Optimal Cluster Head Selection in WSN with Convolutional Neural Network-Based Energy Level Prediction.

Area of Science:

• Computer Science
• Electrical Engineering
• Network Security

Background:

• Wireless Sensor Networks (WSN) are crucial for data collection, but energy consumption and security remain significant challenges.
• Existing clustering protocols like Low Energy Adaptive Clustering Hierarchy (LEACH) face limitations due to arbitrary cluster head selection, impacting network efficiency.
• The need for robust and energy-aware routing protocols is paramount for the sustained operation of WSNs.

Purpose of the Study:

• To develop an optimal cluster head selection (CHS) model for secure and energy-aware routing in Wireless Sensor Networks (WSN).
• To enhance the reliability and efficiency of WSNs by improving cluster head selection criteria.
• To address the limitations of arbitrary cluster head selection in existing protocols.

Main Methods:

• An optimal CHS model was developed, considering factors like distance, energy, security risk, delay, trust (direct and indirect), and Received Signal Strength Indicator (RSSI).
• Energy levels were predicted using an improved Deep Convolutional Neural Network (DCNN).
• The Bald Eagle Assisted Sparrow Search Algorithm (BEA-SSA) was employed for selecting the optimal cluster head in WSN.

Main Results:

• The BEA-SSA model demonstrated superior performance in cluster head selection based on trust, RSSI, and security metrics.
• The proposed model achieved a high Packet Delivery Ratio (PDR) of 0.98 for 100 nodes at 500 rounds.
• Performance was significantly better than established optimization algorithms including GWO, MOFPL, SSA, BES, ROA, HGS, SSO, RCSO, and FCR.

Conclusions:

• The BEA-SSA model provides an effective solution for optimal cluster head selection in WSNs, enhancing both security and energy efficiency.
• The integration of DCNN for energy prediction and BEA-SSA for selection offers a robust approach to WSN routing challenges.
• The study validates the effectiveness of the proposed method, paving the way for more reliable and sustainable WSN deployments.