Improving the Classification of MPSK and MQAM Modulations by using Optimized Nonlinear Preprocess in Flat Fading Channels

kground and Objectives: Intelligent receivers, automatically detect the digital modulation type of the received signals for demodulation purposes where is well known as Automatic Modulation Classification (AMC) module. The performance of AMC algorithms depends on the channel conditions where for example, in fading channel its performance gets worse than the AWGN channel.Methods: We propose a new algorithm for improving the AMC classification accuracy in flat fading channels. The proposed algorithm consists of an optimizable nonlinear preprocess followed by Linear Discriminant Analysis (LDA) technique. Two Lemmas have been found for extracting the optimization rule. And an optimization algorithm has been built based on the previous Lemmas. Results: The simulation results show that the proposed algorithm improves the classification accuracy between ۸-Phase Shift Keying (۸PSK) and ۱۶PSK (as an example of M-array PSK (MPSK) inter-class) for Signal-to-noise ratio (SNR) values greater than ۱۳ dB, and between ۱۶- quadrature amplitude shift modulation (۱۶QAM) and ۶۴QAM (as an example of M-array QAM (MQAM) inter-class) for SNR values greater than ۴ dB.Conclusion: By using the proposed optimization algorithm, the AMC classification accuracy has been improved. Other classification problems can use this algorithm. And other nonlinear preprocess functions or optimization algorithms may be found in future work.

Keywords:

Automatic modulation classification , Linear discriminant analysis , Higher-Order Cumulants , Mahalanobis Distance