Emerging framework for attack detection in cyber-physical systems using heuristic-based optimization algorithm.

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Cyber-Physical Systems (CPS) are increasingly integral to modern life, featuring complex networks for data sharing and remote access.
• The interconnected nature of CPS necessitates robust security measures to protect sensitive information and system resources from network attacks.
• Existing research has focused on detecting insecure networks and attacks, but enhanced solutions are needed for CPS environments.

Purpose of the Study:

• To introduce a novel framework, Deep Bagging Convolutional Neural Network with Heuristic Multiswarm Ant Colony Optimization (DCNN-HMACO), for improving security in Cyber-Physical Systems.
• To enhance the secure transmission of information, efficiency, and convenience within CPS networks.
• To effectively detect and mitigate attacks in Cyber-Physical Systems.

Main Methods:

• Development of the Deep Bagging Convolutional Neural Network with Heuristic Multiswarm Ant Colony Optimization (DCNN-HMACO) framework.
• Implementation of advanced deep learning and optimization techniques for attack detection.
• Comparative analysis against existing methods like Convolutional Neural Networks (CNN) and Fuzzy C-Means (FCM).

Main Results:

• The proposed DCNN-HMACO framework demonstrates a significant improvement in attack detection rates.
• The framework enhances overall system protection in Cyber-Physical Systems.
• Achieved a remarkable accuracy rate of 92.14%, surpassing existing methods (CNN at 72.12%, FCM at 79.56%).

Conclusions:

• The DCNN-HMACO framework offers a superior solution for enhancing security in Cyber-Physical Systems.
• The framework effectively detects attacks, ensuring more secure data transmission and system integrity.
• This approach represents a significant advancement in protecting CPS from cyber threats.