Cluster Based Malicious Node Detection System for Mobile Ad-Hoc Network Using ANFIS Classifier

Abstract Improvement of efficient packet access in a wireless Mobile Ad-Hoc network (MANET) is vital for achieving high speed data rate. The degradation occurs due to identification of malicious node and hence, reducing the severity will be a complex problem due to similar characteristics with trusty nodes in sensing area. In this work, Adaptive Neuro Fuzzy Inference System (ANFIS) classifier based defected node identification system is developed. The conviction parameters to be extract of the reliable and malevolent nodes and these parameters are qualified by ANFIS classifier. Further, the individual nodes in MANET are classified in testing mode of classifier. The network performance will be degraded with the increased number of malicious nodes. Certain conditions like packet delivery ratio, throughput, detection rate, energy consumption, and precision value and link failures occur due to malicious node in the network. The anticipated malicious node detection structure be compare by means of the conservative techniques such as Optimized energy efficient routing protocol (OEERP), Low energy adaptive clustering hierarchy (LEACH), Data routing in network aggregation (DRINA)and Base station controlled dynamic clustering protocol (BCDCP). The proposed ANFIS classifier is designed in Matrix Laboratory (MATLAB) and it can be interfaced with NS2 using “c” programming.