Towards 6G IoT: Tracing Mobile Sensor Nodes with Deep Learning Clustering in UAV Networks

Area of Science:

• Wireless communication networks
• Internet of Things (IoT)
• Robotics and autonomous systems

Background:

• Upcoming 6G networks require advanced solutions for device localization and data relay.
• Unmanned Aerial Vehicles (UAVs) are being explored as flying anchor nodes to support terrestrial IoT sensors.
• Accurate and efficient localization of mobile IoT devices presents a significant challenge.

Purpose of the Study:

• To develop and evaluate a novel algorithm for tracing mobile IoT devices using a swarm of UAVs.
• To enhance the speed and efficiency of localization by optimizing UAV movement and network clustering.
• To leverage deep learning for dynamic network partitioning and target acquisition.

Main Methods:

• Utilized Unmanned Aerial Vehicles (UAVs) equipped with Received Signal Strength Indicator (RSSI) sensors for target localization.
• Implemented a deep learning model based on Graph Convolutional Network (GCN) for dynamic clustering of UAVs.
• Employed a heuristic method for dynamic cluster number determination and optimized an RSSI loss function for partitioning.
• Developed an algorithm to retain effective clusters and return underperforming UAVs to base.

Main Results:

• The proposed deep learning-based UAV clustering algorithm demonstrated improved performance over deterministic approaches.
• Achieved significant reductions in the time required to reach the target mobile IoT device.
• Reduced the total distance covered by the UAVs during the tracing process.
• Validated through simulation experiments showcasing enhanced localization efficiency.

Conclusions:

• The integration of UAVs as flying anchor nodes with intelligent deep learning-based localization offers a promising solution for 6G IoT networks.
• The dynamic clustering algorithm effectively guides UAVs towards the target, optimizing resource utilization.
• This approach enhances the capability for rapid and efficient tracing of mobile IoT devices in complex environments.