In this paper, we present Singular-Value Decomposition-QR with Threshold (SVD-QR-T) algorithm to select a subset of channels in virtual MIMO wireless sensor networks (WSN) in order to reduce its complexity and cost. SVD-QR-T selects best subset of transmitters while keeping all receivers active. The threshold is adaptive by means of Fuzzy C-Mean (FCM). Under the constraint of the same total transmission power, this approach is compared against the case without channel selection in terms of capacity, bit error rate (BER) and multiplexing gain in the presence of water-filling as well without. It is shown that in spite of less multiplexing gain, when water-filling is applied, SVD-QR-T FCM provides lower BER at moderate to high SNR; in case of equal transmission power allocation, SVD-QR-T FCM achieves higher capacity at low SNR and lower BER. In general, it provides satisfying performances compared to the case without channel selection but reduced cost and resource.
展开▼
