A secure worst elite sailfish optimizer based routing and deep learning for black hole attack detection.

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Wireless Sensor Networks (WSNs) are vulnerable to active and passive attacks.
• Black hole attacks pose a significant threat to WSN functionality and data integrity.
• Existing detection methods may lack efficiency in identifying and mitigating these attacks.

Purpose of the Study:

• To design and implement a Deep Learning (DL) based model for detecting and mitigating black hole attacks in WSNs.
• To enhance network security and ensure reliable data transmission in WSNs.
• To optimize routing and attack detection processes.

Main Methods:

• A WSN simulation environment was established.
• The Worst Elite Sailfish Optimization (WESFO) algorithm was employed for optimal routing to the base station (BS).
• An Auto Encoder (AE) model, trained using WESFO, was utilized for black hole attack detection at the BS.

Main Results:

• The proposed model achieved a low delay of 25.64 seconds.
• A high Packet Delivery Rate (PDR) of 94.83% was recorded.
• The model demonstrated effective attack detection with a False-Negative Rate (FNR) of 0.084 and a False-Positive Rate (FPR) of 0.135.

Conclusions:

• The developed DL-based model effectively detects and mitigates black hole attacks in WSNs.
• The integration of WESFO for routing and AE for detection enhances network performance and security.
• The model shows promising results in terms of key performance metrics like PDR, delay, and accuracy.