Enhancing IoT Security in Nuclear Power Plants: Deep Learning Approaches to Detect Mirai Attacks
Ankita Kumari, Deepali Gupta, Mudita Uppal · 2024
Nuclear power facilities are installing a growing number of IoT devices; hence, significant security measures are required to resist against complex cyberthreats like the Mirai botnet. This research suggests a new utilization of artificial neural networks to aid the early detection and modeling of Mirai assaults. We can secure crucial assets by employing artificial neural networks to uncover patterns and operate as a proactive protection system that can identify bad activity in real time. One of the options may be to design a smart chip with an ANN algorithm incorporated in. This gadget is supposed to constantly monitor network traffic and seek for peculiar actions that could be suggestive of Mirai swarm activity. This smart chip decreases the likelihood of catastrophic injury by allowing individuals to respond immediately when they feel oncoming danger. The ANN system can appropriately identify risks since it was trained on a substantial amount of network data, which includes patterns that were equally benign and damaging. The 90% high successful rate ANN-based monitoring system demonstrated its durability and usefulness by identifying Mirai attacks. The F1-score, accuracy, and memory were some of the measures used to assess the model's performance. The findings demonstrated that the technology performed a decent job of differentiating harmful from safe traffic. Furthermore, the system's real-time data processing capabilities guarantees that such threats are swiftly discovered and managed, minimizing the window of opportunity for assaults. The inclusion of this smart chip to the IoT networks of nuclear power plants marks a huge improvement in cyber security as it offers an extra layer of protection against increasingly complicated assaults.The ANN model will be adjusted to increase performance and eliminate false positives in the future trial. To boost the overall efficiency of the system, a combination of various deep learning algorithms will also be studied. By continuing the development of these cutting-edge security technologies, we can assure the stability and safety of nuclear power stations against continuously evolving cyber threats.