Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Agentic AI-RAN Empowering Synergetic Sensing, Communication, Computing, and Control

Lingxiao Sun, Zhaoyang Zhang, Zihan Lin, Zirui Chen, Weijie Zhou, Zhaohui Yang, Tony Q. S. Quek

TL;DR

The paper addresses the challenge of executing closed-loop SC3 tasks on edge nodes for low-altitude wireless networks in 6G, where UAVs and aerial robots face strict latency and SWaP limits. It proposes Agentic AI-RAN, an architecture that combines MIG-based hardware isolation with containerized runtimes to run perception, reasoning, and control in a single edge node, guided by a cognitive Agentic Brain equipped with a CoT-based planner, contextual memory, and a unified tool interface. A case study with an indoor drone demonstrates that a 60/40 memory partition (communication/inference) yields stable operation, with total closed-loop latency in the 500–680 ms range and inference as the dominant latency contributor; OOM or instability occurs under poor partitioning. The work advances edge-native autonomy for embodied intelligence, IIoT, and urban monitoring by enabling tightly coupled SC3 loops with deterministic timing, offering a scalable pathway for mission-critical 6G edge networks.

Abstract

Future sixth-generation (6G) networks are expected to support low-altitude wireless networks (LAWNs), where unmanned aerial vehicles (UAVs) and aerial robots operate in highly dynamic three-dimensional environments under stringent latency, reliability, and autonomy requirements. In such scenarios, autonomous task execution at the network edge demands holistic coordination among sensing, communication, computing, and control (SC3) processes. Agentic Artificially Intelligent Radio Access Networks (Agentic AI-RAN) offer a promising paradigm by enabling the edge network to function as an autonomous decision-making entity for low-altitude agents with limited onboard resources. In this article, we propose and design a task-oriented Agentic AI-RAN architecture that enables SC3 task execution within a single edge node. This integrated design tackles the fundamental problem of coordinating heterogeneous workloads in resource-constrained edge environments. Furthermore, a representative low-altitude embodied intelligence system is prototyped based on a general-purpose Graphics Processing Unit (GPU) platform to demonstrate autonomous drone navigation in realistic settings. By leveraging the Multi-Instance GPU (MIG) partitioning technique and the containerized deployment, the demonstration system achieves physical resource isolation while supporting tightly coupled coordination between real-time communication and multimodal inference under a unified task framework. Experimental results demonstrate low closed-loop latency, robust bidirectional communication, and stable performance under dynamic runtime conditions, highlighting its viability for mission-critical low-altitude wireless networks in 6G.

Agentic AI-RAN Empowering Synergetic Sensing, Communication, Computing, and Control

TL;DR

The paper addresses the challenge of executing closed-loop SC3 tasks on edge nodes for low-altitude wireless networks in 6G, where UAVs and aerial robots face strict latency and SWaP limits. It proposes Agentic AI-RAN, an architecture that combines MIG-based hardware isolation with containerized runtimes to run perception, reasoning, and control in a single edge node, guided by a cognitive Agentic Brain equipped with a CoT-based planner, contextual memory, and a unified tool interface. A case study with an indoor drone demonstrates that a 60/40 memory partition (communication/inference) yields stable operation, with total closed-loop latency in the 500–680 ms range and inference as the dominant latency contributor; OOM or instability occurs under poor partitioning. The work advances edge-native autonomy for embodied intelligence, IIoT, and urban monitoring by enabling tightly coupled SC3 loops with deterministic timing, offering a scalable pathway for mission-critical 6G edge networks.

Abstract

Future sixth-generation (6G) networks are expected to support low-altitude wireless networks (LAWNs), where unmanned aerial vehicles (UAVs) and aerial robots operate in highly dynamic three-dimensional environments under stringent latency, reliability, and autonomy requirements. In such scenarios, autonomous task execution at the network edge demands holistic coordination among sensing, communication, computing, and control (SC3) processes. Agentic Artificially Intelligent Radio Access Networks (Agentic AI-RAN) offer a promising paradigm by enabling the edge network to function as an autonomous decision-making entity for low-altitude agents with limited onboard resources. In this article, we propose and design a task-oriented Agentic AI-RAN architecture that enables SC3 task execution within a single edge node. This integrated design tackles the fundamental problem of coordinating heterogeneous workloads in resource-constrained edge environments. Furthermore, a representative low-altitude embodied intelligence system is prototyped based on a general-purpose Graphics Processing Unit (GPU) platform to demonstrate autonomous drone navigation in realistic settings. By leveraging the Multi-Instance GPU (MIG) partitioning technique and the containerized deployment, the demonstration system achieves physical resource isolation while supporting tightly coupled coordination between real-time communication and multimodal inference under a unified task framework. Experimental results demonstrate low closed-loop latency, robust bidirectional communication, and stable performance under dynamic runtime conditions, highlighting its viability for mission-critical low-altitude wireless networks in 6G.
Paper Structure (20 sections, 5 figures, 1 table)

This paper contains 20 sections, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Typical Low-Altitude Applications of Agentic AI-RAN
  3. Embodied Intelligence
  4. Industrial Internet of Things
  5. Urban Monitoring
  6. Fundamental Challenges in Supporting Agentic AI-RAN
  7. Resource Contention from Task Heterogeneity
  8. Limited Support for Low-Latency Control
  9. Lack of Cross-Domain Scheduling
  10. Architecture Design for SC3 Integration in Low-Altitude Wireless Networks
  11. The Agentic Brain: Architecture and Reasoning Mechanism
  12. Vertical Integration of Partitioning and Containerization
  13. The Agentic Workflow: From Perception to Action
  14. Comparison with Other Related Methods
  15. A Case Study: Closed-Loop Task Execution
  16. ...and 5 more sections

Figures (5)

  • Figure 1: From fragmented processing to edge-native autonomous control.
  • Figure 2: Representative Agentic AI-RAN scenarios featuring SC3 functionality in low-altitude wireless networks.
  • Figure 3: SC3 task execution flow under MIG-based communication and inference partitioning.
  • Figure 4: End-to-end task execution of an autonomous drone in an indoor environment.
  • Figure 5: Experimental evaluation of closed-loop SC3 execution in the proposed Agentic AI-RAN system.