A Novel Recognition Algorithm for XQPSK Based on Phase Difference Annular Statistics and SVM
Xiaoyan Ning, Yuetao Hou, and Zengmao Chen
Harbin Engineering University, Harbin, China
Abstract—A novel recognition algorithm based on phase difference annular statistics and support vector machine (SVM) is proposed to classify BPSK, QPSK, OQPSK and SOQPSK modulation schemes. The proposed algorithm exploring the characteristics of phase difference including the first-order and second-order triangle distance, variance, slope and kurtosis, to analyze the probability density of degree spread, the degree deviation from symmetrical distribution and the normal distribution. The feature parameters are input into a SVM classifier which has the advantages of generality and robustness. Compared with existing classification algorithms, the proposed algorithm can classify these four types of modulation schemes especially for signal with low signal to noise ratio (SNR). The proposed feature extraction method has less computational complexity, more stability. This algorithm improves the reliability and accuracy of the recognition by using the ring statistics and SVM classifier.
Index Terms—modulation recognition, annular statistics, phase difference, SVM, SOQPSK
Cite: Xiaoyan Ning, Yuetao Hou, and Zengmao Chen, "A Novel Recognition Algorithm for XQPSK Based on Phase Difference Annular Statistics and SVM," International Journal of Signal Processing Systems, Vol. 7, No. 1, pp. 1-6, March 2019. doi: 10.18178/ijsps.7.1.1-6
Index Terms—modulation recognition, annular statistics, phase difference, SVM, SOQPSK
Cite: Xiaoyan Ning, Yuetao Hou, and Zengmao Chen, "A Novel Recognition Algorithm for XQPSK Based on Phase Difference Annular Statistics and SVM," International Journal of Signal Processing Systems, Vol. 7, No. 1, pp. 1-6, March 2019. doi: 10.18178/ijsps.7.1.1-6