Research On Small World Theory And Energy-Efficent Algorithm Of Wireless Sensor Networks

Unlike traditional networks, Wireless Sensor Network (WSN) exhibits multi-hop self-organizing property due to deployment flexibility, low cost, small size of sensor nodes and WSN is currently one of the hot spots in the field of wireless communication research. In WSN, the sensor nodes have small communication range and asymmetrical communication load, and are easily affected by external jamming attack. Consequenty, the transmission optimization in WSN is important. In addition, the power supply of the sensor nodes is limited, so it is important to impove the power efficiency under the condition of satisfying the sensor node transmission rate.To solve the above problems, the small world model is introduced to characterize the topology structure and network robustness under the external jamming attack. In this paper, the metrics including global efficiency and local efficiency are introduced to measure the performance of WSN. In addition, a novel multi-dimensional subcarrier power allocation algorithm based on node-cognitive property is proposed to improver the power efficiency of sensor nodes.The main research work of this dissertation are summarized as follows:Firstly, the topology structure of WSN based on small world model is constructed. After introducing super node in WSN, the rewiring probability of small model is used to characterize the propery of super nodes. Then the WSN model with small world propery is achieved. Furthermore, according to optimizing the rewiring probability and the long-range edge weight, the network efficiency of WSN is promoted. At the same time, the best number of super nodes can be obtained to optimize the WSN performance and the simulation result shows that only a few super nodes can remarkly reduce the average path length and improve the network efficiency.Secondly, the anti-jamming-attack propery of WSN is analyzed according to the small world model. In WSN, the jamming attack can be treated as the removal of the node or edge in the topology and the level of jamming attack can be measured by the removal probability of the corresponding node or edge in network topology. In this paper, the focus is on the following three kinds of jamming attack, incluing the attack on the long-range edges, the attack on the edges with highest edge betweenness centrality and the attack on the vertices with highest vertex betweenness centrality. The performance of anti-jamming-attack is promoted by optimizing the rewiring probability and node degree. The simulation results show that our optimization method can improve the network robustmess and provide the solid theory foundation to rebuild the network under external jamming attack.Thirdly, the optimization algorithm of the sensor node power efficiency is proposed. In our research, the sensor nodes have cognitive capability and the multi-carrier transmission structure is applied. In addition, during one time slot, each sensor node is allowed to occupy multiple subcarriers. The power efficiency is defined as the ratio of the transmission power and the information rate. The optimization goal related to power efficiency is shown to be non-convex. In the proposed algorithm, the original non-convex problem can be transformed into a convex problem. In addition, the equivalence of the optimum solution beween the original problem and the transformed problem is proven. Furthermore, the distributive power allocation algorithm is raised to solve the transformed problem. Simulation results show that the proposed distributive algorihm exhibits better convergence than the exsiting algorithm and can approach the optimum solution of centralized algorithm.