GB2350527A

GB2350527A - Modulation identification

Info

Publication number: GB2350527A
Application number: GB9906688A
Authority: GB
Inventors: Christopher Mark Dearlove; Paul George Wright
Original assignee: Marconi Electronic Systems Ltd
Current assignee: BAE Systems Electronics Ltd
Priority date: 1999-03-23
Filing date: 1999-03-23
Publication date: 2000-11-29
Also published as: GB9906688D0

Abstract

Within the HF band, a wide range of different signal types are employed, for example different types of modulation are implemented and different users employ different transmission standards. Therefore, it is necessary to be able to recognise the presence and type of signals in order to be able to receive and demodulate incoming data. The invention provides a device for identifying a received waveform, comprising a first recognition means 1 arranged to correlate a first property of the waveform with that associated with a first type of modulation, and to calculate a first correlation value indicative of the likelihood that the waveform has the first type of modulation. The device further comprises a second recognition means 2 arranged to correlate a second property of the waveform with that associated with a second type of modulation and to calculate a second correlation value accordingly. Decision means 3 are also provided and arranged to compare the first and second correlation values in order to decide which of the types of modulation has been employed.

Description

1 2350527 1 P/61772.GBA/MRC WAVEFORM MENTWICATIO This invention relates to

the identification of the waveform of a received signal.

Within the HF band, for example, a wide range of different signal types are employed, often by the same users. For instance, different types of modulation are implemented, and different users employ different transmission standards. 7lierefore, it is necessary to be able to recognise the presence and type of signals in order to be able to receive and demodulate incoming data. The ability to identify waveforms is a necessary part of an automated frequency management system, in particular for users of the HF band.

7he invention provides a device for identifying a received waveform, comprising a first recognition means arranged to correlate a first property of the waveform with that associated with a first type of modulation, and to calculate a first correlation value indicative of the likelihood that the waveform has the first type of modulation.

is Advantageously, the device further comprises a second recognition means arranged to correlate a second property of the waveform with that associated with a second type of modulation, and to calculate a second correlation value indicative of the likelihood that the waveform has the second type of modulation.

Preferably, decision means are provided and arranged to compare the first and second correlation values and to decide which of the first and second types of modulation is 2 P/61772.GM/MRC the more likely for the received signal.

The invention is described below, by way of example with reference to the accompanying drawing (Figure 1), which is a block diagram of a device constructed according to the invention.

The invention is described using the example of phase shift keying (PSK) signal modulation and frequency shift keying (M) signal modulation. Of course, the invention can be used to identify waveforms having other types of modulation, for example quadrature amplitude modulation (QAM).

Referring to the drawing, the waveform identification device comprises PSK recognition means 1, FSK recognition means 2 and decision means 3. The incoming signal is fed into both PSK recognition means 1 and FSK recognition means 2 for is simultaneous processing.

It is known that PSK signals include known data sequences, in dependence on the transmission standard employed. Different users employ different known standards.

In the PSK recognition means 1, some or all of the known part of a PSK waveform is divided into a number of subsequences. Each of these subsequences is correlated with the received signal. The correlations are combined non-coherently and normalised.

Non-coherent combination is employed because it is faster than coherent combination, which would in any case be subject to residual frequency errors, so that minimal 3 P/61772.GBA/MRC advantage would be gained by using coherent combination. This process works best when carried out over an integral multiple of cycles of the known sequences, but other lengths of time may be employed. An appropriate settling period of time may also be incorporated. The combined and normalised correlation value, the maximum for any position of the known sequences, is in the range 0- 1. A suitable threshold is set, for example 0.4, to balance false recognition and recognition failure probabilities. A correlation value above this threshold is indicative of the likelihood that the received waveform is a PSK type.

Of course, the invention may be employed just to identify waveforms of one type of interest, for example PSK modulated signals, as described above. However, the received signal can be analysed simultaneously M another manner, in order to ascertain whether the signal may be of another type, for example FSK modulated. Thus, the signal is also processed at the FSK recognition means 2.

is The FSK recognition means makes use of the fact that the spectrum of an FSK modulated signal has characteristic peaks at the frequencies corresponding to the known FSK signalling frequencies. There may be some offset from the known signalling frequencies if a frequency error is present. Noise and PSK modulated signals do not tend to have significant spectral peaks, so the presence, position and number of such peaks is a means of identification of FSK modulated signals. The FSK recognition means 2 calculates the spectrum of a burst of the received signal, and calculates the sum of the amplitudes of the spectral components at the expected signalling 4 P161772.GBA/MRC frequencies. Similar sums of amplitudes are also calculated over a range of frequency offsets from the expected signalling frequencies, and the maximum sum is taken in order to allow for possible frequency errors. The maximum sum is then normalised so that, for the expected FSK waveform, a value of unity is obtained. This quantity is the correlation of the signal against the FSK reference, and it can be shown that for other types of waveform and noise, it is significantly less than one. Therefore, threshold value is set in the range 0-1, and if the correlation exceeds the threshold, then it is likely that the received signal is an FSK type.

The respective correlation values from the PSK recognition means 1 and the FSK recognition means 2 are then input to the decision means 3.

The decision means 3 combines the individual correlation values for each possible waveform, and then makes a decision on the identity of the waveform. The decision algorithm can be summarised as follows:- If no correlations exceed their corresponding threshold, then none of the possible waveforms are present.

If exactly one correlation value exceeds its decision threshold, then that is the waveform present. If more than one correlation value exceeds its corresponding threshold, then for each P/61772.GBA/MRC one that does, the proportion that the correlation is over the threshold is calculated, given by the following formula:

(correlation, - threshold)1(1.0 - thresholc.) 5 ]be waveforTn with the biggest value of this quantity is chosen as the likely identity of the waveform.

Mie invention has been described in relation to the HF band, as frequency management for such a band is needed. The invention is applicable to other frequency bands and to other types of modulation. Provided that a modulation imposes a characteristic on,a signal, a waveform transmitted according to this modulation can be identified by the above invention.

is While the components of the device have been described in terms of hardware, the invention is also applicable to the case where they are implemented by software.

Claims

6 P/61772.GBA/MRC CLAIMS

1. A device for identifying a received waveform, comprising a first recognition means arranged to correlate a first property of the waveform with that associated with a first type of modulation, and to calculate a first correlation value indicative of the likelihood that the waveform has the first type of modulation.

2. A device as claimed in claim 1, further comprising a second recognition means arranged to correlate a second property of the waveform with that associated with a second t)W of modulation, and to calculate a second correlation value indicative of the likelihood that the waveform has the second type of modulation.

3. A device as claimed in claim 2, further comprising decision means arranged to compare the first and second correlation values and to decide which of the first and second types of modulation is the more likely for the received signal.

4. A device as claimed in any preceding claim, wherein one of the first and second types of modulation is PSK and the respective property is a data sequence.

5. A device as claimed in claim 4, wherein the other of the first and second types of modulation is FSK and the respective property is a frequency spectrum.

6. A device as claimed in claim 4 or 5, wherein the respective type of modulation 7 P/61772.GBA/MRC is associated with at least one predetermined standard.

7. A device, substantially as hereinbefore described with reference to, or as illustrated in, the accompanying drawing.

8. A method of identifying a received waveform, comprising correlating a first property of the waveform with that associated with a first type of modulation and calculating a first correlation value indicative of the likelihood that the waveform has the first type of modulation and correlating a second property of the waveform with that associated with a second type of modulation and calculating a second correlation value indicative of the likelihood that the waveform has the second type of modulation and comparing the first and second correlation values to decide which of the first and second types of modulation is the more likely for the received signal.