Method for the automatic detection and classification of digital quadrature-amplitude-modulated signals having unknown parameters

摘要

This method is intended for monitoring the radio traffic, for example by the post office. The received signal is digitised and, with prior knowledge of the frequency bandwidth in one or otherwise in a number of parallel channels of different frequency bandwidth, is separately processed further. In each channel, the signal element timing is extracted and the instantaneous values of amplitude, frequency and phase are sampled at the signal element timing points. Mean values and variances of the amplitude and frequency values are determined. Using the amplitude values, the frequency values and the differences of successive phase values, histograms are formed. The histogram forms typical of the wanted signal types are automatically detected by cross correlation with predetermined reference function sets which are typical of the modulation. To increase the reliability of decision, the mean values and variances between amplitude and frequency are simultaneously checked for the maintenance of predetermined value ranges. The classification of the signal type is obtained from the reference function set which leads to the highest correlation results in the histogram analysis. The signal element rate is estimated from the frequency bandwidth of the associated channel.