Sunday 16 March 2025
The quest for efficient communication networks just got a boost, courtesy of researchers who have developed a novel network slicing framework designed specifically for advanced air mobility (AAM) systems.
As the demand for seamless connectivity in urban air mobility continues to grow, ensuring reliable and efficient communication between various airborne devices and ground stations has become increasingly important. The challenge lies in allocating limited resources, such as bandwidth, beam alignment, and computing capabilities, among multiple eVTOLs (electric vertical take-off and landing vehicles) that fly at varying speeds and altitudes.
To tackle this complex issue, researchers proposed a low-altitude intelligent network slicing framework that dynamically allocates heterogeneous resources based on real-time flight patterns and task requirements of eVTOLs. The framework consists of three key components: access pairing, resource pre- assessment, and slice orchestration.
The first component, access pairing, ensures optimal eVTOL-BS (base station) pairings by pre-scheduling the most suitable combinations of eVTOLs and BSs based on factors like flight trajectory, velocity, and altitude. This minimizes the risk of channel congestion and improves overall network efficiency.
Resource pre-assessment is the second critical component, which evaluates available resources at each BS before allocating them to eVTOLs. By predicting resource requirements and availability in real-time, this module avoids exhausting limited resources and ensures that eVTOLs receive the necessary bandwidth, computing capabilities, and beam alignment for optimal performance.
The third component, slice orchestration, leverages a multi-agent deep reinforcement learning (MADDPG) algorithm to optimize task offloading within a layered AAM system. MADDPG learns from its exploration phase to progressively optimize resource allocation, minimizing operation and violation costs while improving offloading efficiency.
Simulations demonstrated that the proposed framework significantly outperformed existing solutions in terms of resource utilization, user satisfaction, and cost savings. The results show that the intelligent network slicing framework can effectively allocate resources to meet the diverse needs of eVTOLs flying at varying speeds and altitudes.
This research has significant implications for urban air mobility systems, as it enables more efficient communication networks that can support the growing demand for seamless connectivity in AAM applications. By allocating limited resources more effectively, this framework can improve overall network performance, reduce costs, and enhance user satisfaction.
The development of this intelligent network slicing framework marks a major step towards realizing the full potential of urban air mobility systems.
Cite this article: “Intelligent Network Slicing Framework for Advanced Air Mobility Systems”, The Science Archive, 2025.
Airborne Devices, Network Slicing, Advanced Air Mobility, Evtols, Urban Air Mobility, Communication Networks, Resource Allocation, Real-Time Flight Patterns, Base Stations, Multi-Agent Deep Reinforcement Learning
