FR2573594A1 - Spread spectrum method of transmission in a communications network, and transceiver set intended for the implementation of this method - Google Patents

Abstract

The spread spectrum, in the method of transmission according to the invention, is obtained by setting up the communications within the network on carrier signals modulated by the information, having frequencies which are continuously variable in accordance with pseudo-random sequences of successive continuous functions fi(t) selected from a predetermined group of possible functions, each sequence being completely determined by a key dependent on the set requested in the network and by the time in the network. Each of the transceiver sets intended for the implementation of this method of transmission includes frequency generators 10, 110, with variable frequency, which may be frequency synthesizers which can be modulated at high speed, and controlled by clocks 12, 120 giving the local time, by pilot oscillators 10, 100 giving a reference frequency, and by key generators 13, 130 determining the pseudo-random sequence of continuous functions giving, for a communication, the frequency of the carrier signal for transmitting information or the frequency of the local signal for demodulating the information. Application, in particular, to multiaccess networks in which it is desired to set up discreet links, or to have good protection against interference.

Description

Spread spectrum transmission method in
a communications network, and transceiver station
intended for the implementation of this process.

 The invention relates to networks of transceiver stations with multiple access, and more particularly to a spread spectrum transmission method in such a network, and to stations intended for the implementation of this method.

For the transmission of communications in a frequency band allocated between transceiver stations of a network, several transmission methods allowing multiple access to the network are currently known: multiple access to the network can be obtained by time division (TDMA) ) between the stations, by frequency distribution (AMRF), or even in simultaneous access, and distribution by coding (AMRè), this coding being able to consist in establishing for each transmitting station a law of frequency jumps (EVF, for frequency evasion) , or coding by rapid phase modulation (PN, for pseudo-noise according to English terminology). The most classic processes, AMRT or
AMRF does not provide sufficient protection against interference to the communications system; moreover, they require extensive coordination of the network.

 On the contrary, CDMA type processes lend themselves well to the establishment of links protected against interference. Indeed, the links are protected due to frequency hopping (EVF) or phase hopping (PN) or combinations of the two, these hopping appearing as random hops for third parties who do not have the law of variation. This type of process leads to a widening or "spreading" of the spectrum occupied by the signal which makes it possible to improve the protection against interference. Such methods can be used in particular in tactical satellite telecommunications systems, but they require sophisticated procedures and control systems, in particular because the different stations of the network must be synchronized.

 In addition for demodulation, these methods use either passive correlators, charge transfer devices (CCD), surface acoustic waves (SAW), or fast digital processors, or active correlators. In the first case, the processing gain is limited, due to technological limitations, and the frequency errors lead to reserving this type of transmission at high bit rates, there is little protection against interference; in the other case the acquisition times are high.

Finally, the distribution methods by coding based only on frequency evasion (EVF), if they allow to have fairly short acquisition times, are however sensitive
- passive countermeasures, by which a third party seeks to detect and locate the transmitting station,
- to active countermeasures, by which a third party seeks to scramble the communication by transmitting an interfering signal consisting of several pure frequencies (line comb jammer),
- as well as intrinsic interference, due to the fact that the frequency hopping laws of two transmitting stations on the network can, on a given level, be superimposed.

The characteristics of these spread spectrum transmission methods are listed in an article in the technical journal.
THOMSON-CSF, volume 14, no. 4, December 1982, pages 81S to 896.

 The transmission method according to the invention also spreads the spectrum of the useful signal, but makes it possible to avoid the drawbacks of coding methods by rapid phase hopping due to the complexity of synchronization, and also makes it possible to obtain times. acquisition of the same order as those obtained by a frequency evasion transmission method, but with information rates which can be reduced at will to increase protection against interference, while retaining great insensitivity to the spectrum of an interfering signal. The method according to the invention applies more particularly in a satellite communications network comprising small transceiver stations, for tactical links at low bit rates, 75 to 2400 bitslsecond for example.

 According to the invention, a spread spectrum transmission method, in a communications network comprising several transceiver stations, is characterized in that the communications are established there by modulation, by information, of carrier signals whose frequency varies according to successive continuous functions of ranks i, fi (t), each of these functions being defined for a period T, the resulting carrier signals also being continuous signals, the series of functions fi being a pseudo-random series of functions chosen from a predetermined set of possible functions, this sequence being entirely determined, for a given communication, by a key specific to the requested station and by a time reference.

 The invention also relates to a transceiver station intended for the implementation of the transmission method thus defined.

 The invention will be better understood and other characteristics will appear with the aid of the description which follows with reference to the appended figures.

Figure 1 is a diagram illustrating the transmission method according to the invention;
- Figure 2 is a block diagram of the network transceiver station;
- Figure 3 is a diagram illustrating a lack of synchronization between transmission and reception;
FIG. 4 is the block diagram of a first embodiment of the synchronization circuit, on reception;
- Figure 5 is the block diagram of the reception part of a transceiver station, with a second embodiment of the synchronization circuit
FIGS. 6 and 7 are diagrams illustrating the synchronization method implemented in this second embodiment of the synchronization circuit,
According to the invention, the frequency of the information-carrying signal is continuously variable. FIG. 9 illustrates the frequency of the information-carrying signal as a function of time for a communication in the network between a calling station and a called station. In the method illustrated in the figure, the frequency varies according to successive continuous functions fi (t ), each defined for a period T of rank i, the period
T being preferably equal to the rhythm of the symbols of the modulation.

bearer of information. The frequency variation law resulting from the succession of these functions is itself continuous, the value reached by the preceding function corresponding to a period of rank i - 1 being the value taken for the initialization of the following function f. of the following period. These functions can be linear or not. They are chosen according to a key specific to the called station, and a time reference given by the network clock. For a given subscriber number, the sequence of functions fi is predetermined, but this sequence is pseudo-random, that is to say that the functions are chosen pseudo-randomly from a predetermined set of possible functions. The different pseudo-random sequences of functions f. different subscribers are known to other stations on the network. A station generates its own law of variation in reception and in transmission the law of frequency variation of the station with which it wishes to communicate, according to local time.

FIG. 2 represents a block diagram of a network transceiver station intended for the implementation of the transmission process according to the invention. The transmitting part includes a pilot oscillator 10 which supplies a pilot frequency applied to the frequency generator 11. The frequency generator also receives a time reference from a clock 12, and a key from a dice generator 13. From of local time supplied by the clock 12 and of the supplied by the key generator 13 the frequency generator chooses the function fi to be used at all times. The carrier signal from this frequency generator is applied to the input of a modulation circuit, which can for example be an angular modulator 14, for modulation of this carrier signal by the information signal to be transmitted. The modulation method used can be one of the following methods: MSKs QPSK, BPSK, DPSK,
FSK, ........ The output of the modulator 14 is connected to the input of transmission circuits 15, which carry out the transposition at the transmission frequency and the amplification and transmit the signal to be transmitted to an antenna d 16. Typically, in this transmission method, the spectrum of the angularly modulated variable frequency carrier signal occupies a much wider band than that of the angular modulation signal.

 The reception part includes a pilot oscillator 100, connected to a frequency generator 110 which also receives the clock signal from a clock 120 and a key from a key generator 130. The output of the frequency generator 110 is connected to the input of a mixer 140 which also receives the output signal from the reception circuits 150 connected to the reception antenna 160. The output of the mixer 140 is connected to the input of a bandpass filter 170 whose output is connected to the input of an angular demodulator 180. The reception part further comprises a synchronization circuit 200 whose input is connected to the output of the bandpass filter 170, and which optionally provides a signal error signal applied to the frequency generator 110, and a phase error signal applied to the control input of the clock 120, as described below.

 Several conditions are necessary for the implementation of the transmission method according to the invention by means of such transceiver stations.

 First of all, the frequency generators must be such that the phase of the information-carrying signal is controlled and continuous at all times. In addition, the frequency variation or phase control of these generators must be easy. Frequency synthesizers with fractional division, wobble at high speed, are particularly suitable for such use. Such synthesizers are described in French patent applications nbs 8115808 of August 17, 1981, 83 20844 and 83 20845 of December 27, 1983 in the name of the applicant. In a preferred but not exclusive embodiment, the successive continuous functions fi (T ) are linear, the frequency variation laws then being made up of straight line segments of slopes varying pseudo-randomly from one period T to another, according to a law defined by the key provided by the key generator and a function of the subscriber number requested. As indicated above, the rate of change of slope is fixed and equal to the rate of the symbols of the information-carrying modulation, so as to facilitate synchronization, but this arrangement is not necessary.

A link is established in the network between transceiver stations of the type described with reference to FIG. 2 as follows: the transceiver station transmits on the frequency law of the subscriber requested, using the key provided by the key generator linked to subscriber number and local time. The subscriber station reception circuits always generate the frequency variation law specific to the subscriber, from the same frequency given by the pilot oscillator 100, and from the same key provided by the key generator 130, own key to the subscriber. The signal received by the reception antenna 160 is mixed in the mixer 140, after possible transpositions and amplifications in the reception circuits 150, with the variable frequency signal supplied by the frequency generator 110. The locally generated signal and the signal received are affected by the same pseudo-random frequency variations. Let SE be the received signal and SL the locally generated signal. SE and SL can be expressed as follows according to their respective pulsations, 1) l and W2 of the phase variation t (t) and the angular modulation tut):
- 5E = Cos (lt + t (t) ± o (t))
- 5L = Cos (cu2t + (t))
The filter 170 tuned to the frequency Cu 1- 02 delivers a signal devoid of the pseudo-random frequency modulation expressed in the above relationships by the corresponding phase variation "(t), but affected by the angular modulation 0 (t ) constituting the information. The angular demodulation circuit 180 therefore makes it possible to restore the received information signal. The influence of an interfering signal on the behavior of this reception circuit is a beat between the wave generated locally in the receiver and the interfering signal, which does not include pseudo-random frequency modulation, and therefore produces a signal whose spectrum width is determined by pseudo-random modulation, S (t), which cannot be eliminated in the mixer Consequently, the energy of the interfering signal entering the bandpass filter 170 is reduced in the ratio of the bands of the pseudo-random modulation and of the angular modulation constituting the information to be transmitted, relative to the power present at the input of the receiver. An interfering signal can be as well a signal emitted by a hostile generator, as signals emitted by other transceivers of the network using the same band in an access system multiple.

 In practice, due to the transmission delays between the transmitter and the receiver, there may be a time difference between the locally generated variable frequency signal and the variable frequency signal modulated by the information received by the receiver. In addition, there are also generally frequency drifts. It is therefore necessary to provide a synchronization circuit making it possible to shift the signal generated locally in the receiver over time and to correct its frequency.

This is the object of the synchronization circuit 20Q provided for in the receiver in FIG. 2. An example of possible offset between the locally generated signal S1L, and the signal transmitted and received by the receiver, SR, is shown in FIG. 3 , on a diagram where the frequencies of these signals are represented as a function of time In the example shown, the two signals have an AF frequency offset and a time offset #t.

où A est un coefficient constant et S la puissance du signal reçu.Le circuit de transformation de Fourier rapide effectue son analyse sur une bande de fréquence B inversement proportionnelle à la fréquence du signal d'échantillonnage HE La raie à la fréquence F peut être détectée tant que cette fréquence reste dans la bande B, et tant que le rapport signalibruit n'est pas trop faible. Soit bt,,x l'erreur maximale pour que les deux conditions ci-dessus soient respectées, et soit AH l'erreur maximale initiale entre les horloges du modulateur et du démodulateur.Le circuit de calcul de la transformation de Fourier rapide 202 utilise un calculateur qui fonctionne en deux phases : dans une première phase dite phase d'acquisition, le démodulateur explore la plage + AH au pas de Atmax jusqu'à ce qu'une réponse apparaisse en sortie du calculateur du circuit de transformation de
Fourier. Soit Flla fréquence de la raie détectée en sortie du calculateur.FIG. 4 represents a first embodiment of the synchronization circuit 200. It includes a sampling circuit 201 receiving the signal from the bandpass filter 170, and controlled by a sampling clock signal HE. The samples from this sampling circuit 201 are applied to the input of a fast Fourier transformation circuit, 202 controlled by a clock signal, advantageously chosen at the rate of the symbol clock H5, of period T. Either at the instant when the frequency of the received signal F (SR) changes the slope, to + T being the instant of the next slope change, or to + #t the instant of the slope change in the frequency of the signal generated locally t0 + T + At being the instant of the next slope change in the frequency of the locally generated signal. In the time interval [to + Ilt,, to + TJ, the frequency of the filtered signal from the mixer and applied to the fast Fourier transform circuit has a constant value as a function of the frequency deviation åF, and the time deviation At, of value:
F = fEW #F + R. # t, f being the beat frequency filtered by the filter 170. R is the slope of the function fitt) giving the frequency during this symbol period. Outside the interval defined above, the frequency of the signal applied to the Fourier transform circuit takes other values which depend on the slopes of the preceding and following functions, fi - 1 (t) and 9 +1 (t). The power of the line at frequency F calculated by k Fourier transform circuit over the interval to + "t, t4 + At + is:

where A is a constant coefficient and S the power of the received signal. The fast Fourier transform circuit performs its analysis on a frequency band B inversely proportional to the frequency of the sampling signal HE The line at frequency F can be detected as long as this frequency remains in band B, and as long as the signal-to-noise ratio is not too low. Let bt ,, x be the maximum error for the above two conditions to be met, and let AH be the initial maximum error between the clocks of the modulator and the demodulator. The circuit for calculating the fast Fourier transformation 202 uses a computer which operates in two phases: in a first phase known as the acquisition phase, the demodulator explores the range + AH at the step of Atmax until a response appears at the output of the computer of the transformation circuit of
Fourier. Let Fll be the frequency of the line detected at the output of the computer.

The second phase, known as the synchronization phase, then consists in detecting the following line, ie F2 the frequency of this line, and in calculating from the successive frequencies F1 and F2 the time offset At This is carried out as follows: assuming the variation of negligible AF from a time interval Ti to the time interval Ti + 1, the frequencies F1 and F2 are then given by the following expressions:
F1 F AF + RlAt
F2 = # F + R2At,
R1 and R2 being the slopes of the two successive ramps.

F1-F2
#t is then equal to: # t = R1-R2
The AF frequency error can then be calculated from one or other of the measured values F1 and F2. This circuit therefore allows the synchronization error on the symbol rate, given by
At, and the frequency error AF.

 However, in general, the frequency error can be corrected simply by introducing a time shift, that is to say a corresponding desynchronization, the corresponding variation being so slight that the performance of the demodulator is not degraded. This is due to the fact that the excursion RiT is large in front of AF, even if AF is itself large in front of the spectrum width of the useful signal.

For example, in a military satellite telecommunications system, the following values can be used:
- the modulation frequency can be equal to 100 Hz, the period of the TS symbols being equal to 10 ms;
- The slope of the successive linear frequency variation functions can be chosen in the range i07 Hz / s to 2.107 Hz / s, the frequency variation over a modulation period of 10 ms then being between 100 KHz and 200 KHz.

- F max can be equal to 1 KHz;
-H can be equal to one second.

 AF compensation introduces a synchronism error of less than 1% of Ts. The Fourier transformation is calculated on a 5 kHz band, the sampling period being 0.1 ms.

During the acquisition phase, the range + H = + is is explored at step AtmaX = 0.2 ms. At the rate of 20 ms of test per position, the acquisition phase lasts at most 200s.

 This embodiment of the synchronization circuit is not the only possible one. FIG. 5 shows the reception part of a transceiver in which another synchronization mode is implemented.

In this figure the same elements as in FIG. 2 have been designated by the same references, and the elements constituting the synchronization circuit 200 have references bearing 2 as a digit for hundreds.

The reception antenna 160 feeds the reception circuits 150 connected to the first input of a mixer 140 which receives the locally generated signal in the same manner as previously by means of the pilot oscillator 100, the clock 120, the dice generator 130 and frequency generator 110. The output of the mixer 140 is connected as previously to the input of a bandpass filter 170 whose output is connected to the angular demodulator 180. However, to allow synchronization, this embodiment locally generates no longer a single local signal as previously but three time-shifted signals, the intermediate signal being that used for demodulation. For this purpose, the output of the frequency generator 110 is connected directly to the input d a second mixer 210, at the input of the first mixer 140 via a delay line 205, and at the input of a third mixer 220 via the delay lines 205 and 2I2 in series .

The mixers 210 and 220 are connected like the mixer 140 to the output of the reception circuits read 0. The outputs of these mixers 210 and 220 are respectively connected to the inputs of two bandpass filters respectively 211 and 221, the outputs of which are connected respectively to the inputs of two frequency discriminators 212 and 222. The outputs of these discriminators are connected to the inputs of detectors respectively 213 and 223, the outputs of which are connected to the inputs "+" and "-" of a differential amplifier 230. The output of this differential amplifier is connected to the low-pass filter bun input 240 whose output provides the signal synchronization control input applied to the clock synchronization control input 120.

FIG. 6 illustrates the operation of the synchronization circuit 200 as it emerges from the description of the reception circuits above. In this figure have been represented the frequencies as a function of the time of the signal from the frequency generator 116,
F (SLXt), and signals from delay lines 205 and 213, i.e. the same signal shifted by t and 2T respectively. If the errors of frequency AF and of time At are zero, the signals from the mixers 210 and 220 have the same frequency F ,. If the local clock is shifted, early or late, the signals from mixers 210 and 140 have a frequency respectively lower or higher than those from mixer 220. These signals undergo bandpass filtering to limit the noise band , then are applied to the two frequency discriminators 212 and 222 which supply signals whose amplitude is proportional to the frequency. Differential amplifier 230 provides an error signal
D, obtained by subtracting the levels detected at the output of the discriminators 212 and 222, represented in FIG. 7 as a function of the corresponding time offset At. This signal is only significant over the time interval Et, + t3. Synchronization is acquired by searching the range -H, + AH, with a maximum step equal to 2t. Keeping the numerical values indicated above, the step would be equal to 2r = 0.2 ms.

A frequency error AF changes the error signal by adding a quantity proportional to RAF, R being the slope of the frequency ramp corresponding to the symbol period in progress As indicated above the error introduced on time is in generally weak, moreover, in this embodiment it has a zero mean value because the mean of R is also zero
In the same way that in frequency evasion two links using the same frequency lead to intrinsic interference on a frequency step, in the invention described above, when two links use the same ramp simultaneously, and starting from the same final value for the previous function, this results in so-called intrinsic interference. Likewise, an interfering signal which emits frequency ramps as a function of time at tgasard can cause a bad demodulation of a symbol. If the error rate thus induced is considered unacceptable, an error correcting code used before the modulation will reduce the error rate to an acceptable value.

 The invention is not limited to the embodiments precisely described and shown. In particular, other synchronization means can be used and in particular synchronization means external to the demodulator.

 The spread spectrum transmission method according to the invention and the transceiver stations intended for its implementation can be applied in all networks comprising a certain number of subscribers using the same frequency band at the same time. This is particularly the case for satellite telecommunications networks, radio-telephone networks and even tactical military networks. This method is particularly applicable to make the links discrete, to facilitate multiple access to the networks, or to obtain a good protection against interference. This method is particularly recommended for the links at low bit rates where the frequency errors are important compared at the doinform221OnF speed and where relatively short acquisition times are sought.

Claims (8)

 1. A method of spread spectrum transmission in a communications network comprising several transceiver stations, characterized in that communications are established there by modulation, by information, of carrier signals whose frequency varies according to continuous functions successive ranks i, fi (t), each of these functions being defined for a period T, the resulting carrier signals also being continuous signals, the sequence of functions fi being a pseudo-random sequence of functions chosen from a predetermined set of possible functions, this sequence being entirely determined for a communication given by a key specific to the requested station and by a time reference.  2. Transmission method according to claim 1, characterized in that the functions fi are linear functions of time, the  slope varies in a pseudo-random manner and whose starting value is the value reached by the previous function at the end of the period Previous T.  3. Transmission method according to one of claims 1 and 2, characterized in that for the establishment of a communication a requesting station transmits on a pseudo-random frequency law specific to the requested station, each station generating, on reception, a local signal according to its law of pseudo-random variation of natural frequency, for the demodulation of the received signals.  - And a reception part comprising a generator (110) of local signal with continuously variable frequency according to a pseudo-random sequence of successive continuous functions chosen from the same predetermined set of possible functions and entirely determined by a key specific to the station and by a reference time provided by a clock (120), a mixer (140) having an input coupled to the output of the local generator (110) and an input connected to the output of reception circuits (150), a demodulator (180) coupled to the output of the mixer (140) by a bandpass filter (170) and a synchronization circuit (200) controlling an offset of the reception clock (120) in the event of desynchronization between the local signal and the received signal.  - a transmission part comprising a generator (11) of carrier signal with contindment variable frequency according to one of several pseudo-random sequences of successive continuous functions chosen from a predetermined set of possible functions, this sequence being entirely determined by a de provided by a key generator (13) determining for each communication a key depending on the requested station dice, and by a time reference provided by a clock (123, and a modulator (14) receiving this carrier signal at variable frequency and the signal d information to be transmitted, the output of which is connected to transmission circuits (15),  4. Transceiver station for implementing a spread spectrum transmission method, characterized in that it comprises:  5. transceiver station according to claim 4, characterized in that the frequency generators (11, 110) are wobble high speed frequency synthesizers.  6. transceiver station according to one of claims 4 and fi, characterized in that the synchronization circuit (200) estimates the desynchronization At between the clock associated with the received signal and the reception clock (120) by analysis of the Fourier transform of the received signal, the synchronization circuit (200) comprising for this purpose a sampling circuit (201) whose output is connected to a transformation circuit of Fourier (202), itself connected to a synchronization calculation circuit (203) providing on an output an offset control signal to the reception clock (120).  7. transceiver station according to claim 6, characterized in that the calculation circuit (204) also provides a frequency shift control signal on a second output connected to a frequency shift control input of the frequency generator reception (110).  8. transceiver station according to one of claims 4 and 5, characterized in that the synchronization circuit estimates the desynchronization At between the clock associated with the received signal and the reception clock (120) by comparing the resulting lines of the mixture between the received signal and respectively time-shifted signals by + and -T with respect to the signal transmitted by the local reception generator (110) to the reception mixer (140), the synchronization circuit comprising for this purpose two delay lines (205, 21S) placed in series at the output of the local frequency generator (110), the reception mixer (140) having its input connected to the point common to these two delay lines, the synchronization circuit comprising further two mixers (210, 220) having inputs respectively connected to the output of the frequency generator (110) and to the output of the delay lines and of the inputs connected to the output of the reception circuits (lys0), the outputs of these mixers (210, 220) being connected via bandpass filters (211, 221), frequency discriminators (212, 222) and detectors (213, 223) to the inputs of a differential amplifier (230) whose output is coupled to the control input of the reception clock (120) by a low-pass filter (240).