Optimizing Power Allocation in Sensor Networks with Application in Target Classification

In the present work, a system of distributed wireless sensor networks with applications in target detection and classification is investigated. In particular, aspects of mathematical modeling, methodologies for resource management, and performance analyses are studied. The ultimate goal of present investigations is to find analytical solutions in closed-form for optimizing the power allocation and hence minimizing the energy consumption of the entire network while keeping the classification performance constant. Since closed-form solutions are very difficult to obtain and only achievable in particular cases, this work presents first novel results for dealing with power allocation in distributed radar systems. Passive and active radar systems are investigated under sum-power and individual power constraints per sensor node. In addition, the corresponding classification performance, exemplarily for the active radar system, is determined. This leads to the investigation of a certain class of integrals concerning the evaluation of the average classification probability for communication over fading channels. Since the evaluation of average classification probability and average symbol-error probability for communication over fading channels are closely related, the corresponding mathematical integrals are thoroughly scrutinized. Finally, new equivalent equations and novel mathematical expressions as well as accurate analytical bounds in closed-form are presented.