Development and Validation of a Cyber-Physical System Leveraging EFDPN for Enhanced WSN-IoT Network Security

Area of Science:

• Cyber-physical systems security
• Network intrusion detection
• Machine learning applications in IoT

Background:

• Wireless Sensor Networks (WSNs) and the Internet of Things (IoT) are integral to modern interconnected environments.
• Existing security models for WSN-IoT networks face challenges including lengthy training times and complex attack classification.
• Robust security solutions are crucial for protecting these evolving cyber-physical systems.

Purpose of the Study:

• To propose a novel cyber-physical system, the Emphatic Farmland Fertility Integrated Deep Perceptron Network (EFDPN), for enhanced WSN-IoT security.
• To reduce computational complexity in attack identification and classification.
• To improve the accuracy and efficiency of intrusion detection in WSN-IoT networks.

Main Methods:

• Development of the EFDPN framework integrating Deep Perceptron Network (DPN) for classification.
• Introduction of the Farmland Fertility Feature Selection (F³S) technique to minimize feature set complexity.
• Utilization of the Tunicate Swarm Optimization (TSO) model to refine the sigmoid transformation function for improved prediction accuracy.

Main Results:

• The EFDPN system demonstrated superior performance in classifying cyber-attacks compared to traditional methods.
• The F³S algorithm effectively reduced irrelevant features, enhancing classifier prediction accuracy.
• The TSO model significantly improved the classification performance of the DPN algorithm, leading to higher F1-score values.

Conclusions:

• The proposed EFDPN-based system offers a robust and efficient solution for securing WSN-IoT networks.
• The integration of F³S and TSO models provides a significant advancement in intrusion detection capabilities.
• This research contributes a promising approach to bolstering the security and resilience of interconnected cyber-physical systems.