Research Selfish Behavior IEEE 802.15.4 Wireless Sensor Network Nodes

IEEE802.15.4is the most popular standard in MAC layer of Wireless Sensor Networks (WSNs). With the development of wireless network technology and programmable network equipments, it provides convenient for MAC layer selfish behavior, while the selfish behavior of the nodes may seriously affect network performance. Therefore, it is very necessary to study nodes selfish behavior in IEEE802.15.4Wireless Sensor Networks.At first, this paper studies nodes selfish behavior for network performance and gives NS-2simulation experiments by modifing IEEE802.15.4backoff mechanism.Secondly, selfish behavior detection is the key to solve nodes selfish behavior. To solve the detection problem of selfish behavior in IEEE802.15.4Wireless Sensor Networks, an improved Sliding Window Non-parameter Cumulation Sum (SWN-CUSUM) algorithm based on statistics is proposed. By tracing the delay characteristic sequence between successful transmissions, a node can distinguish if there is a selfish behavior in the WSNs.The experimental results on NS-2show that the improved algorithm can effectively and quickly detect the nodes selfish behavior of the IEEE802.15.4Wireless Sensor Networks.Once more, this paper further presents a Network Performance Characteristic Serial based-Clustering Detection Method (NPCS-CDM) for multi selfish nodes networks. NPCS-CDM takes the average delay and the average throughput of each node link as the network performance characteristic, and then establishes statistic sequence, based on which K-means clustering algorithm is used to classify the data and distinguish the selfish behavior from the normal behavior. The NS-2simulation tool was conducted to validate the feasiblity of the proposed algorithm.The simulation results indicate that:The NPCS-CDM algorithm not only can determine if there is a selfish behavior in the network, but also can distinguish the normal nodes and selfish node. When the numbers of selfish nodes increase, the algorithm can maintain high classification efficiency.At last, in order to control the nodes selfish behavior,an Incentive Detection Mechanism(IDM)for cooperation of nodes selfish behavior is proposed. This mechanism consists of two parts which are detecion module based on retransmission numbers and punishment module. The detection module determines if there is a selfish behavior in the networks by collecting data from retransmission numbers of all nodes, and the punishment module change the strategy of selfish nodes for cooperation with normal nodes so that the network performance can improve. The NS-2simulation tool was conducted to validate the feasiblity of the proposed algorithm. The experimental results indicate that the IDM can improve the detecting rate and lower the false detecting rate in appropriately selected detection mechanism parameters, and every node link can obtain a relatively fair throughput to improve the performance of overall networks when following the mechanism of proposed.