INTRUSION DETECTION SYSTEM FOR INDUSTRIAL IOT USING MACHINE LEARNING

Abstract

The Internet of Things (IoT) is a term that refers to all of the gadgets that may connect to the Internet in order to collect and share data. Cybersecurity has been widely used in a variety of applications like including intelligent manufacturing processes, homes, personal gadgets, and automobiles, and has resulted in new advances that continue to confront hurdles in tackling problems linked to IoT device security approaches. The increasing number of cyber security attacks against IoT devices and intermediary media for communication supports the assertion. Attacks against the IoT, if undetected for a long period of time, inflict serious service interruption and financial loss. It also poses the risk of identity theft. This project addresses the security difficulties that the Internet of Things IoT and industrial IoT devices confront, and it presents a machine learning technique-based intrusion detection solution for IoT devices. The proposed testbed is divided into seven levels, each of which contains new technologies that meet the critical requirements of basic IoT and industrial IoT applications. The dataset used in this study is Edge-IIoTset, which is a complete, realistic cyber security related dataset of IoT and industrial IoT applications that can be used by machine learning technique driven intrusion detection systems. The dataset comprises sensors for humidity, temperature, water level detection, level of pH, moisture in the soil, pulse rate, and flame detection, among other things. The collection also identifies and assesses fourteen attacks on IoT and industrial IoT communication procedures, which are divided into five categories. The study examines the performance of the different machine learning algorithms after processing and analysing the dataset. Real-time intrusion-based detection on Internet of Things devices is critical for making IoT-enabled services dependable, safe, and profitable.