A novel extended Internet of things (IoT) Cybersecurity protection based on adaptive deep learning prediction for industrial manufacturing applications

The use of the Internet of things (IoT) devices is growing rapidly in the current industrial market. Integrating these devices into the network system has subjected the information to Cybersecurity attacks putting the market of IoT technology at risk. As a result, accurate protection algorithms and efficient monitoring systems need to be developed. This paper proposes a new framework design relays on a stochastic measurement of determining parameters based on a novel adaptive deep learning (ADL) algorithm for industrial manufacturing such as transportation, energy. The new framework will be designed to incorporate an element of randomness as opposed to deterministic frameworks. The implemented method considered the network forensic systems and intrusion detection (ID); therefore, the prediction model is set up to alert the administrator about the system status. The framework is proposed based on five prioritized protection stages. The developed model tested on the Bot-IoT dataset, which is well-structured and valid for training, as confirmed by literature. The result has shown the ability to simulate the IoT network traffic along with risk assessment against different types of attacks by using the ADL algorithm. The new framework will be designed to improve computational efficiency by being able to adjust their structure based on the inputs. The research contributions are that it will help in the development of an algorithm which can minimize traffic accidents, improve traffic flow, minimize energy wastes, and save costs of controlling traffic flow. Findings show that using this new algorithm has 100% accuracy, which makes it an excellent choice for the detection of threats and errors.