A Reduced-Complexity Blind Detector for MIMO System Using K-Means Clustering Algorithm

This paper proposes a clustering-based blind detector for multiple-input multiple-output system using space shift keying modulation. First, we convert the blind detection problem to a clustering problem while considering block fading channel. Second, we use the well-known k-means clustering algorithm to design the blind detector. Third, the proposed k-means clustering detector for a blind receiver can provide comparable performance to that of the optimal receiver with perfect channel state information under the conditions of sufficient channel coherent time and sufficient random initializations of the k-means clustering algorithm. Simulations are conducted to demonstrate the performance of the proposed detector.