Smart grid is a cyber-physical system that enhances the capability of conventional power networks leveraging functional automation information and communication technologies. These systems allow energy provider companies to deliver low cost reliable with minimal losses. Despite advantages, such are prone heterogeneous attacks leading breach data integrity confidentiality. A significant part research aims tackle weaknesses smart grids, suggests intrusion detection (IDS) as an effective solution. However, robustness, accuracy, adaptability new major concerns in systems. Therefore, we proposed intelligent for uses deep reinforcement learning. Our IDS robust highly accurate false alarm rate. model based on novel CVAEDDQN architecture, combines generative along Due lack specific datasets, have used benchmark network-based NSL-KDD dataset cloud ISOT-CID dataset. The experimental results show effectiveness our terms accuracy positive rate well network attack capabilities. We also evaluated adaptiveness changes patterns against critical types.
- Dinesh Mohanty
Indian Institute of Technology Bhubaneswar
- Kamalakanta SethiCorresponding
Indian Institute of Technology Bhubaneswar
- Sai Prasath
Indian Institute of Technology Bhubaneswar
- Rashmi Ranjan Rout
National Institute of Technology Warangal
- Padmalochan Bera
Indian Institute of Technology Bhubaneswar