Power Control In Wireless Sensor Networks Based On Network Utility Maximization

Wireless sensor network (WSN) is composed by the wireless sensors with the ability of sensing, data processing and wireless communications which can work in a self-organization manner. Since WSN can monitor the targets and collect the information of environment through the cooperation of large amount of sensors, the sensing performance and reliability have been improved significantly. So it is widely used in military sensing, environmental monitoring, traffic, medical treatment, even surveillance, etc.However, the energy of sensor nodes are severely energy-constrained , the energy consumption caused by wireless transmission is the main source of energy consumption. How to control the power during transmission to satisfy data transmission requirement with low energy consumption becomes an important problem in WSN. As a very effective method to design the network architecture, network utility maximization (NUM) plays an important role in researching network resource allocation algorithm, this method also is suitable to design the power control algorithm in WSN.This thesis firstly analyzes the characteristic of power control in WSN. Secondly, the basic model of NUM and the fruits of using NUM to do research in network architecture and resource allocation algorithm research have been introduced. Thirdly, a NUM model to describe WSN has been proposed. This model has taken the collaboration of transmission layer and physical layer into consideration, can adjust the transmission power and rate very efficiently to attain the maximum of network utility. In this model, the convex function of power is taken into consideration to measure the network utility. As a result of that, the proposed model gives a more accurate description on the energy consumption in WSN comparing with related frameworks. Dual decomposition technology is adopted to solve this optimization problem. Through solving this model, a distributed power and rate control algorithm in WSN has been achieved, numerical simulation shows the convergence of the proposed algorithm. The discussion of simulation result also has given some conclusions on power, rate and node allocation in WSN.