A MANET Intrusion Detection System Based On Improved Clustering Algorithm

Mobile Ad Hoc network, (MANET, growing independently), is a multi-hop temporary autonomous system which is composed of a set of Mobile terminal with radio transceiver. It connects network and transfers data by the mobile nodes' collaboration and self-organization not by any fixed infrastructure. This kind of network can work independently or connect with the Internet or cellular wireless network. MANET has the great strategic significance and vast commercial prospects for its flexible networking form, convenient configuration and low cost.Since MANET is widely used in many fields, it becomes more and more necessary to make research on its security. The traditional network security measures no longer apply to MANET. It has been a necessary measure to use intrusion detection technique in MANET to guarantee the network security. Intrusion detection technique is the core of second network security technology, so it becomes a hot research on how effectively set up suitable intrusion detection sysytem for MANET.This thesis is committed to research a suitable intrusion detection system for MANET to deal with security problems and threats faced by the MANET. Firstly studies basic principle of MANET, analyzes the defects of the present clustering algorithms and proposes an improved clustering algorithm for the intrusion detection system by showing its specific calculation method. Then by using NS2 simulation software platform simulating a comparative experiment between improved clustering algorithms and typical clustering algorithms and makes a detailed analysis about the experiment result to testify validity and superiority of the improved clustering algorithm. Finally, On the basis of the above analysis, the thesis designs a MANET Intrusion Detection System based on the improved clustering algorithm and explains the details of structure and principle of the system as well as analyzes and designs the main functional modules in the system.However, the improved clustering algorithm in the thesis still has the room for improvement, required further study and improvement in the future work.