Research On Repeated Game Modeling For Intrusion Detection In Wireless Sensor Network

Information security guarantee technology is one of the key technologies of wireless sensor network (WSN). Due to the unique properties and severe energy constraints of WSN, security mechanisms used in traditional networks cannot be applied directly in WSN. So studying intrusion detection of information security guarantee technologies of WSN has very important academic value and significant to the application systems construction of WSN. According to the analyses of the security requirements and the existing intrusion detection algorithms of WSN, we analysis the attack behaviors and defense behaviors respectively between attacker and intrusion detection system (IDS). Based on game theory, we propose a static game model and a repeated game model for intrusion detection in Wireless Sensor Network. Simulation results indicate that the proposed frameworks are effective.The main research works are as follows:1,Introduce the status of domestic and overseas research in information security guarantee technology of WSN briefly, and put forward the research background, meaning, and main task.2,Analyze the security requirements in wireless sensor network and the researches of existing intrusion detection algorithms, prepared for the future work.3,According to the analyses of intrusion detection approaches in cluster-based WSN,we propose a improved static game model: the attacker and the IDS choose to attack and defense the network respectively with some probability. And show that this game achieves mixed Nash Equilibrium, thus leads to a defense strategy for the network, and significantly increases the chance of detecting intrusions.The game takes the severe energy constraints of WSN into full consideration and effectively prevent the attack against nodes in the network with an energy-saving mode.4,The lack of cooperation among the nodes of a network may seriously affect the network functioning. In order to solve the problem, we propose a repeated game framework to optimize packet forwarding probabilities of distributed nodes by detecting and responding to the selfish behavior. By the repeated game among nodes and broadcasting the monitored nodes' utility change, the framework could detect the selfish behavior in the network in time. We provide a punishing mechanism to prevent the selfish nodes' deviation from cooperation. From the simulation results, we verify that with the punishment mechanism in the model, selfish nodes can be forced to participate to the network operation and maintain the connectivity of network.