Decision fusion schemes for wireless sensor networks operating in a Nakagami-m fading environment

In this paper, we study the fusion of decisions in a wireless sensor network in which a number of independent geographically distributed local sensors transmit hard decisions to a fusion center. We assume that the transmission channel is unreliable due to the presence of channel fading and noise. In particular, for a Nakagami-m fading environment, we consider two optimum fusion rules based on the likelihood ratio; one that relies on instantaneous channel state information and another that uses only the channel fading statistics. Since the optimum likelihood ratio based fusion rules are quite complex and their performances difficult to evaluate in practice, we also consider some suboptimum fusion rules which are very easy to implement and whose performances are not substantially worse than those of the optimum fusion rules. Numerical results presented show that the Nakagami-m fading parameter has a significant effect on the detection performances of these well known fusion rules.