An algorithm to mitigate channel distortion in blind modulation classification

This paper presents a method for classification of digital modulation schemes in an asynchronous reception scenario. The proposed classification is based on likelihood principles and without prior knowledge of channel response. We propose blind channel estimation in non-cooperative scenario using conventional Decision Directed Least Mean Square algorithm while the signal is subjected to channel distortion and unknown gain. Convergence characteristics of the LMS algorithm is modulation dependent and after choosing the best equalizer, the final decision for Modulation Classification is made by likelihood principles. Experimental results are presented to compare the performance with the optimal classifier. The scope for further improvement is outlined.