IOT Routing Attack Detection Using Deep Neural Network

Abstract

The internet of things (IoT) devices becomes omnipresent as IOT resources become all-encompassing. Their success has not gone unnoticed and there are still growing numbers of attacks and assaults on IOT products and services. Cyber-attacks aren’t new to IOT, but since IOT is profoundly embedded within our lives and cultures, cyber security becomes a must. There is also a real need to protect IOT, and therefore the risks and assaults on IoT networks must be grasped thoroughly. This thesis aims to categories threat categories as well as evaluates and describes intruders and assaults that IoT devices and services face. The objective of this thesis has been aimed a deep-learning-based machine learning addition to analyzing and characterizing intruders and attacks in IOT devices and services, this study is an attempt to identify category of threat. A deep learning machine method for the revelation of routing attacks for IoT was provided as the purpose of this thesis. The Cooja IoT simulator was utilized in our research for the development of high-reliance attack results. For detecting deep learning attacks for IoT path attacks, we provide a highly scalable solution. They have high precision and high precision Kiru and Mirai.ng technique for the revelation of routing attacks for IoT. Consequently, the Cooja IoT emulator was utilized for high resolution attack data generation. We propose a highly scalable, deep-learning-based attack revelation techniqueology for the revelation of IoT routing attacks which are Kiru, and Mirai with high accuracy and precision. The application of deep learning for cyber-security in IoT requires the availability of substantial IoT attack data and we believe that the IoT attack dataset produced in this work can be utilized for further research. The results analysis offers a deep Neural Network model to robustly classify routing attacks with a 99.9 percent accuracy assessment scoring for Kiru and Mirai attacks, which forms a robust model for preventing most attacks from impacting subsequent layers in the sensors or actuators. The Cooja network simulator based on the Contiki OS is utilized to demonstrate in real-time simulation the functioning of this routing and compression protocol. The results show that the utilize of the Deep learning technique in IoT security is a promising solution to the challenges facing security.