Abstract:
With more sophisticated and integrated systems, the new trend of smart cities has transformed into more smart and intelligent services. Infrastructure for smart cities has adopted several new technologies for sustainability and improvement. The Internet of Things (IoT) is a rapidly developing technology for enhanced smart city infrastructure for better services. IoT applications are a crucial part of smart cities due to their wide range of benefits and simplicity. The Industrial Internet of Things (IIoT), where numerous services relating to operation technologies, manufacturing, utilities, and machine monitoring have been applied to connected devices, is another use of IoT in the industrial sector. The Routing Protocol for Low Power and Lossy Networks (RPL), based on IPV6, is the best option for ensuring efficient data transfer in IIoT by utilizing limited resources. Due to this phenomenon, these networks are vulnerable to several critical security issues that must be resolved. This research proposes an Efficient Detection of Unauthorized Internet of Things (ED-UIoT) method to identify security vulnerabilities in Destination Oriented Directed Acyclic Graph (DODAG) and able to detect DODAG Information Solicitation (DIS) flooding attacks in the RPLbased IoT networks. It utilizing fundamental notions from genetic programming and Machine Learning (ML) techniques. The suggested framework can identify attacks using the dataset’s classification through a Support Vector Machine (SVM). In this research, a system is tested and trained on the generated dataset and then measured its accuracy by using several performance metrics, including attack detection accuracy and confusion matrix, which indicated the false positive and false negative ratio of work. By using these techniques system achieved 98.7% of accuracy. Positive findings validate the suggested framework, making it the best option for RPL-based IoT environments
