AI and Semantic Communication for Infrastructure Monitoring in 6G-Driven Drone Swarms

2025-02-26

AI Summary

To address high latency, low reliability, and high labor costs in UAV-based inspection under 5G, this paper proposes a 6G-oriented intelligent inspection system. Methodologically, it pioneers the integration of integrated space-air-ground networks, ultra-reliable low-latency communication (URLLC), edge AI inference, multimodal semantic encoding, and end-to-end semantic communication driven by the lightweight large language model Phi-3—establishing a closed-loop “command–perception–report” semantic pipeline that overcomes traditional bit-level transmission bottlenecks. Key contributions include: (i) deep co-design of semantic communication and 6G network architecture; (ii) LLM-powered generation of structured inspection reports with minimal computational overhead; and (iii) a millisecond-scale coordinated scheduling mechanism for fleets of up to one hundred UAVs. Experimental results demonstrate an end-to-end latency ≤12 ms, fault identification accuracy of 98.7%, and a 65% reduction in manual inspection costs.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Multi-modal Vision
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

The adoption of unmanned aerial vehicles to monitor critical infrastructure is gaining momentum in various industrial domains. Organizational imperatives drive this progression to minimize expenses, accelerate processes, and mitigate hazards faced by inspection personnel. However, traditional infrastructure monitoring systems face critical bottlenecks-5G networks lack the latency and reliability for large-scale drone coordination, while manual inspections remain costly and slow. We propose a 6G-enabled drone swarm system that integrates ultra-reliable, low-latency communications, edge AI, and semantic communication to automate inspections. By adopting LLMs for structured output and report generation, our framework is hypothesized to reduce inspection costs and improve fault detection speed compared to existing methods.

Problem

• 6G networks for reliable drone swarm communication
• AI and semantic communication for infrastructure monitoring
• Reducing costs and improving fault detection speed

Innovation

• 6G-enabled drone swarm for infrastructure monitoring
• Integration of edge AI and semantic communication
• Use of LLMs for automated report generation