CR-IOT based selfish attack detection via RSSI-LSTM

Abstract

Internet of Things-based technologies rely on cooperating between nodes to increase network capacity. A selfish or malicious node is a node that does not cooperate with other nodes in the network. Selfish nodes raise their interests by using facilities provided by other nodes. Selfish cognitive radio attacks significantly degrade the performance of Cognitive Radio networks, which pose a serious security threat and malicious nodes often abuse the network and damage its facilities. To overcome this issue, a novel Cognitive radio on-demand distance vector (CRODV) is proposed. The proposed CRODV technique utilize (Received Signal Strength Indicator) RSSI-based Long Short-Term Memory (LSTM) to detect attacks. This is accomplished by counting the number of active channels to and from the test node in a specific area. If the values don't match, the test node in question is self-centered or introverted. If both values are the same, the test node is normal. Simulations were carried out in MATLAB to test the proposed CRODV. A comparison is made between the suggested framework and existing methods in terms of packet delivery ratio, route dis-connectivity ratio, throughput, detection time, and end-to-end delay. The experimental results demonstrates that the CRODV technique reduce the end-to-end delay by 27.4%, 33.24%, and 40.14% when compared with UKF, HMM, and ALAD techniques.