Research On Key Technologies Of Intrusion Detection For Wireless Sensor Network

With the rapid development of wireless sensor technologies, applications based on wireless sensor network have been widely used. Wireless sensor network has great military value and wide business application prospect. However, owing to the resource constraint and wireless communication characters in wireless sensor networks, the security situation is more rigorous than that in the traditional network.Intrusion detection system provides deep-seated protection for wireless sensor network. Intrusion detection can make up for the shortages of the traditional defense mechanism such as authentication and encryption. Therefore, the research on wireless sensor network intrusion detection technology has great signification.According to the inherent characteristic of wireless sensor network, attacks and intrusion detection features are deeply researched. Based on the analysis results, taking advantages of several tools such as artificial immunity, data mining, fuzzy theory, and propagation model, this thesis proposes intrusion detection algorithms, models and intrusion response policy with high efficiency and performance. These methods are verified with experiments.The main works are as follow:1. This thesis summarizes the recent development in wireless sensor networks and security issues both at home and abroad. Deeply analysis the security requirements of wireless sensor network and its unique security threats.2. An intrusion detection model based on danger theory is proposed. The proposed model adopts distributed and cooperative manner which suitable for wireless sensor network. Anomaly pattern recognition process been triggered only when sensing dangers. The danger sensing process utilizes local knowledge which avoids massive communications between sensor nodes. Nodes are not required to deploy the whole intrusion detection instance and detect intrusion behavior cooperatively. This model can decreases energy consumption with good detection performance.3. Based on the sensing data of wireless sensor network, an intrusion detection algorithm is proposed by mining anomalies in sensing data. Intrusion detection based on sensing data anomalies is not restricted by the network protocol and node type. As a result, the method shows great scalability and suitability. Not only suitable for different wireless sensor networks but also be able to solve the intrusion detection problem in other complex and heterogeneous networks.4. Negative selection algorithm is researched and applied to intrusion detection system of wireless sensor network. Traditional negative selection algorithm has been optimized by adopting fuzzy logic, and the new algorithm called fuzzy negative selection algorithm (FNSA). The proposed algorithm provide a good solution for hole problem of NSA, therefore, improve the detection performance. The experiment results verify that FNSA shows stability and better detection performance when the attacks becoming more deceptive.5. On the basis of intrusion detection results, epidemic attacks control strategies are researched. This thesis takes Worm attack as breakthrough point, analysis propagation characteristic and immune method. Based on the analysis, an environmental adaptation worm control algorithm is been proposed. The proposed algorithm can efficiently stop the epidemic of worm with a minimal number of immunizations which enhance the wireless sensor networks’ ability to resist worm attack.