Logo image
Energy-Efficient Resource Allocation for Satellite-UAV-Ground Integrated Compute First Networking
Journal article   Peer reviewed

Energy-Efficient Resource Allocation for Satellite-UAV-Ground Integrated Compute First Networking

Yue Zhang, Gaopan Hou, Jingmin Liu, Jie Feng, Penghao Chen, Chen Chen and Qingqi Pei
IEEE transactions on consumer electronics, Vol.71(4), pp.10986-11000
01/11/2025

Abstract

Autonomous aerial vehicles Computational modeling deep reinforcement learning Delays LEO satellite constellations Low earth orbit satellites Multi-UAV computational offloading Real-time systems Resource management Satellite constellations satellite-UAV integrated networks Satellites Servers Trajectory trajectory path planning
In multi-UAV networks with limited ground edge nodes, computational offloading poses significant challenges. The absence of a robust terrestrial network exacerbates these issues. To address this, we propose a framework based on satellite-UAV integrated systems, incorporating LEO satellite constellations. A two-layer UAV architecture is designed to categorize UAVs by function. Our study models communication dynamics and offloading processes, aiming to jointly optimize decisions, ratios, trajectories, and LEO frequency allocation. To tackle high-dimensional complexity and dynamic network conditions, we develop a multi-intelligent body deep deterministic policy gradient algorithm. Each task-generating UAV acts as an agent, adjusting its trajectory and offloading decisions in real-time. Simulations validate our framework, showing significant improvements over baseline approaches in minimizing delay, energy consumption, and computational costs.

Metrics

1 Record Views

Details

Logo image