Signal Identification using a new high efficient technique

Abstract

Abstract: Automatic signal type identification (ASTI) is an important topic for both the <br><br>civilian and military domains. Most of the proposed identifiers can only recognize a few <br><br>types of digital signal and usually need high levels of SNRs. This paper presents a new high <br><br>efficient technique that includes a variety of digital signal types. In this technique, a <br><br>combination of higher order moments and higher order cumulants (up to eighth) are <br><br>proposed as the effective features. A hierarchical support vector machine based structure is <br><br>proposed as the classifier. In order to improve the performance of identifier, a genetic <br><br>algorithm is used for parameters selection of the classifier. Simulation results show that the <br><br>proposed technique is able to identify the different types of digital signal (e.g. QAM128, <br><br>ASK8, and V29) with high accuracy even at low SNRs.