Convolutional Neural Network Based IOT Intrusion Detection System using Edge-IIoTset
Mahima Singh, Naveen Chauhan · 2024
The world of IoT security is rapidly evolving, especially with the rise of advanced technologies. The paper is a study of how Convolutional Neural Networks into Intrusion Detection Systems (IDS) for Edge Industrial Internet of Things (EdgeIIoT) datasets can greatly improve the security of IoT devices. By utilizing deep learning and CNNs in IoT security, we can develop advanced anomaly detection algorithms capable of identifying unusual patterns and behaviors that may indicate potential security threats. The primary benefit of utilizing CNNs for Intrusion Detection Systems (IDS) in EdgeIIoT security lies in their capacity to autonomously learn and extract features from unprocessed data. This capability proves especially valuable when managing the extensive streams of varied data produced by EdgeIIoT devices. This paper explores how CNN and its variations, Xception and VGG16 (Visual Geometry Group 16), help improve the security of an Edge-IIoTset. Generic CNN tops both Xception and VGG16 models with the highest accuracy of$98.98\%$. The vast field of IoT security is growing rapidly, especially because of advanced technologies. This case study delves into the application of CNNs in IDS for EdgeIIoT datasets, displaying a boost in security measures for IoT devices.