CELLULAR AUTOMATA-BASED ANOMALY NODE DETECTION IN WIRELESS SENSOR NETWORK

Wireless Sensor Networks (WSNs) are used in a broad range of applications where monitoring and detecting some events play a vital role. The aforementioned application scenario needs a higher level of security. Since the mode of communication is wireless and the scenario of deployment is sensitive, WSN is more vulnerable to major attacks. This study is intended to detect the nodes that behave abnormally and to isolate them from the network. For the detection of such nodes, a Cellular Automata (CA) based anomaly detection model is proposed here. A special CA called Single Attractor Cellular Automata is used for the classification of nodes. Initially the rules that form the attractor is analyzed and selected accordingly. Each node in the network transmits its trust value to the base station where its bits are set accordingly. Based on the trust values, the rule vector is formulated. Thus attractor is run for detecting and locating the anomaly nodes. Since CA could be applied in WSN without any computation overhead, the proposed model will detect the abnormal nodes within the minimum bandwidth and time. The detection performance of the proposed study is proved by determining the metrics such as detection rate and false alarm rate. The network performance of the proposed study is measured in terms of packet delivery ratio and routing overhead. It is proved that it performs well even in the presence of more number of anomaly nodes.