Bispectrum Classification of Multi-User Chirp Modulation Signals Using Artificial Intelligent Techniques

摘要

Automatic Digital signal type classification (ADSTC) has many important applications in both of the civilian and military domains. Most of the proposed classifiers can only recognize a few types of digital signals. This paper presents a novel technique that deals with the classification of multi-user chirp modulation signals. In this paper, the peak of the bispectrum and its bi-frequencies are proposed as the effective features and different types of classifiers are used. Simulation results show that the proposed technique is able to classify the different types of chirp signals in additive white Gaussian noise (AWGN) channels with high accuracy and the neural network classifier (NN) outperforms other classifiers, namely, maximum likelihood classifier (ML), the k-nearest neighbor classifier (KNN) and the support vector machine classifiers (SVMs).