Enhanced SVD-QR-T FCM for Adaptive Channel Selection in Virtual MIMO based WSNs

This paper proposes a practical adaptive version of the algorithm SVD-QR-T, which was proposed in [11]. We call this new proposed algorithm ‘Adaptive SVD-QR-T FCM’, in which the fuzzy c-means (FCM) adaptively adjusts the number of clusters it uses, compared with the SVD-QR-T in which only two clusters are employed. The proposed algorithm selects a subset of channels in virtual multiple-input–multiple-output (MIMO) wireless sensor networks (WSNs), to balance the MIMO advantage and complexity of sensor cooperation. In the proposed model, WSN is organized in a manner of cluster to cluster multihop, the singular-value decomposition-QR with threshold (SVD-QR-T) algorithm selects the best subset of transmitters while keeping all receivers active. The threshold is updated adaptively by means of Fuzzy C-Means (FCM). Moreover, and more important than updating the threshold, For betterpresentation of data in clusters, we apply statistical method called Elbow, in which the number of clusters in FCM is determined adaptively. Our proposed algorithm differs from the previous algorithm SVD-QR-T FCM, in terms of number of clusters used in FCM. To validate and compare the performance of this algorithm with previous work, Extensive Monte Carlo simulations are presented and demonstrated that despite of no difference between these two algorithms in terms of capacity, BER and Multiplexing Gain (MG) at low number of transmitters, the Adaptive SVD-QR-T FCM reveals significant improvements in the capacity with a slight degradation in BER at high number of transmitters.