Automatic Modulation Recognition in Wireless Multi-carrier Wireless Systems with Cepstral Features

• Published in: Wireless personal communications Vol. 81; no. 3; pp. 1243 - 1288
• Main Authors: Keshk, Mohamed El-Hady M., El-Naby, Mohammed Abd, Al-Makhlasawy, Rasha M., El-Khobby, Heba A., Hamouda, W., Abd Elnaby, Mustafa M., El-Rabaie, El-Sayed M., Dessouky, Moawad I., Alshebeili, Saleh A., El-Samie, Fathi E. Abd
• Format: Journal Article
• Language: English
• Published: Boston Springer US 01.04.2015
• Subjects: Algorithms; Automation; Communications Engineering; Computer Communication Networks; Engineering; Learning theory; Modulation; Networks; Neural networks; Recognition; Signal,Image and Speech Processing; Support vector machines; Transforms; DT; SDR; AWGN; OFDM; ADMR; SVM; MFCC; RPROANN
• ISSN: 0929-6212; 1572-834X
• DOI: 10.1007/s11277-014-2183-3

Automatic digital modulation recognition (ADMR) has become an interesting problem in wireless communication systems with various civil and military applications. In this paper, an ADMR algorithm is proposed for both orthogonal frequency division multiplexing and multi-carrier code division multiple access systems using discrete transforms and mel-frequency cepstral coefficients (MFCCs). The proposed algorithm uses one of the discrete cosine transform, discrete sine transform, and discrete wavelet transform with MFCCs to extract the modulated signal coefficients, and uses also either a support vector machine (SVM) or an artificial neural network (ANN) for modulation classification. Simulation results show that the proposed algorithm provides higher recognition rates than those obtained in previous studies, in addition to a superiority of SVM performance compared to ANN performance at low signal-to-noise ratios.