WO1997015986A1

WO1997015986A1 - A method for direct sequence spreading of a data sequence (dsss)

Info

Publication number: WO1997015986A1
Application number: PCT/SE1996/001371
Authority: WO
Inventors: Arne Lindblad; Anna Wik
Original assignee: Försvarets Forskningsanstalt
Priority date: 1995-10-25
Filing date: 1996-10-25
Publication date: 1997-05-01
Also published as: JP2001513950A; AU712151B2; SE9503752L; SE506622C2; EP0873598A1; CA2235737A1; AU7511496A; SE9503752D0

Abstract

The present invention relates to a method for direct sequence spreading of a data sequence (d(t)) or a modulated data sequence (s(t)). The method yields better spectral properties than conventional methods for data transmission using direct sequence spread spectrum (DSSS) and is also suitable for systems where one wants to conceal the existence of transmission, so-called stealth radio. This is achieved by one or more spreading sequence generators (21) in the transmitter and the corresponding sequence generators (31) in the receiver emitting new output data at a rate which essentially exceeds the one required for obtaining in the transmitter a certain desired spread bandwidth, the spreading sequence or spreading sequences (c(t)) being caused to bandspread the data sequence (d(t)) or the modulated data sequence (s(t)), and the desired spread bandwidth being obtained by filtering in band-limiting filters (22).

Description

A METHOD TOR DIRECT SEQUENCE SPREADING OF A DATA SEQUENCE (DSSS)

The present invention relates to a method for direct sequence spreading of a data sequence. The method yields better spectral properties than conventional methods for data transmission using direct sequence spread spectrum (DSSS) and is also suitable for systems where one wants to conceal the existence of transmission, so- called stealth radio. Better spectral properties mean in this case that the power out¬ side the desired spread bandwidth is reduced and that the power distribution within the spread bandwidth can be controlled. This is good in frequency-divided multiple user systems since different signals will interfere with each other to a minimum extent.

Direct sequence spread spectrum is a well-known technique. The method was in¬ vented in the first place to obtain a good protection against jamming. Besides, it decreases the possibilities of an unauthorised receiving or even perceiving that the signal is emitted, which is a common reason for using DSSS. This property is usually called stealth radio or LPI, Low Probability of Intercept. The suggested method adds two important properties for stealth radio signalling. On the one hand, it reduces the risk of detection in most common radio intelligence detectors owing to the shaφ filtering of the signal and, on the other hand, a spreading code is pro¬ duced which is more difficult for an undesired receiver to reproduce. A further reason for using DSSS can be the possibility of letting many different users utilise the same frequency band without interfering with each other. Each user is then given his unique code sequence such that only the intended receiver is able to receive the message. This technique is called Code Division Multiple Access

(CDMA) and attracts much interest for e.g. cellular radio. A further reason for using DSSS is the possibility of improving the resistance to frequency-selective fading caused by multipath propagation.

DSSS works on the principle that a spreading sequence (also called spreading code) is multiplied by the more slowly varying data sequence and is allowed to modulate a carrier wave, usually by applying the method BPSK (Binary Phase Shift Keying). Alternatively, one lets the data sequence first modulate a carrier wave, which is then multiplied by the spreading code. Also QPSK (Quaternary Phase Shift Keying), or other forms of PSK (Phase Shift Keying) are used. The spreading code, the spreading rate and the carrier-wave frequency are known to the receiver, who retrieves the transmitted data sequence by demodulating and de-spreading the signal. The method DSSS is described in more detail in, for instance, Marvin K. Simon et al, "Spread Spectrum Communications", Computer Science Press Inc., 1985, which is hereby incoφorated by reference.

The spreading codes in DSSS systems are often produced by binary feed-back shift registers. The codes are of the type binary pseudorandom sequences (PN sequences) and are available in various designs. The best properties will be found in the so-called maximum length sequences. The disadvantage therewith is that for a given length of sequence there is a limited number of different sequences. To be able to choose between a greater number of sequences, one generally uses com¬ binations of two or more maximum length sequences, so-called gold sequences. These sequences, however, will have inferior correlation properties.

To increase the amount of sequences having good correlation properties, chaos- generated sequences have recently been suggested for communication systems, see Heidari-Bateni and McGillem, "A Chaotic DSSS Communication System", IEEE Transaction on Communication, Vol. 42, February/March/April 1994. The generated chaos sequences are non-binary, i.e. they can assume many different values within an interval. An advantage of chaos-generated sequences is that it is possible to produce a very large number of different sequences having a low crosscorrelation, which is good in CDMA applications and for stealth radio. A further advantage is that a chaos sequence is very easy to generate. In many cases, it is only necessary to save a constant and the preceding value in order to generate the next value, see H.G. Schuster, "Deterministic Chaos, An Introduction", Physick-Veriag GmbH, 1984, which is hereby incoφorated by reference. In the above-mentioned article by Heidari-Bateni and McGillem, the values of the chaos sequence have been allowed to amplitude-modulate a carrier wave and the resulting signal has been found to be better from the LPI viewpoint than traditional DSSS.

Pulse forming of the data pulses and the spreading code has been suggested in, for instance, Mark A. Wickert et al, "Practical Limitations in Limiting the Rate-Line Detectability of Spread Spectrum LPI Signals", Proceedings of MILCOM, Monterey, 1990, in order to decrease the detecting probability of the DSSS signal in LPI con¬ texts. It has been established that it is, above all, the frequency components over half the spreading sequence rate that make the signal easy to discover by means of a squaring detector or a delay-and-multiply detector, which are the most probable detectors for radio intelligence against DSSS signals. One therefore tries to filter off these higher frequency components.

The present invention provides a different method for improving the spectral proper- ties for different applications in the fields of stealth radio (LPI) and frequency shar¬ ing in multiple user systems, but is not restricted to these applications and is instead intended to comprise all fields where there are similar problems. The features of the invention are stated in the claims.

The invention will now be described in more detail with reference to the accompany¬ ing drawings, in which:

Fig. 1 illustrates schematically a known principle for BPSK DSSS, Fig. 2 illustrates schematically an embodiment of a transmitter according to the invention, and

Fig. 3 illustrates schematically an embodiment of a receiver according to the invention.

The basic idea of the invention is that the spreading sequence is generated at a rate that exceeds the need to a considerable extent, in order to obtain a certain desired spread bandwidth, and that the sequence or the modulated signal is then filtered to the desired spread bandwidth.

According to the present invention, very powerful filtering is carried out for the pur- pose of having the spreading symbols, after filtering, assume values between the sequence values generated by the spreading code generator and create a depend¬ ence between many successive sequence values. This is something quite different from the moderate filtering of the emitted signal that is frequent in prior-art DSSS systems and which is carried out to improve the spectral properties. Such moderate filtering only leads to the emitted spreading pulses being slightly rounded without being too much deformed.

Fig. 1 shows schematically an example of a system using prior-art technique for direct sequence spread spectrum. The data source 12 outputs binary data d(t) to a modulator 13, which by BPSK modulates a carrier wave. The modulated signal s(t) is multiplied by a spreading sequence in a multiplier 14. The spreading sequence consisting of the symbols +1 is generated by a PN generator 11 at a rate which roughly seen yields the spread bandwidth W_ss for the system. The spread spec- trum signal x(t) is transmitted via an arbitrary channel to the receiver, where the received signal y(t) is multiplied in a multiplier 16 by a sequence, generated in a PN generator 19, which is identical with the spreading code. The effect of the spreading code on the signal then disappears since c²(t) = 1. The de-spread signal r(t) is de¬ modulated in a demodulator 17 and regenerated data d(t), which is an estimate of the original data, is feed to the data sink 18.

If the known DSSS system is carried out in analog technique, the BPSK modulator 13 and the multipliers 14 and 15 may consist of diode ring mixers. To the modulator 13 there is supplied, in addition to data, the desired carrier wave generated in an oscillator. The demodulator 17 of the receiver is also supplied with a sine signal which is locked to the carrier wave in y(t). Besides, the demodulator may also con¬ sist of a diode ring mixer, an integrator and a decision circuit. The PN generators 11 and 19 can be designed with digital circuits as feed-back shift registers. The feed¬ back pattern and the start value are the same for the PN generators of the trans- mitter and the receiver.

Alternatively, the signal processing of the system in the units 11, 13, 14, 16, 17 and 19 is earned out in digital technology, using, for instance, a digital signal processor (DSP) or an application-specific integrated circuit (ASIC). By such an implementa- tion, also digital/analog conversion of x(t) and analog/digital conversion of y(t) will be required. An example of a circuit containing most of the functions in a receiver as described above is PA-100 "Spread Spectrum Demodulator ASIC" supplied by Loral Coφoration. It is important that the signal x(t) is amplified and possibly filtered before being sup¬ plied to the channel 15. The receiver amplifies y(t) and separates it by frequency selective filtering before it is de-spread in 16.

A transmitter according to the invention can be designed fundamentally as illus¬ trated in Fig. 2. Data to be transmitted from the data source 23 modulates a carrier wave in a modulator 24. The modulator can be designed for various usual modula¬ tion forms, for instance BPSK, QPSK or MSK (Minimum Shift Keying). The phase of the signal s(t) will be changed in the phase rotator 25 controlled by the signal c(t). The phase shifting is in the range +π. The thus modified signal x(t) is transmitted via some medium to the receiver. The bandwidth of the signal c(t) is considerably greater than s(t), which results in the desired bandspread. The spreading signal c(t) is generated by the sequence generator 21 coacting with the filter 22. The se¬ quence generator 21 may, however not necessarily, be designed as a chaos gen- erator. The sequence generator emits a new output value at a rate which is k times greater than the rate corresponding to the desired spread bandwidth W_ss. The filter 22 low pass filters the sequence and forms the signal c(t), which has a bandwidth corresponding to the spread bandwidth W_ss. Independently of whether the se¬ quence generator emits binary or multilevel sequences, the signal c(t) will have many possible levels after the filtering.

If the transmitter is designed as described above and the modulation method in 24 is selected to be one having a constant amplitude, for instance BPSK or MSK, the modulated and bandspread signal x(t) to be transmitted will also have a constant amplitude. This is a property which is most desirable in contexts where one wants to use power efficient non-linear amplifiers. If the ratio of the bandwidth of the signal s(t) to that of the signal c(t) is high (high processing gain), the bandwidth of x(t) will largely be determined by the transfer function in the filter 22. This design is very convenient in connection with cellular radio.

The receiver can be designed fundamentally as shown in Fig. 3. The sequence generator 31 and the filter 32 are identical with the corresponding units 21 and 22 of the transmitter. The sequence generator 31 has also the same code key as the sequence generator 21 of the transmitter. The sequence generated in the se- quence generator 31 is synchronised to y(t) which is the received signal x(t) delayed and transmitted in the transmission medium. The generated replica of the signal c(t) is sign-inverted in the inverter 33 before it is allowed to control the phase rotator 36. Alternatively, the inverter 33 is included in the phase rotator 36. The bandspread signal y(t) will be de-spread in the phase rotator 36, and the signal r(t) contains merely the original modulated carrier wave. In the demodulator 35, data is repro¬ duced by prior-art technique for the selected modulation method and is fed to the data sink 34.

The technical implementation of the suggested methods can largely be designed with digital signal processing in digital signal processors or ASIC. In the same way as in the description of the prior-art system, the following will be required: digital/analog conversion of x(t), analog/digital conversion of y(t) and the condition that the signal x(t) is amplified and possibly filtered before being fed to the channel. The receiver amplifies and separates by frequency selective filtering y(t) before it is de-spread in the phase rotator (36).

The chaos generators 21 and 31 can, for instance in a DSP, be designed by using the so-called logistic function

*-+! = **»(! - *.).

see the above-mentioned book by H.G. Schuster. The calculation accuracy in the used DSP decides how long the sequence can be before repeating itself. The filters 22 and 32, which decide the spread bandwidth W_ss, can be digital low pass filters of the type FIR (Finite Impulse Response) or MR (Infinite Impulse Response). It is important that the chaos generators and filters of the transmitter and the receiver be designed in exactly the same manner and with exactly the same numerical accuracy.

The phase rotation in 25 and 36, respectively, can be carried out mathematically as a complex multiplication of the l-Q divided signal s(t) and y(t), respectively. Alterna¬ tively, the carrier-wave signal which is fed to the modulator 24 and the demodulator 35, respectively, can be phase-shifted. The required sine signal can be generated in a circuit for Direct Digital Synthesis (DDS) and there its phase is shifted, con¬ trolled by c(t). An example of a circuit having the desired functions complex multipli¬ cation and phase-controllable signal generation is HSP45116 "Numerically Con¬ trolled Oscillator/Modulator supplied by Harris Semiconductor.

An altemative embodiment of the transmitter is to generate two independent spreading signals c(t) and c'(t), respectively, which are allowed to modulate the real part and imaginary part of s(t) (l-Q modulation) in a complex multiplier. This can be carried out by doubling the sequence generator 21 and the filter 22, thereby gen- erating two signals c(t) and c'(t), respectively. The phase rotator 25 is replaced by a complex multiplier, in which the complex signal s(t) is multiplied by c(t) as real value and c'(t) as imaginary value. In this embodiment, the signal x(t) will not have a con¬ stant amplitude, but the possibilities of forming the spectrum of x(t) will be improved. W_ss is determined unambiguously by the transfer function of the filters 22. This embodiment, especially using chaos-type sequence generators, is particularly suitable for LPI systems.

Claims

Claims:

1. A method for direct sequence spreading of a data sequence (d(t)) or a modu¬ lated data sequence (s(t)), having one or more spreading sequence generators (21) in the transmitter and the corresponding sequence generators (31) in the receiver, c h a r a c t e r i s e d in that the sequence generators emit new output data at a rate essentially exceeding the one required for obtaining in the transmitter a certain desired spread bandwidth, that the spreading sequence or spreading se¬ quences (c(t)) are caused to bandspread the data sequence (d(t)) or the modulated data sequence (s(t)), and that the desired spread bandwidth is obtained by filtering in band-limiting filters (22).

2. The method as claimed in claim 1, c h a r a c t e r i s e d in that the spreading sequence or spreading sequences (c(t)) are chaos sequences generated by a chaos generator (21,31).

3. The method as claimed in claim 1 or 2, c h a r a c t e r i s e d by using only one sequence generator (21) in the transmitter, whose generated sequence, after filtering in said band-limiting filters (22), is caused to modulate the data sequence carrier-wave phase, whereupon, in the receiver, the corresponding sequence-controlled phase rotation takes place with the opposite sign.

4. The method as claimed in claim 1 or 2, c h a r a c t e r i s e d by using only one sequence generator (21) in the transmitter, whose generated sequence is caused to modulate the data sequence carrier-wave phase, whereupon the modu¬ lated data sequence (s(t)) is filtered in said band-limiting filters (22), whereupon, in the receiver, the corresponding sequence-controlled phase rotation takes place with the opposite sign.

5. The method as claimed in claim 1 or 2, c h a r a c t e r i s e d by using two sequence generators in the transmitter, whose generated sequences, after filtering in said band-limiting filters, are caused to modulate the data sequence camer-wave real part and imaginary part, respectively (I and Q modulation), where- upon, in the receiver, the corresponding sequence-controlled phase rotation takes place with the opposite sign.

6. The method as claimed in claim 1 or 2, c h a r a c t e r i s e d by using two sequence generators in the transmitter, whose generated sequences are caused to modulate the data sequence carrier-wave real part and imaginary part, respectively (I and Q modulation), whereupon the modulated data sequence (s(t)) is filtered in said band-limiting filters, whereupon, in the receiver, the corresponding sequence-controlled phase rotation take place with the opposite sign.