK, L = 2 et on a g(0) = -1 et g(1) = 1

The STR signal sent to the receiver is of the form:

where g is the mapping function performed by the modulator 28, Ts the symbol time and h _e (t-iTs) the emission filtering. For exemple:

• - in BPSK modulation, L = 2 and we ag (0) = -1 and g (1) = 1

avec a_i = 0 ou 1

• - en modulation QPSK, L = 4 et
  • 9(0) = 1, g(1) ⁼
  • g(2) = -1 et g(3) =
• - en modulation 8PSK, L = 8 et g(k) = e^2jkπ/8

In this case the relation 1 is written:

with a _i = 0 or 1

• - in QPSK modulation, L = 4 and
  • 9 (0) = 1, g (1) ⁼
  • g (2) = -1 and g (3) =
• - in 8PSK modulation, L = 8 and g (k) = e ^{2jkπ / 8}

Generally, in MPSK modulation, L = _M and g (k) = 2 ^{jkπ / M.}

lorsqu'un étalement par séquence directe est mis en oeuvre (G > 1).Note that the mapping function g of the modulator must respect the relationship:

when a spread by direct sequence is implemented (G> 1).

et

pour k *- 0 (critère de Nyquist).The impulse response h _e of the emission filter is assumed such that:

and

for k * - 0 (Nyquist criterion).

The means 26 for spreading spectrum by direct sequence are of course optional in the invention and are therefore shown in broken lines.

The transmission means 25 can also include frequency escape means 29, 30, also optional and therefore shown in broken lines, capable of modifying the carrier frequency of the signal transmitted to the receiver. Frequency evasion consists of frequently changing the carrier frequency in order to further broaden the spectrum of the signal transmitted to the receiver. The modulated signal SM, in baseband or in intermediate frequency, is applied to a multiplier 29 receiving a carrier frequency signal from a generator 30.

It can be seen that the phase hazard generator 23 allows low-frequency coding of the signal to be transmitted and makes it possible to pseudo-randomly modify the phase of the signal transmitted when the modulation is of MPSK type. It can thus be considered that the generator 23 and the combination means 22 provide a phase escape function performed at low frequency. An amplitude modulation, also pseudo-random, of the signal to be transmitted is combined with this phase escape when the modulation is of the QAM type (modification of the phase and of the amplitude of the transmitted signal). This is how the transmission system of the invention makes it possible to obtain significant resistance to ECM interference.

The output signal STR of the transmission means 28 is transmitted over the air to the receiver 31, the block diagram of which is given in FIG. 3.

The receiver 31 receives a signal STRr corresponding to the signal STR noisy by the transmission medium. It comprises reception means generally referenced by 40 restoring the sequences SQ of q whole numbers, denoted SQr at the level of the receiver. The reception means 40 here comprise means 32 for suppressing the carrier frequency controlled by a local oscillator 33. The means 32 conventionally comprise two mixers controlled by clock signals in quadrature and two signals are obtained at the output of these means. quadrature. When a frequency escape is used at the level of the transmitter 20, the local oscillator 33 operates in synchronism with that of the transmitter, referenced 30. This synchronization can be obtained by known means. The output signal of the means 32 is noted SMret corresponds to the signal SM of the transmitter.

The signal SMr is applied to spectrum compression means 34 intended to suppress spreading by direct sequence possibly carried out at the level of the transmitter 20. Spectrum compression means are notably described in "Digital Communications" by JG PROAKIS, McGraw-Hill ^TM chapter 8. Those represented in FIG. 3 include a sampler 35 controlled at the frequency chip Fc followed by a module 36 for spectrum compression. The module 36 comprises a complex multiplier 37 followed by an adder 38. The multiplier 37 receives a direct sequence SE from a generator 39, this direct sequence SE being identical to that generated by the generator 27 of the transmitter 20. The setting phase of these two direct sequences is obtained by known means.

où e_sk est la valeur du chip à l'instant k de la séquence directe SE et ^* désigne le complexe conjugué. Cette sommation permet de supprimer l'étalement spectral par séquence directe.The summator 38 calculates, for each block of G consecutive samples r _k from the multiplier 37, the following sum:

where e _sk is the value of the chip at time k of the direct sequence SE and ^* denotes the conjugate complex. This summation eliminates spectral spreading by direct sequence.

Each sum U _k therefore corresponds to a sample α _i of the signal STR transmitted to the receiver. At the output of module 36, there are therefore suites SQr identical to the suites SQ originating from the means 24 for generating signals from the transmitter 20.

These SQr suites are applied to means 45 for processing that have the function of performing a demodulation of the received signal and removing the random phase E _is introduced at the transmitter 20 by the generator 23 hazards.

pour s = 0 à M-1.In the embodiment shown, the processing means 45 comprise correlation means 41 which calculate, for each block of Q successive sums U, the following value:

for s = 0 to M-1.

The correlation means 41 receive for this a reference signal SR constituted by the various sequences SQ which can be generated at the level of the transmitter 20, that is to say those for example represented in Tables 1 or 2. The advantage of generating orthogonal or quasi-orthogonal sequences using the generator 24 of FIG. 2 (and not any sequences) is that it is easy to detect a correlation of these signals.

The calculated correlations provide sums C ⁰ to C ^M-1 which each correspond to one of the integers from the combining means 22 of the transmitter 20. These sums are applied to a demultiplexer 42 receiving from a generator 43 a signal E _a identical to that generated by the generator 23 of the transmitter, and in phase with it.

The demultiplexer 42 selects N sums C ^S from M as a function of the value of the hazard E _a . In general, the demultiplexer42 performs an inverse function f ^-1 to suppress the phase hazard introduced at low frequency on transmission.

le démultiplexeur 42 fournit en sortie les signaux:

pour i = 0 à N-1 et E_a appartenant à l'ensemble {0, 1, ..., M-1}. Le démultiplexeur42 sélectionne ainsi les échantillons C^S en fonction de l'aléa E_a.By way of example, if the combining means 22 produce:

the demultiplexer 42 outputs the signals:

for i = 0 to N-1 and E _a belonging to the set {0, 1, ..., M-1}. The demultiplexer42 thus selects the samples C ^S as a function of the hazard E _a .

Each sample d _i therefore corresponds to a sample E _c of the emitter. These samples d _i are then applied to decoding means 44 performing an inverse operation to that of the coding means 21 of the transmitter 20. They can also carry out a deinterlacing of the decoded samples if the coding means perform an interleaving of the samples coded. The output signal SNr of the decoding means 44 then corresponds to the digital signal SN of the transmitter.

Of course, other embodiments of the processing means 45 can be envisaged. It is for example possible to calculate only the samples d _i according to the relation:

This direct calculation makes it possible not to use a fast correlation algorithm and therefore to simplify the practical implementation of the receiver. Only the useful correlations are then calculated. The means of treatment 45 then only comprise correlation means such as 41, receiving the signal E _a .

The present invention applies for example to transmission systems where error correcting codes are used and where an alphabet of orthogonal signals of very large size, greater than the alphabet used by the error correcting code, is available. Elements of the alphabet not used by the code can be used for low-frequency pseudo-random coding of the signal to be transmitted, thus making it possible to improve the robustness of the system with respect to interception at low cost.

Claims (8)

1. System for transmitting a digital signal (SN) between a transmitter (20) and a receiver (31), characterized in that: * said transmitter (20) successively comprises: - coding means (21) receiving said digital signal (SN) and supplying, for each block of k bits of said digital signal (SN), a coded sample (Ec) taking an integer value included in the interval [0, N -1], each integer value (Ec) being representative of the k bits of the block from which it comes; - means (22) for combining said coded samples (Ec) with samples (Ea) coming from a generator (23) of pseudo-random phase hazards, said means (22) for combining providing an integer (s) included in the interval [0, M-1] for each combination of a coded sample (Ec) and a sample (Ea) of phase hazard from said generator (23) of phase hazards, M being greater than N; - means (24) for generating signals providing, for each integer (s) included in the interval [0, M-1], a sequence (SQ) of q integers corresponding to this integer (s), the different sequences (SQ) being orthogonal or quasi-orthogonal to each other; - Means (25) for transmitting said sequences (SQ) of q whole numbers to the attention of said receiver (31), said means (25) for transmitting comprising a phase shift modulator whose number of states is equal to M; * said receiver (31) successively comprises: - reception means (40) restoring said sequences (SQr) of q whole numbers; - processing means (45) receiving on the one hand said sequences (SQr) of q whole numbers of said receiving means (40) and on the other hand samples (Ea) of phase hazards originating from a generator ( 43) of phase hazard synchronized with said generator (23) of phase hazards of said transmitter (20), said processing means (45) ensuring a demodulation of said sequences (SQr) of q whole numbers and performing a reverse operation of that of said means (22) for combining to restore said coded samples (d _i ); - decoding means (44) restoring said digital signal (SNr) from said samples supplied by said processing means (45). 2. System according to claim 1, characterized in that said M sequences (SQ) of q integers are Hadamard sequences. 3. System according to one of claims 1 and 2, characterized in that said transmission means (25) comprise means (26, 27) of spread spectrum by spread sequence (SE) and in that said reception means (40) comprise spectrum compression means (34) operating in synchronism with said spectrum spreading means (26, 27) of said transmission means (25). 4. System according to one of claims 1 to 3, characterized in that said transmission means (25) comprise frequency escape means (29, 30) capable of modifying the carrier frequency of said signal transmitted to said receiver ( 30) and in that said reception means (40) comprise means (32, 33) performing a function opposite to that of said frequency escape means (29, 30), capable of suppressing said frequency escape introduced to said transmitter (20). 5. System according to one of claims 1 to 4, characterized in that said coding means (21) also perform an interleaving of the bits of said digital signal (SN) and in that said decoding means (44) also perform a deinterlacing of decoded samples (d _i ). 6. System according to one of claims 1 to 5, characterized in that said combining means (22) of said transmitter (20) provide, for each coded sample (Ec), an integer (s) equal to:

or: - s is said integer supplied by said combining means (22); - E _c is said coded sample; - E _a is a sample of phase hazards from said generator (23) of phase hazards of said transmitter (20); ^- M ⊕ denotes the addition modulo M, with integer M; and in that said means for removing said phase hazard from said receiver (30) provides, for each sequence (SQe) of q bits originating from said processing means, an integer (d _; ) equal to:

where E _a is a sample of phase hazards from said generator (43) of phase hazards from said receiver (31). 7. A spread spectrum transmission method of a digital signal between a transmitter (20) and a receiver (30), characterized in that it consists in: * at said transmitter (20): - generate, for each block of k bits of said digital signal, a coded sample (Ec) taking an integer value included in the interval [0, N-1], each integer value being representative of the k bits of the corresponding block; - combining said coded samples (Ec) with phase random samples (Ea) to generate an integer (S) in the interval [0, M-1] for each combination of a coded sample (Ec) and a phase random sample (Ea), M being greater than N; - generate for each integer (s) in the interval [0, M-1], a sequence (SQ) of q corresponding whole numbers, according to a one-to-one transformation, the different sequences (SQ) being orthogonal or quasi-orthogonal between them; - transmitting said sequences (SQ) of q integers to said receiver (30); * at said receiver (30): - reconstruct said sequences (SQr) of q integers from the signal received from said transmitter (20) and generate, for each sequence (SQr) of q reconstituted whole numbers, an integer according to a reverse transformation from that carried out at said transmitter ( 20); - Combine each integer generated with a phase random sample (Ea) identical to that which made it possible to obtain this integer at the level of said transmitter (20), so as to restore the corresponding coded sample (d _; ), said combination thus eliminating said phase hazard (E _a ); - decode each coded sample (d _; ) so as to restore said digital signal (SNr). 8. Method according to claim 7, characterized in that said sequences (SQ) of q integers are Hadamard sequences.

Images