This paper presents a verity of K-Nearest Neighbors (KNN) classifiers along with different statistical features in order to classify MPSK and MQAM signals. Further, the performance of proposed KNN classifiers with different values of âKâ and distance functions are analyzed under non-ideal channel conditions. Finally, to prove the superiority of the proposed KNN classifiers the performance is compared with that of the literature approaches at various SNR values.
Modulation Recognition, Distance Function, Nearest Neighbors, Majority Selection