GB2349544A

GB2349544A - Covert communication system

Info

Publication number: GB2349544A
Application number: GB9007410A
Authority: GB
Inventors: Robert G Wilkinson
Original assignee: UK Secretary of State for Defence
Current assignee: UK Secretary of State for Defence
Priority date: 1989-05-30
Filing date: 1990-04-02
Publication date: 2000-11-01
Anticipated expiration: 2010-04-02
Also published as: GB2349544B; GB9007410D0; GB8912337D0

Abstract

Data is transmitted covertly using <WC 1>a large number (4096) of narro<WC 1>wband channels spaced over a 1MHz bandwidth <WC 1>at a signal level below that of background noise. Input data to be transmitted is demultiplexed so that sections of data are transmitted in groups of channels at 25 baud,DPSK modulated using difference phase shift key (DPSK) and added for transmission as an analog signal. The inventor envisages that 16DPSK can be used because of the coherent and semi-coherent channel addition signal processing which can be achieved. In the receiver the signal is converted to digital form then processed by an FFT circuit to produce 4096 frequency bins. Differential phase detection is carried out and running averages used in each channel to determine the proportion of times that the phase difference falls within allowed limits. A discrimination level determines whether individual channels should be excised because of noise corruption. The remaining channels in groups are then added together to determine each data phase/received. The <WC 1>measured phase vectors are added vectotially using semi-coherent addition and the vector sum is used to determine the data bit. In a second arrangement the channels of any one group are arranged in bunches of about 10 spread over a bandwidth of 10 kHz (the experimentally determined coherence bandwidth) with the bunches spread over the 1MHz bandwidth. The bunch channels are added coherently and the resultants are added semi-coherently as before to determine the group data. The group data is then multiplexed to reproduce the high data rate input data. The arrangement allows very high data rates to be transmitted with each of the 4096 channels carrying only 25 baud.

Description

B Wideband High Frequency Covert Communications System The invention relates to low probability of intercept (LPI) communications for the high frequency (HF) band.

The purpose of anti-intercept communications is to deny any enemy electronic warfare (EW) interceptors the ability to detect a transmitted message. If the interceptor is entirely unable to detect these LPI signals, even using feature extraction techniques, then the signals are said to be covert. If all the transmitted signals of a task force are sent covertly the eneny EN units will be unable to exploit the communications traffic : a. locate transmitters (task force unit) ; b. jam reception; c. monitor traffic levels; d. intercept messages; or e. falsify messages The aim of an LPI waveform is to conceal the existence of a transmitted signal from a determined interceptor. Ideally this should be done in away that will always put the signal below the background noise level in the radio channel.

This ensures the signal is'covered'and therefore indistinguishable from all the other RF noise energy also being received. Interceptor detection under these conditions can only be achieved using special signal processing techniques which are only able to detect the presence of the signal (by feature extraction) rather than decoding the content of it.

Fortunately, at HFs these feature extraction techniques are made virtually ineffective because of the variability of the background noise from atmospheric, galactic and man made interference.

Present day communications waveforms can offer no protection against interception because they must be received, by the intended recipients with an excellent signal to noise ratio ( > 10 dBs) or otherwise the messages will contain numerus errors. This makes detection by an enemy EW interceptor very easy because interception can be successfully achieved at much lower signal to noise ratios ( < OdBs) than is necessary for acceptable message quality.

The aim for LPI is therefore to tip the balance of signal to noise ratio in favour of the communications requirement and against that for the interceptor. This can be achieved when the transmitted energy of a communications waveform is'spread'over a much greater frequency bandwidth because this reduces the signal power received in a given bandwidth. If this frequency spreading is made large enought the transmitted signal power will be lower than the background noise.

Unfortunately, although frequency spreading can make interception virtually impossible it also makes communications very difficult because the quality of the received signal will also be very poor for the intended recipients.

Hweverf this can be overcome if the waveform is properly designed since signal processing techniques can be used to reconstitute the frequency spread signal (de-spreading) and interference excision algorithms can remove the hundreds of very much larger interferences. Such a technique is described in GB Patent Application No (Agents ref: P0723).

A special de-spreading process is essential for the successful operation of the communications system because it must recover the very low power signals from the higher background noise which is used to conceal it from the enemy interceptor. The interceptorr however, will be unable to use the same process for interception because the frequency spreading will be done in a cnptoc, raphic r, lanner.

In the cited reference LPI waveform frequency spreading was achieved using a direct sequence spreading pseudo random code to modulate the carrier frequency at a very high chip rate. By this means the data signal and pseudo random code were combined to create the wideband transmitted signal. Unfortunately, although the transmitter was very efficient and simple to implement the reception of these signals was ineffective. This problem arose because the propagating radio path can produce considerable phase and frequency distortions on the received signal and because very much larger interfering signals are also received. This made it impossible to recover the wanted signal. Various methods for correcting the phase and frequency distortion were investigated but an efficient cost effective solution could not be found.

The obj ect of the present invention is to provide a wideband hf communications system which overcomes the known problems.

The invention provides: A wideband high frequency communications system comprising : a transmitter comprising : a) input means to receive a digital signal data for transmission; b) the input means connected to a multiplicity of separate narrowband frequency channels, each preferably less than 100 Hz, distributed over a discrete broad transmission spectral region of the hf band whereby each channel is modulated at a low data rate; and c) means to transmit the multiplicity of low data rate frequency channels ; and a receiver comprising: a) means to receive the multiplicity of low data rate frequency channels; b) means to identify channels corrupted by noise ; and c) means to recombine the channels so as to remove the effect of identified channel corruption and recover the transmitted digital data.

Throughout the specification the bandwidth of the individual channels is described as narrowband since it is much smaller than the bandwidth of the overall system, referred to as broadband. In addition, the rate of data transmission, typically 25 bps per channels is selected to be low enough such that a signal period is longer than a typical multipath impulse response (about 8ms). The plurality of channels enable the system to achieve a significant signal processing gain in order to operate the system specifically in much higher ambient background noise levels than the signals themselves.

Advantageously the channels are divided into groups and a demultiplexer is provided in the transmitter, connecting the input means to each group whereby a high data rate input signal is demultiplexed to produce a multiplicity of parallel low data rate channels in each group, and in the receiver the means to recombine the channels includes a multiplexer to reproduce the high data rate transmitted signal. By this means transmission data rates of 9. 6 kbps can be achieved while preserving multipath discrimination.

In one arrangement the channels of any one group are interleaved with the channels of every other group such that the channels of each group are spread throughout the broad transmission bandwidth.

The number of sub-channels used is optional, but in the limit it depends on the total bandwidth available and the required data rate. In general, the more sub-channels used the greater will be the data rate and/or the higher will be the signal processing gain. Thus generally the highest possible bandwidth is used to achieve the highest processing gain. This in turn will allow the system to be operated at the lowest signal to noise ratio.

In a preferred arrangement the communications system has 4096 narrowband channels spread over a lMHz bandwidth. Preferably interference-corrupted channels are excised in the receiver before the channels are recombined. By this means the de-spreading and detection mechanisms in the receiver are optimised and processing gains are produced to enable the signals to be recovered from well below the received noise floor.

Thus by using a very large number of channel frequencies to transmit each data bit it is possible to recover the transmitted message from a very noisy received signal spectrum. Those channels with unacceptably high noise or interference can be removed without producing serious degradation because each data bit is transmitted using so many channels and it only takes one good S/N channel to receive a data bit correctly.

In the preferred arrangement the data is differential phase shift key (DPSK) modulated in each channel. The transmitter channels are then preferably summed and converted back into an analogue signal before transmission.

Advantageously differential phase shift key (DPSK) modulation is employed and this has been shown to be operable fro M-ary number 2 to at least 16 (where M is the modulation level used). In the preferred arrangement the difference phases relevant to an element of data, in all channels in a group, are added vectorially and the resultant vector is used to determine the data element.

Thus the combined signal is assumed to be the same as the wanted signal. On the face of it this would appear to be an incorrect assumption since it is derived from all channels and the preponderance of these may be noisy.

However, computer simulations have shown that this arrangement works and the combined output from all the channels is a good representation of the signal.

The phase of this signal can therefore be used to measure more accurately the phase PDF in each channel. This is done by counting the number of times the channel signal phase falls within defined limits of theconbined signal phase (hits). Channels which fail to achieve a desired hit count are then excised from the conbined signal and this in turn improves the accuracy of the combined signal phase. This results in a rapidly converging excision system. Preferably therefore there is included in each channel an error detector and an exciser, the resultant phase vectors for each signal being connected to the error detector where they are compare to the corresponding measured phases in the channel, an output signal being produced therein corresponding to the proportion of channel phases which lie within predetermined phase limits of the resultant phase vestors, the output signal being connected to the exciser where the channel is accepted or excised for data decoding in dependence on whether the measured proportion exceeds a further predetermined limit.

Advantageously the error detector output is also used to control a channel exciser at the input to the phase vector adder whereby only good channel signals determine the resultant phase vectors. In this arrangement the output from the error detector may be connected to a decorrelator which compares the proportion of channels for excision to a predetermined limit and prevents channel excision exceeding the limit.

The receiver preferably includes an analogue to digital converter connected to a fast Fourier transform (FFT) processor which has a number of frequency bins equal to the number of transmitted frequency channels. The interference excision means preferably includes means to measure the signal phase in each frequency bin and to determine the quality of each received frequency channel, and to excise channels where interference is subsequently detected. Each channel is preferably measured for quality by comparing its phase with all its expected values and a count is made of the number of times it looks like the expected phase (a hit). If this count exceeds a fixed threshold level then the channel is used. If the channel count is less than the threshold then the channel output is excised.

In a direct sequence system. excision or frequency weighting can be accomplished by comparing signals in symmetrically placed frequency bins, one frequency with a signal exceeding the other by more than a predetermined ratio threshold can be either excised or weighted to match the other. In an active arrangement the threshold value could be varied according to the quality of the received data.

The invention will now be described by way of example only with reference to the accompanying Drawings of which: Figure 1 illustrates a known FEK HF communications system ; Figure 2 illustrates a known PSK HF communications system; Figures 3a and 3b illustrate a known spread frequency communication system and the resulting broadband frequency spectrum ; Figures 4a and 4b show a block diagram of a wideband transmitter used in the present invention and the resulting broadband frequency spectrum and Figure 4c is a block diagram of a system for randomising a DPSK signal channel ; Figure 5a illustrates 2 and 4 phase shift key (PSK) modulation processes and Figures 5b and 5c contrast PSK and DPSK waveforms; Figure 6 is a schematic block diagram of the waveform generator; Figure 7 is a schematic block diagram of the receiver; Figure 8 shows diagrammatically the semi-coherent channel addition; Figure 9a illustrates graphically channel vector addition of phases; Figure 9b is a graph of the phase probability density function (PDF) for random noise outputs from a DPSK detector; Figure 9c is a similar graph to Figure 9b for differing signal to noise ratios; Figure 9d illustrates how the Figure 9c PDF can be used for interference excision; Figure 9e is a block schematic diagram illustrating an arrangement using phase PDF for interference excision; Figure 10 shows the effect of channel addition on the received signal error rate (BER) against the channel signal to noise ratio (SNR); Figure 11 shows the effect of M-ary number on BER against SNR ; Figure 12 is a block diagram of of interference excision and data decision making in a bunched and grouped frequency transmission system ; Figure 13 shows a block diagram of an alternative phase detector and channel exciser; and Figure 14 is a vector diagram illustrating operation of the Figure 13 arrangement.

As shown in Figure 1 a known 3k Hz bandwidth frequency exchange key (FEK) communications system operates (11) on a stream 12 of digital data to produce one of two frequency tones 13. 14 depending upon whether a data"1"or"0"is present. Two receiver channels 15, 16 are included in the remote receiver, tuned to the respective"1"and"0"frequency tones. Because this system includes some frequency diversity (2 frequencies) some protection is provided against interference and multipath distortion. Good signal-to- noise is required to enable the receiver to detect the signal and this makes interception of the communications link relatively easy, hence making the link susceptible to jamning.

Figure 2 illustrates a phase shift key (PSK) modulator 21 operating on an input data flow 22 of 25 bits per sec to produce an output signal which shows a single peak 23 in the frequency spectrum. This is because the phase of the transmitted signal is either = 0O or 1800 depending upon whether a data"1"or "0"is to be transmitted. Unlike the FEK transmission there is no frequency diversity, making the transmission highly susceptible to multipath interference and thus PSK is not used in this application.

Figure 3a illustrates a conventional previous art broadband communications system employing frequency spreading. Data 31 is modulated (32) by a pseudo random code produced by a pseudo random data generator 33 r the pseudo random data being selected by a clock input 34. The modulated data is then connected to a RF modulator 35 where it is RF modulated by a carrier signal 36 of frequency Fc. The RF modulator output is then connected to a radio transmitter 37. Figure 3b shows that for a pseudo random chip rate of 500 kbps a 1mi2 broadband frequency spectrum 38 is produced, centred on the carrier frequency Fc. By spreading the communications waveform over a wide range of frequencies the transmitted signal power can be lower than the background noise. Interference excision must then be used to attempt to remove the large number of narrowband interference signals which will be received by a 1 MHz bandwidth receiver. As mentioned previously, signal reception by this technique is inefficient because of distortions of the received signal caused by variations in the propagating medium and because of the large signal interferences received.

A wideband communications system according to the present invention uses a new waveform generated by a transmitter as shown in Figures 4a and 4b. In the example shown, 4096 frequency channels or tones 41 are generated by an inverse fast Fourier transform (FFT) circuit 42 for connection to a channel bit phase modulator 43. Input data 44 is connected to a demultiplexer 45 having a plurality of outputs 46 connected to selected groups of the 4096 frequency channels41, aswill bedescribed below, formodulationin the phase modulator 43. Each tone within a group is modulated by the same sequence of data bits, with different data bit sequences being used to modulate the other respective groups. The output channels 47 from the phase modulator 43 are combined in an RF modulator 48 to modulate a carrier signal of frequency Fc for transmission by a radio transmitter T x 49. The 4096 channels are arranged such that the frequency spectrum 410 of the transmitter is very flat and is spread over a 1 MHz bandwidth centered at Fc. Each channel 411 is narrowband (ie less than lOOHz) and, as shown, this can form a constant low level signal across the spectrum. The data demultiplexer 45 is arranged such that the channel data rate is say 25 bps. Hence for an identical input (44) data rate of 25 bps each of the 4096 channels can be modulated by the input data without demultiplexing. Thus there would be 4096-fold redundancy. At higher input data rates the data is demultiplexed to provide a number of outputs 46 with a low baud rate of 25 bps at each output. These outputs 46 are connected to respective groups of channels in the phase modulator 43 with each channel in a group being modulated in like manner. The groups may be selected such that the frequencies of any one group are interleaved with the frequencies of each other group and thus spread over the lMHz bandwidth. Thenr f or exampler for transmission of input data at 100 bits per sec, the data would be demultiplexed into four groups of 1024 channels and then there would be 1024- fold redundancy for each data to it with each channel transmitting at 25 transmission bps.

The flat frequency spectrum 410 can be made to look exactly like gaussian white noise (thermal) by using special modulating techniques as shown in Figure 4c. Each channel DPSK signal can be randomised to look exactly like gaussian white noise. The data channel 412 from the data source 413 after DPSK modulation (414) is connected sequentially to phase (415), frequency (416) and amplitude (417) modulators, each connected to outputs from a pseudo random number generator (PNG) 418. Random numbers produced by the pseudo random generator 418 synchronised to the time of day (419) would be used such that Phase, Frequency and Amplitude modulate the channel signal. The PNG clock is used to control the data rate from the data source 413.

At the receiver the channel signal would be unwhitened using an inverse/reciprocal process before being phase demodulated.

A further refinement might be to also time j itter the transmitted baud symbols (variable baud rate). This would prevent the signal being detected by an interceptor using a symbol rate analyser. Because there are very many narrow-band channels being transmitted for each data bit, large numbers can be removed if corrupted by interference, while still maintaining the communications link.

At the receiver each narrowband channel is separately received, filtered and detected before being re-combined (de-spread) to reproduce the multiplexed higher speed data output. Interference in any of the channels can be readily identified so the corrupted channels can be removed before the signal recombining. This optimises the de-spreading and detection mechanisms.

In a conventional frequency diversity communications system each channel is sampled and compare and then an algorithm is used to select the best channel identified as not being corrupted by interference. Such an approach effectively throws away any advantage that could be achieved by using channel signal coherent summation.

The inventor has discovered that by using simple DPSK modulation with the phase shifts of 7t, or a more complex DPSK modulation with phase shifts of 1 (/2 or 5t/4 etc. then an effective semi-coherent channel addition can be carried out in the receiver.

Then for the same channel redundancy the transmission rates can be increased by the following relationship for each transmission group.

PSK (2 phase)-25 bps QPSK (4 phase)-50 bps 8PSK (8 phase)-75 bps 16 PSK (16 phase)-100 bps Thus by using DPSK modulation, higher data rates are possible for the same channel bandwidth and the same channel redundancy.

Figure 5a illustrates the 2 phase and 4 phase PSK modulation where the phase of the carrier is altered from upwards (0) (51) to downwards (t) (52) as shown depending on whether a digital"1"or"0"is to be transmitted. In the QPSR arrangement the carrier is modulated by one of the phases 53-56. Thus on detection, the phase windows for discriminating between"1"and"0"are as follows : PSK- (/2 QPSK- tut/4 8 PSK- (/8 16 PSK- 1 (/16 The inventor has found that differential phase coding leads to an improved performance through channel signal combining since it relies on relative phases and not on absolute phase. When the high frequency radio path causes phase changes detection is then possible without a prior knowledge of absolute phase. Figure 5b shows PSR modulation of the data stream 011001010 and Figure 5c shows differential phase modulation. In the phase difference modulation a change of phase of 1800 represents a"1"and zero phase change represents a"0". Figure 6 illustrates the transmission waveform generation using difference PSK (DPSK) modulation. As shown, the input data 61 is demultiplexed (62) to a number of groups N with each group having 4096/N channels. Each channel is DPSR modulated (63), and modulated by time window (64) before being added (65) and then converted to analogue form (66) for transmission. On reception (Figure 7) the received analogue signal is first converted to digital form (71) then connected to a Fast Fourier Transform circuit 72 with 4096 frequency bins or channels 73. As will be described below, interference excision is then carried out in a circuit 74 where channels carrying identified interference are discounted (as indicated by the intelligent switch 75). The parallel channel outputs 73 after interference excision are connected to a channel addition circuit 76 which adds channels in the respective groups 1 to N. Channel addition is done using a semi-coherent addition process as will also be described below. The groups are then decoded by a DPSK decoder 77 and the grouped output data is then multiplexed (78) to provide the output data.

Figure 8 illustrates the semi-coherent addition process. Differential Phase detection 81 is done in each baud period (40ms) in each of the 4096 frequency channels 82. The frequency channels of each group which should all be identicalt are then vectorially added (83) and a decision algorithm (84) is employed to determine the transmitted data bit. Figure 9a illustrates the vector addition of three such differential phase detector outputs 91-93 with the dashed line 94 representing the sum. Ideally for perfect signal-tonoise all these vectors should be in line but noise and interference varies the detected phase. In a DPSK modulation system where the phase difference should be either 0 or 1 as indicated in Figure 5a, the decision algorithm (84) can be arranged such that the data bit is resolved as"0"if the vector sum 94 falls within the range of angles 0 + (/2, and is"1"if the vector sum is within-+ Tk/2. Where a higher order PSK modulation is used, QPSK for example, the decision window ti} (/2 above) is made appropriately smaller about the expected phase differences.

Where 2-phase DPSK is employed we expect the DPSK detector output to be 0 or 1800 at the baud rate of 25. Thus the DPSK output is sampled 25 times a second. If interference is present this will not be at exactly the same frequency or expected phase as any one channel and will cause the measured signal to phase rotate in that channel. If there were just random noise phase shifts at the DSPK output this would lead to a uniform phase probability density distribution (PDF) 95 (Figure 9b) at the DSPK detector output (81).

For the 2-phase DSPK system we can define the detection windows for a data"0" as 00 + 900 and"1"as 1800 900 because the phase PDF distribution is centred on 00 and 1800 although the detected phase can be deviated away from the true position by random noise as shown in Figure 9c. For S/N = # the received phase will be either Oo or 1800 (96, 97). For lower S/N ratios the phase will increasingly (for falling S/N) have a probability of being further away from 0 and 180 . A good S/N PDF is shown by 98, 99 and a poor S/N PDF by 910, 911.

For detection therefore a signal phase which lies within 900 of the phase positions Oo and 1800will be detected as 00 and 1800 respectively (i. e."0"or "1"). Figure 9c indicates a method for directly measuring signal quantity and its potential for use in an interference excision algorithm without any prior knowledge of the data being transmitted. Figure 9d shows the PDF of phase for a DPSK signal of modest S/N ratio. It can be seen that the phase has a higher probability of being within a small range of (either) 00 or 1800. If the received signal was just random noise/interference this would not be true. as indicated by Figure 9b. For noise or interference discrimination, therefore, we reduce the phase windows (914, 915), about the expected phase directions by say half, to + 45 0. A running average is then taken, for each channel over a number of baud periods, of how many times the detected phase falls within the expected windows 914., 915 (Hit) and the number of times it falls outside (Miss) the expected windows as indicated in Figure 9d. If the ratio of Hits to Misses is much greater than 1 then the channel may be considered good Forlower S/N ratios the Phase PDF ratio will be much lower.

For random noise, for example, it will be an average, If because there is an equal probability of being a Hit or a Miss.

Figure 9e shows how the Phase PDF detector 916 is connected via a filter 917 to a receive channel 918 to operate as an interference/noise exciser system. A Hit/Miss count in the exciser 919, as described with reference to Figure 9d, would be made at the detector output 920 and if it exceeds a specified S/N threshold the signal in the receive channel 918 will be connected to a subsequent vector summing circuit 921, asdescribed above, by the switch 922.

The setting of the S/N threshold in the exciser 919 is critical to the proper optimum operation of the system because this decides what the quality of the channel signal shall be before it is added to the vector summing circuit 921.

If, for example, the number of channels being added is very larger say 1000, then it is necessary to include all these channels having a S/N of > -lOdBs because when added together they can produce a resulting equivalent signal of > +7dBs (see Figure 10). When the number of channels is much smaller. say 10, then only channels of > +2dBs can be used to produce a resulting signal of > +7dBs (1 in 1000 BER).

The time taken (the number of bauds) to measure the Hit to Miss ratio will depend on the acceptable S/N ratios for the signals and the required reliability of excision process. In general the poorer the signal-to-noise ratio, then the longer it will take to measure the Phase PDF ratio.

For small numbers of channel vector combining the S/N ratio has to be good so the number of bauds needed for calculating the Phase PDF can be quite small.

Typically this may be done over a 20 baud period (0.8s). If we count 20 Hits (very low noise probability) then the switch 922 in the excision circuit (Figure 9e) is closed while if the number is 10 (random noise) then the switch 922 is opened. A threshold for channel acceptance is therefore set between 10 and 20 hits in this example. Using thebinomial theory, the probability of 20 hits being due just to noise is 1 in 106 so the probability of a false signal is very small. The probability of exceeding a specified threshold number of 20 hits being due just to noiseisl inlO6 sotheprobability of a false signal is very small. The probability of exceeding a specified threshold number of hits due to noise can be reduced by counting over a larger number of baud samples.

The inventor has shown that even though there may be poor demodulation signal in the individual channels (with a poor estimate of phase change), when the vector sum is done followed by detection there is a significant processing gain.

The effect of channel addition can be seen with reference to Figure 10. This shows graphs 101-104 of the required channel signal to noise (dBs) against bit error rate (BER) for the addition of If 10, 100 and 1000 channels respectively for DPSK (M-ary = 2). Surprisingly this shows a significant improvement in using more than one channel despite the fact that DPSK demodulation is used prior to channel addition (See Figure 8).

The table below shows the waveform performance in terms of the number of channel/groups (degree of multiplexing) and the M-ary number.

Table

NO OF GROUPS (G) 1 2 4 8 16 32 64 128 256 CHANNELS/GROUP 4096 2048 1024 512 256 128 64 32 16 DATA RATE (BPS) 25 50 100 200 400 800 1600 3200 6400 (M-ary=2) DATA RATE 50 100 200 400 800 1600 3200 6400 12800 (M-ary--) DATA RATE 75 150 300 600 1200 2400 4800 9600 19200 (M-ary--8) DATA RATE 100 200 400 800 1600 3200 6400 12800 25600 (M-ary =16) DATA RATE=25. G. LOG (M-ary) b/s This shows that with an M-ary number of 16 and N = 256, 16 channels per group, a data rate of 25. 6 kbps can be achieved. The highertheM-ary number, however, the greater the channel signal-to-noise ratio that is needed for the same BER performance. Figure 11 shows graphs of BER against SNR for M-ary numbers 2 r 4 r 8 and 16. Thus, as can be seen, for a bit error rate of 10-2 (1%) the SNR is approximately 6,9,14 and 21 dBs respectively for M-ary numbers 2, 4,8 and 16.

In a conventional PSK system employing 16 PSK and one channel then for 10-4 BER we would need +25 dB SNR if there were no phase distortion or interference.

Hence 16 PSK is normally never used for HF communications. With our system and using 128 channels the SNR needed is approximately 1-2 dBs for the same BER.

Experiment has shown that there are a correlation bandwidths of greater than lOkHz in the HF spectrum. This is the extent of the frequency band over which the transmission dispersion characteristics are correlated and the phases of the received signals are correlated. To take advantage of this property the transmission of the 4096 channels may be arranged such that same of the channels of any one group are bunched in frequency bands of about lOkHz across the frequency spectrum. True correlation techniques can then be applied to each of the channels in a bunch. Thus for example, each bunch might comprise 100 channels. True coherent processing would be used for each bunch of 100 channels (20 dBs of processing gain) in a group and then semi-coherent processing will be applied to the bunches (as in Figure 8) to r -combine all the frequency channels in the group. This approach will give much greater processing gain and allow even lower SNR reception and therefore will lead to improved covert communicatons.

Figure12 shows this process more clearly and shows how interference excision would be implemented. The input channels 121 in the lOkHz bandwidth are separately excised (122) before being coherently summed in 123, to form the 'bunch channel'sum signal. If more channel bunches are used to form the groups then these are semi-coherently added together in 125 after the signals from 123 are differentially phase demodulated in 124. The vector sum output from 125 is decoded in the phase detector 126 to produce the data output 127, for the group, which would then be multiplexed together with other groups as in Figure 7, to form the user data signal. Since the signals in the lOkHz 'bunch'channels have the same phase characteristics (being in a correlated bandwidth) the signal at the output of the coherent summer 123 must represent the same signal but with a greatly enhanced S/N. In gaussian noise this improvement is 10 log (N) where N is the number of channels in the'bunch'.

The measurement of the signal phase 128 in the phase detector 124 is therefore an accurate measure of the actual phase of each channel signal. This can therefore be used to increase the accuracy of the excision modules (122) phase PDF Hit/Miss ratio counter (129) because we now know what the phase of each channel signal was. This increased accuracy will reduce the probabil ity of a false phase PDF count and also reduce the time taken tomake a channel excision decision.

The performance of a 1000 channel group system (say) based on Figure 12 would yield a much greater processing gain than the previously described semicoherent system (Figures 7 and 8). Figure 10 shows. for example. that the semi-coherent system would give a gain of about 18dBs for 10-3 BER.

For a system like Figure 12 where there could be 100 channels/bunch and 10 bunches/group the gain would be about 26dBs (20 + 6dBs).

An alternative arrangement for the channel excision circuit shown in Figure 9e has been found to improve the phase PDF detector and the channel excision process in general. Figure 13 shows the improved arrangement for one of many channels (say 100). The received channel input 131 is differentially detected by a phase detector 132 to find the change in relative phase of the signal since the last baud. The data output 133 from the circuit is determined using a data group detector 134 compose of a data channel exciser 135r a vector summer 136 and an MDPSK decoder 137.

In this arrangement the MDPSK decoder 137 converts the phase of the vector summer 136 output to a binary code (gray scale) as is known in the art. Each channel input to the vector sommer 136 is excised if the channel quality falls below a predefined threshold. In this particular channel the exciser 135 will be activated if the count of a BPSK error detector 138 exceeds a set threshold 139 in a given period of time (or baud count). The BPSK error detector 138 thus replaces the phase PDF detector in the previous arrangements. The purpose of the BPSK error detector 138 is to compare the phase of the channel signal 1310 with that produced by an estimated group vector detector 1311 composed of a channel exciser 1312 and a vector summer 1313. The vector summer 1313 produces the estimated signal vector 1314 from the combined inputs of the unexcised channels in the group (as is also done in the data group detector 134). The estimated signal vector 1314 therefore represents the sum of all the good channels in the group and will therefore have a better signal to noise ratio (S/N) than any one individual channel in the group.

In the BPSK error detector 138 the phase of the channel signal 1310 is compared to the phase of the estimated group vector 1314 as shown in Figure 14. The BPSK error detector 138 will count an error whenever the phase of the channel signal 141 exceeds the estimated signal vector 142 phase by more than Pi/2.

The signal 143 is considered correct if it lies within Pi/2 of the estimated signal vector 142.

This alternative system is better than the previous hit/miss arrangement because a direct comparison is now made between each measured channel phase and what it should actually have been. In the previous arrangement a test was made (Figure 9d) to see if the signal phase fitted the expected phase (PDF) for the given modulation level. In the Figure 13 arrangement the BPSK error count is always made as shown in Figure 14, irrespective of the modulation level used to transmit the data (eg 8DPSK) whereas in the previous arrangement the phase PDF hit and miss windows have to be changed to suit the modulation level being used.

A refinement to the channel excision process is also necessary (in the estimated group vector detector 1311) to prevent erroneous channel capture.

This is done by using a channel decorrelator 1315. The decorrelator 1315 is used to prevent a small number of channels"capturing"the excision process and locking out all the other channels. The decorrelator 1315 operates to stop the number of unexcised channels falling below a def ined ratio 1316 r f or example 50%. The decorrelator 1315 unexcises some of the poor channels to the summer 1313 to prevent capture. Adding these poor channels back to the excision process produces a negligible effect on the data output bit error rate.

Claims

Claims 1. A wideband high frequency communications systemhaving : a transmitter comprising : a) input means to receive a digital signal data for transmission ; b) the input means connected to a multiplicity of separate narrowband frequency channels, where narrowband is defined hereinbefore distributed over a discrete broad transmission spectral region of the hf band whereby each channel is modulated at a low data rate; and c) means to transmit the multiplicity of low data rate frequency channels; and a receiver comprising : a) means to receive the multiplicity of low data rate frequency channels ; b) means to identify channels corrupted by noise ; and c) means to recombine the channels so as to remove the effect of identified channel corruption and recover the transmitted digital data.
2. A wideband high frequency communications system as claimed in claim 1 wherein each channel is less than 100 Hertz.
3. Awideband high frequency camunications system as claimed in claim 1 or 2 wherein the channels are divided into groups and a demultiplexer is provided in the transmitter, connecting the input means to each group whereby a high data rate input signal is demultiplexed to produce a multiplicity of parallel low data rate channels in each group, and in the receiver the means to recombine the channels includes a multiplexer to reproduce the high data rate transmitted signal.
4. A wideband high frequency ccammications system as claimed in any one preceding claim wherein the channels of any one group are interleaved with the channels of every other group such that the channels of each group are spread throughout the broad transmission bandwidth.
5. A wideband high frequency communications system as claimed in claim 4 wherein the channels within a group are distributed in bunches with the channels within a bunch being adjacent in frequency. the receiver then being arranged to process the bunch channels by a semi-coherent addition process.
6. A wideband high frequency communications systemas claimed in any one preceding claim wherein the communications system has 4096 narrowband channels spread over a lMHz bandwidth.
7. A wideband high frequency communications system as claimed in any one preceding claim wherein differential phase shift key (DPSK) modulation is used.
8. A wideband high frequency communications system as claimed in any one preceding claim wherein interference-corrupted channels are excised in the receiver before the channels are recombined.
9. A wideband high frequency communications system as claimed in claim 8 wherein the phase probability density function (PDF) is used to determine whether individual channels are corrupted by noise/interference.
10. A wideband high frequency communications system as claimed in claim 9 wherein each receiver channel includes a phase window defined around the values expected from the receiver DPSK output and there is also included a counting circuit is included to count the proportion of values that fall within the window (Hits to Misses) and means to excise statistically determined noisy channels.
11. A wideband high frequency communications system as claimed in any one of claims 7 to 10 wherein digital data is DPSK modulated and the transmitter DPSK channels are summed and converted back into an analogue signal before transmission.
12. A wideband high frequency communications system as claimed in any one of claims 7 to 11 wherein DPSK modulation is employed with an M-ary number in the range of 2 to 16.

13. A wideband high frequency communications system as claimed in any one of claims 3 to 12 wherein difference phases relevant to an element of data, in all channels in a group, are added vectorially in a phase vector adder and the resultant vector is used to determine the data element.

14. A wideband high frequency communications system as claimed in claim 13 wherein there is included in each channel an error detector and an exciser, the resultant phase vectors for each signal being connected to the error detector where they are compared to the corresponding measured phases in the channel, an output signal being produced therein corresponding to the proportion of channel phases which lie within predetermined phase limits of the resultant phase vestors the output signal being connected to the exciser where the channel is accepted or excised for data decoding in dependence on whether the measured proportion exceeds a further predetermined limit.

15. A wideband high frequency communications system as claimed in claim 14 wherein the error detector output is also used to control a channel exciser at the input to the phase vector adder whereby only good channel signals determine the resultant phase vectors.

16. A wideband high frequency communications system as claimed in claim 15 wherein the output from the error detector is connected to a decorrelator which compares the proportion of channels for excision to a predetermined limit and prevents channel excision exceeding the limit.

17. A wideband high frequency communications system receiver as claimed in any one preceding claim wherein the receiver includes an analogue to digital converter connected to a fast Fourier transform (FFT) processor which has a number of frequency bins equal to the number of transmitted frequency channels.

18. A wideband high frequency communications system wherein direct sequence data modulation is used and as claimed in any one of claims 8 to 17r the interference excision means includes means to compare signals in symmetrically placed frequency bins, one frequency with a signal exceeding the other by more than a predetermined ratio threshold can be either excised or weighted to match the other.

19. A wideband high frequency communications system as claimed inclaim 18 wherein the threshold value is varied according to the quality of the received data.

20. A wideband high frequency communications system as claimed in any one of claims 7 to 19 wherein each DPSK channel is randomised prior to transmission.

21. A wideband high frequency communications system as claimed in claim 20 wherein the randomising is done by connecting the DPSK signal sequentially to phases frequency and amplitude mLdulators, each modulator being connected to the digital output of a pseado-random number generator (PNG).

22. A wideband high frequency communications system as claimed in claim 21 wherein the PNG is synchronised to the time of day.

23. A wideband high frequency communications system as claimed in any one of claims 20 to 22 wherein the receiver includes inverse means to demodulate the randomised signals.

Amendments to the claims have been filed as follows 1. A covert wideband high frequency communications system having : a transmitter comprising: a) input means to receive a digital data signal for transmission; b) a multiplicity of separate narrowband frequency channels distributed over a discrete broad transmission spectral region of the high frequency band; c) means to modulate each narrowband frequency channel with the digital data using phase shift key (PSK) modulation; and d) means to combine the multiplicity of frequency channels as a broadband signal for transmission; and a broadband receiver comprising: a) means to receive the broadband signal and separate the narrowband frequency channels; b) means to identify channels corrupted with noise; and c) means to recombine the channels so as to remove the effect of identified channel corruption and recover the transmitted digital data using a semi-coherent processor to determine the phase of the PSK signal from the measured phases in each signal channel; the arrangement being such as to enable the receiver to recover the transmitted signal when the signal level in each receiver channel is no greater than the order of the background noise level.

2. A wideband high frequency communications system as claimed in claim 1 wherein the modulation means uses differential phase shift modulation.

3. A wideband high frequency communications system as claimed in claim 1 or 2 wherein the transmitter includes a digital to analogue converter to produce a broadband analogue signal for transmission and the receiver includes an analogue to digital converter to digitise the received signal.

4. A wideband high frequency communications system as claimed in any one of claims 1 to 3 wherein each channel is less than 100 Hertz.

5. A wideband high frequency communications system as claimed in any one preceding claim wherein the semi-coherent processor comprises means for the vector addition of the measured phases.

6. A wideband high frequency communications system as claimed in claim 5 wherein the semi-coherent processor includes threshold means to determine the transmitted phase from the vector addition.

7. A wideband high frequency communications system as claimed in any one preceding claim wherein the channels are divided into groups and a demultiplexer is provided in the transmitter. connecting the input means to each group whereby a relatively high data rate input signal is demultiplexed to produce a multiplicity of parallel relatively low data rate channels in each group, and in the receiver the means to recombine the channels includes a multiplexer to reproduce the high data rate transmitted signal; the difference phases relevant to an element of data, in all channels in a group being added vectorially in a phase vector adder and the resultant vector used to determine the data element.

8. A wideband high frequency communications system as claimed in claim 7 wherein the channels of any one group are interleaved with the channels of every other group such that the channels of each group are spread throughout the broad transmission bandwidth.

9. A wideband high frequency communications system as claimed in claim 8 wherein the channels within a group are distributed in bunches with the channels within a bunch being adjacent in frequency and within a frequency band of less than or equal to the order of 10 kHz, the receiver then being arranged to process the bunch channels by a coherent process; the measured phase outputs from the bunches within a group then being added by the semi-coherent processor to determine the phase of the transmitted signal.

10. A wideband high frequency communications system as claimed in any one preceding claim wherein interference-corrupted channels are excised in the receiver before the channels are recombined.

11. A wideband high frequency communications system as claimed in claim 10 wherein the phase probability density function (PDF) is used to determine whether individual channels are corrupted by noise/interference 12. A wideband high frequency communications system as claimed in claim 11 wherein each receiver channel includes a phase window defined around the values expected from the receiver DPSK output and there is also included a counting circuit to count the proportion of values that fall within the window (Hits to Misses) and means to excise statistically determined noisy channels.
13. A wideband high frequency communications system as claimed in any one of preceding claim wherein DPSK modulation is employed with an M-ary number in the range of 2 to 16.
14. A wideband high frequency communications system as claimed in claim 13 wherein there is included in each channel an error detector and an exciser, the resultant phase vectors for each signal being connected to the error detector where they are compared to the corresponding measured phases in the channel, an output signal being produced therein corresponding to the proportion of channel phases which lie within predetermined phase limits of the resultant phase vestors, the output signal being connected to the exciser where the channel is accepted or excised for data decoding in dependence on whether the measured proportion exceeds a further predetermined limit.
15. A wideband high frequency communications system as claimed in claim 14 wherein the error detector output is also used to control a channel exciser at the input to the phase vector adder whereby only good channel signals determine the resultant phase vectors.
16. A wideband high frequency communications system as claimed in claim 15 wherein the output from the error detector is connected to a decorrelator which compares the proportion of channels for excision to a predetermined limit and prevents channel excision exceeding the limit.
17. A wideband high frequency communications system receiver as claimed in any one preceding claim wherein the receiver includes an analogue to digital converter connected to a fast Fourier transform (FFT) processor which has a number of frequency bins equal to the number of transmitted frequency channels.
18. A wideband high frequency communications system wherein direct sequence data modulation is used and as claimed in claim 10 or L-, the interference excision means includes means to compare signals in symmetrically placed frequency bins, one frequency with a signal exceeding the other by more than a predetermined ratio threshold can be either excised or weighted to match the other.
19. A wideband high frequency communications system as claimed in claim 18 wherein the threshold value is varied according to the quality of the received data.
20. A wideband high frequency communications system as claimed in any y one of claims 7 to 19 wherein each DPSK channel is randomised prior to transmission.
21. A wideband high frequency communications system as claimed in claim 20 wherein the randomising is done by connecting the DPSK signal sequentially to phase, frequency and amplitude modulators, each modulator being connected to the digital output of a pseado-random number generator (PNG).
22. A wideband high frequency communications system as claimed in claim 21 wherein the PNG is synchronised to Cite time of day.
23. A wideband high frequency communications system as claimed in any one of claims 20 to 22 wherein the receiver includes inverse means to demodulate the randomised signals.
24. A wideband high frequency communications systemas claimed in any one preceding claim wherein the communications system has 4096 narrowband channels spread over a lMHz bandwidth.