Research On Intrusion Detection In WSN Based On Projection Pursuit

The wireless sensor network(WSN) as a new kind of mobile network more and more widely applied in military deployment and commercial areas. WSN generally work in the deserted area, so the nodes in Wireless Sensor Network security is urgent problem to be solved. Intrusion detection system is an effective means of defense in WSN, collecting of network nodes of a variety of network data and analyzing the collected information state, effectively capturing malicious access behavior.Due to the high complexity and high power consumption, the intrusion detection model in traditional network is not completely suitable for WSN. Therefore, the intrusion detection system with high detection accuracy in resource constrained nodes is the key to the research of WSN. In this paper, first we analyze the security threats of WSN, and summarize the common WSN intrusion detection technology, we designed and implemented a novel intrusion detection mechanism based on WSN. The main achievements are as follows: Firstly, This paper introduces the WSN development and application background, and points out that the WSN joined the importance of security defense mechanism, analysis the WSN faces security problems with means of defense, summarizes the current domestic and foreign WSN intrus io n detection mechanism research status.Secondly,putting forward using projection pursuit algorithm(project pursuit, PP) model is applied to WSN intrusion detection, the PP model of high-dimensional data processing advantages, effectively reduce the computational complexity and reduce network energy consumption and the simulation experiment proved that the detecting system based on GA-PP algorithm to improve the resolution of real-time and emergencies has better detection rate.Finally, we analysed the deficiencies of traditional optimization algorithm, then detailed introduction of the national competitive mechanism of a new optimization algorithm-Imperialist Competitive Algorithm(ICA)and summed up the advantages of ICA algorit hm compared to the traditional algorithm. The PP model is optimized using ICA algorithm, and simulation experiments are used to prove the ICA algorithm to optimize the PP model is better than the traditional optimization model,it has higher detection rate and higher convergent efficiency, real-time better, the algorithm consumes energy network is relatively lower.