Attack Detection Techniques For Wireless Sensor Networks

Abstract

Security is a vital problem in many real-time applications of wireless sensor networks (WSNs), which is one among the exploring technologies and it is always susceptible since it is normally exploited in hostile environments. WSN comprises more amounts of sensor nodes (SNs) that gather the data. The gathered data are subjected to sink by intermediate nodes. While SNs data are uncovered to internet, there is a probability of susceptibility in WSN. WSN is executed utilizing low priced SNs, which is inhibited with respect to memory, and battery life. WSNs monitors environmental situations and are generally exploited in places which cannot be accompanied by humans. The SN is attacked in several ways and is susceptible where a clone attack is the dangerous attacks. In a clone attack, the SNs are attained and copy more sensors with similar ID. Personal information such as credentials is utilized in other networks. Due to the complex identification of clone node, the attacker deploys these situations and utilizes it as base for black hole attacks. Therefore, predicting such loose points in WSN is a difficult task. Therefore, this thesis explores two contributions of attack detection techniques in WSN to ensure that data is not tampered with or stolen, and maintain its integrity and accuracy. The first contribution describes trust based model for detecting attack in WSN. Here, two step verification methods are designed to predict malicious nodes. When energy consumption is abnormal, Cluster Heads (CHs) identify untrusted node. The additional nodes utilized in this network observed the network traffic. Based on the observations, the node is considered as trusted or untrusted. If CH and observer node considered the node as untrusted one, the SN will be assumed as malicious node. newline