DESIGN AND DEVELOPMENT OF AI BASED INTRUSION DETECTION AND MONITORING DASHBOARD FOR PID- CONTROLLED SYSTEMS

Abstract

This paper describes the design and development of an AI- based intrusion detection and monitoring system for PID- controlled industrial environments. The project addresses the increasing issues of cybersecurity in Industrial Control Systems (ICS) as a result of the convergence of operational technologies and networked systems. We constructed a simulated cyber-physical environment, which produces real-time operational data from PID-controlled processes like temperature and water-level regulating. The system includes attack simulation techniques such as replay attacks, noise injection and false data injection (FDI) attacks to analyse the industrial processes security vulnerabilities.

Machine Learning methods such as Random Forest and Logistic Regression are used for real time cyber intrusion detection and anomaly detection. Finally, the reliability and efficacy of the detection performance are evaluated by accuracy, precision, recall, F1-score and confusion matrix analysis. A web-based dashboard is combined with a real-time detection engine providing live monitoring, visual analytics, alarms and logging of historical events to improve situational awareness.

The suggested system makes a contribution in the field of industrial cybersecurity by combining artificial intelligence, intrusion detection, and PID control system dynamics into a single framework. This research presents a scalable and effective strategy for safeguarding current ICS and Industrial IoT infrastructures through adaptive real-time anomaly detection inside a simulated industrial setting, contrasting with classic static IDS methods.