Distributed incumbent estimation for cognitive wireless networks
Vibhav Kapnadak, Murat Senel, Edward J. Coyle · 2008
We investigate distributed incumbent estimation algorithms for cognitive wireless networks, focusing on the 2.4 GHz ISM band. Our goal is to estimate the mean of the SNR distribution of the transmitted signals from incumbent networks that interfere with users at the edges of our APs' coverage area. This estimate is then used to determine the distance to the incumbent device. The wireless sensors we distribute along the edge of our network measure the SNRs of incumbents and communicate their measurements to a clusterhead (CH) over a noisy channel. Both the measurement and communication tasks are thus affected by noise. Each sensor sends one bit, plusmn1, that indicates whether the measured SNR is greater/less than that sensor's threshold. The clusterhead fuses these measurements to produce a maximum-likelihood estimate (MLE) of the mean of the SNR distribution. Unlike other approaches in distributed estimation, we assume each sensor uses a different threshold. We numerically compare the performance of our approach with others that assume all sensors use the same threshold. Our comparisons show that the multi-threshold approach performs: (1) as well as a single-threshold approach in which every node uses the same optimal threshold; and (2) much better than the single-threshold approach when its threshold deviates from the typically unknown optimal value. We also derive an approximate Cramer-Rao lower bound on the variance of our estimator.