Dair-mlt: detection and avoidance of IoT routing attacks using machine learning techniques

The Internet of Things (IoT) aims to increase the physical device’s intelligence. These devices are capable of exchanging data without human intervention. But, IoT devices are resource-constrained and also prone to attacks during routing. These attacks deplete the energy and lifetime of each node in the network thereby gradually degrading the performance of the network and possibly bringing it to a halt. To overcome these problems, this paper proposes a novel Detection and Avoidance of IoT Routing Attacks using Machine Learning Techniques (DAIR-MLT) to detect and avoid Hello Flooding attacks, Rank attacks, and Version Number attacks. The Cooja simulator is used for simulating the proposed DAIR-MLT Techniques. In the DAIR-MLT, the detection and avoidance of attacks are carried out in two stages namely Detection of IoT routing Attacks and Avoidance of detected attacks. Two different datasets are used namely DA_IoT_Routing Normal Datasets and DA_IoT_Routing Abnormal Datasets to test and analyze the performance of proposed DAIR-MLT Techniques. From the simulation results, it is inferred that the proposed algorithm increases the packet delivery ratio by 41.55%, throughput by 39.56%, and network lifetime by 43.2% compared to existing algorithms. Further, it is found that the proposed DAIR-MLT algorithm decreases the energy consumption of nodes in IoT by 40.16% and End-to-End delay by 45.26%.