Enhancing RPL Security in the Internet of Things for Preventive Network Threats with Machine Learning and NANTAR

Abstract

"Internet of Things" refers to a collection of physical devices that are connected to the Internet and can communicate with one another. Rank, Sinkhole, and Wormhole attacks can seriously impair the performance of the Routing Protocol for Low-Power and Lossy Networks (RPL), which is essential for effective data transfer in Internet of Things networks. Due to its lightweight core, RPL cannot support resource-intensive and computationally intensive security implementation techniques. As a result, security attacks, which may be roughly divided into RPL-specific and sensor-network-inherited assaults, can affect both IoT and RPL. They take advantage of RPL resources and components, such as maintenance methods, routing settings, administrative messages, and network sensor components. The research presented here suggests an innovative machine learning-based method to increase RPL's security. We present a Novel Algorithm for Network Traffic Analysis and Response (NANTAR). This innovative approach that uses AI-based techniques to optimize the security and efficacy of RPL routing, in conjunction with a reinforcement learning module for dynamic and adaptive routing. This framework significantly improves packet false positive  rate, lowers latency, and boosts network throughput while maintaining minimal jitter.

 It effectively secures the RPL protocol in IoT-enabled wireless sensor networks by achieving a high detection rate, few false positives, and quick reaction to security problems.