AI Enabled Digital Twin Models to Enhance Security in Smart Cities
Aamir Mayan, S Krishanveni, B. Jothi · 2024
As smart cities continue to expand, the security of interconnected systems, such as smart grids, water distribution, and waste management, is paramount. Traditional security measures often lack the predictive capabilities needed to identify and counter cyber threats effectively, posing risks to these critical infrastructures. This research investigates the use of AIenabled digital twin models to strengthen security in smart city environmental management systems. Utilizing Autodesk Tandem as the digital twin platform, we simulate these subsystems and integrate a machine learning (ML) model to monitor network fluctuations, simulate cyber-attacks, and evaluate response strategies. Specifically, a Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) model, trained on the CICIDS2017 dataset, is applied to detect threats in real time. By observing subsystem behavior under attack conditions, we identify vulnerabilities and develop enhanced preventive measures. This approach seeks to create a robust security framework capable of predicting, detecting, and mitigating cyber threats in smart city infrastructures.