Sensor Clustering Using a K-Means Algorithm in Combination with Optimized Unmanned Aerial Vehicle Trajectory in...

Area of Science:

• Wireless Sensor Networks (WSNs)
• Unmanned Aerial Vehicle (UAV) Communications
• Machine Learning in Networking

Background:

• Wireless sensor networks (WSNs) face challenges in large, complex areas regarding deployment flexibility, cost, and reliability.
• Efficient data relay and signal processing are crucial for maintaining network performance.

Purpose of the Study:

• To propose a novel framework for enhancing WSNs using UAVs as flying relays.
• To optimize sensor clustering and UAV positioning for improved network efficiency.
• To develop a robust signal decoding mechanism to mitigate interference.

Main Methods:

• Application of unmanned aerial vehicles (UAVs) as relays employing non-orthogonal multiple access (NOMA).
• K-means unsupervised machine learning combined with the gap statistic method for sensor clustering and UAV positioning.
• Development of a centroid-to-next-nearest-centroid (CNNC) path algorithm for UAV trajectory optimization.
• Implementation of a diagonal matrix phase-shift framework for signal separation and decoding at the UAV.

Main Results:

• The K-means algorithm effectively optimizes sensor clustering and UAV deployment.
• The proposed CNNC path algorithm enhances UAV trajectory efficiency.
• The diagonal matrix framework successfully mitigates cochannel interference, improving signal decoding.
• Monte Carlo simulations validate the performance improvements in outage probability.

Conclusions:

• The integrated approach of UAVs, NOMA, K-means clustering, and a novel decoding framework significantly enhances WSN performance.
• The study demonstrates a practical and efficient solution for deploying and managing large-scale WSNs.
• The findings highlight the potential of machine learning and aerial platforms in future wireless communication systems.