Return
Optimizing Energy Consumption for IoV in Remote Areas via Space–Air–Ground Integrated Networks: A DRL-Based Wireless Power Transfer Strategy
Y
张
X
H
DOI:10.1109/JSAC.2026.3705346.png)
Abstract
En 中文
With the rapid expansion of connected vehicles, Internet of Vehicles (IoV) technology has enhanced driving safety and travel experiences. However, the implementation of this technology heavily relies on robust infrastructure, yet faces challenges in remote areas due to insufficient ground communication coverage and limited vehicle range. To address this, the space-air-ground caching Wireless Power Transfer (WPT) model has been constructed, providing an integrated caching-communication-power supply solution for energy-deprived remote areas. The satellite transmits radio energy to the Unmanned Aerial Vehicle (UAV) via radio frequency and schedules cloud caching. The UAV locally caches high-demand files and serves vehicles through a terahertz link, while vehicles can harvest energy from multiple sources. To reduce frequent requests for satellite remote resources by vehicle users, we propose a Feature-Enhanced Hybrid Recommendation (FEHR) algorithm. This approach integrates user profiles with area scene characteristics, leveraging UAVs to recommend personalized cached files tailored to vehicle needs, thereby minimizing satellite backhaul energy consumption. An Energy-Aware UAV Trajectory Optimization (EATO) algorithm has been proposed to reduce the high flight energy consumption of UAVs caused by remote terrain. Simultaneously, our proposed Decoupled Intelligent Wireless Power Transfer (DIWPT) algorithm based on DRL resolves the coupling problem between charging and communication within the space-air-ground framework through timeslot allocation and power splitting. Simulation results show that the proposed solution effectively lowers system energy consumption while ensuring a satisfactory user experience for vehicles, and provides a reliable technical foundation for communication and power supply in remote areas emergency scenarios.
Keywords:
Internet of Vehicles
wireless power transfer
content caching and recommendation
deep reinforcement learning
energy optimization
Journal
IF:
17.2
Papers:
6.4K
Citations:
3.1W
