When there is no feedback channels in wireless sensor networks or when the sensors are deaf (cannot listen to the feedback channels), the conventional and practical method for power allocation is to use fixed power transmission. This method has many limitations such as near-far problems where the relatively far away sensors are blocked by near sensors and also the unnecessarily high power consumption for good channel sensors. One recent proposal to overcome such limitations is to use uniform random power allocation for terminals. In that work the system performance is analysed for static channels (i.e. snapshot assumption), where it shows that there is a considerable obtained gain when using uniform random power allocation over fixed power. An important question is about the performance of such transmission power strategy in case of a more realistic channel scenario such as Rayleigh fading channels. This paper answers this question. Detailed mathematical analysis as well as Monte-Carlo simulations for the system performance in Rayleigh fading channels are given in this paper. We show that the performance of uniform power allocation is much worse in case of fading channels than in static channels. Almost there is no gain of using uniform random power over fixed power in terms of average probability of outage. However, uniform random power may considerably reduce the total power consumption at small degradation of the average performance. This can be critical feature in energy constrained networks.
ASJC Scopus subject areas