Cyber- Physical System for AI - Based Models

The need CPSs, or cyber-physical systems, have expanded across various domains, including intelligent transportation, smart grids, and vital infrastructures. Consequently, the extensive data network and communication layers that CPSs rely on render them susceptible to cyber threats and attacks. Artificial intelligence (AI) techniques are commonly used to address these vulnerabilities. However, AI models frequently find it difficult to stay up with rapidly evolving threat landscapes. The present research examines the use of extended both extreme gradient boosting (XGBoost) and short-term memory (LSTM) artificial intelligence models in detecting cyberattacks in critical infrastructure systems. The evaluation metrics used to validate the approach included accuracy, precision, recall, and the F1-score. The suggested method was evaluated using a gas pipeline industrial control system dataset and two additional benchmark datasets, namely NetML- 2020 and IOT-23, which feature a range of cyberattacks. The performance of the XGBoost and LSTM models were superior to other models, such as support vector machine (SVM) and artificial neural networks (ANN), on several evaluation metrics. Finally, recommendations for future studies in this field are provided.