Blind Detection of Burst Signals in the Gaussian Channel

Today, the design of an intelligent receiver is essential for identification and blind detection of signals in the communications that are based on the burst signal. The first and most important step in this type of receiver detection, is signal presence detection and determination of the start and end points of the bursts for the purpose of extracting information. In this paper, a new method for blind detection of the burst signal is provided. In the proposed method, the beginning and end of bursts, which are actually the edges of the signal, are obtained by wavelet transform. Dice similarity criterion, burst error criterion, silence time error criterion and total error criterion are applied to investigate the detector performance. For long bursts taken as an example, the results of the wavelet transform detector simulation in the Gaussian channel, indicate that in the signal to noise ratio of -2 dB, the dice coefficient is around 0/9768 which shows the proper performance of the proposed method compared to the existing methods in this field.