Hybrid Block-Based Lightweight Machine Learning-Based Predictive Models for Quality Preserving in the Internet of...

Area of Science:

• Computer Science
• Network Engineering
• Wireless Communication

Background:

• Wireless Sensor Networks (WSNs) are essential for modern applications, including the Internet of Things (IoT).
• Efficient neighbor discovery (ND) in WSNs is critical for communication, balancing energy efficiency, latency, and discovery rates.
• Existing indirect ND methods have drawbacks, either disrupting wake-up schedules or creating inefficient slots.

Purpose of the Study:

• To develop a novel hybrid approach for neighbor discovery in WSNs.
• To overcome the limitations of existing indirect ND methods by combining their strengths.
• To enhance the performance of WSNs through an optimized neighbor discovery protocol.

Main Methods:

• A hybrid approach, termed HAND (Hybrid Approach to Neighbor Discovery), was developed by integrating two categories of indirect ND methods.
• Virtual nodes were introduced to maintain wake-up schedule integrity and incorporate efficient active slots.
• The approach incorporates a lightweight intrusion detection system (IDS) using feature selection and classification for IoT security during cloud-based medical data transfer.

Main Results:

• The HAND protocol effectively combines the benefits of existing indirect ND methods while mitigating their respective drawbacks.
• Simulation studies demonstrate that HAND significantly outperforms current indirect ND models in WSNs.
• The integrated IDS enhances IoT security for sensitive data transmission.

Conclusions:

• The proposed HAND protocol offers a superior solution for neighbor discovery in WSNs compared to existing indirect methods.
• HAND achieves improved energy efficiency, reduced latency, and a higher percentage of discovered neighbors.
• This research contributes to more robust and secure WSNs, particularly for critical applications like IoT and medical data transfer.