Deep Learning-Based Privacy-Preserving Data Transmission Scheme for Clustered IIoT Environment.

Area of Science:

• Computer Science
• Artificial Intelligence
• Cybersecurity

Background:

• The Industrial Internet of Things (IIoT) is crucial for Industry 4.0, generating massive data from diverse devices.
• Security and privacy of IIoT data stored in cloud servers present significant challenges.
• Existing solutions struggle with the scale and heterogeneity of IIoT data.

Purpose of the Study:

• To develop a novel deep learning-based privacy-preserving data transmission (BDL-PPDT) scheme for clustered IIoT environments.
• To ensure secure data transmission and protect data privacy in IIoT systems.
• To enhance intrusion detection and data security within the IIoT infrastructure.

Main Methods:

• A two-stage process involving enhanced moth swarm algorithm-based clustering (EMSA-C) for cluster formation.
• Utilizing a multi-agent system (MAS) for secure inter-cluster communication.
• Employing a multi-head attention with bidirectional long short-term memory (MHA-BLSTM) model for intrusion detection, optimized by stochastic gradient descent with momentum (SGDM).

Main Results:

• The BDL-PPDT technique demonstrated superior performance compared to existing methods in simulation studies.
• Achieved a high accuracy of 98.15% in intrusion detection and secure data transmission.
• The proposed scheme effectively addresses security and privacy concerns in clustered IIoT environments.

Conclusions:

• The BDL-PPDT scheme offers a robust solution for secure data transmission in IIoT.
• The integration of MAS and deep learning models significantly enhances IIoT security.
• The technique is recommended for its effectiveness in protecting sensitive IIoT data.