LYRA: Label-Free Structural Synchronization and Resource Allocation for UAV Edge Networks

2026-08-07

AI Summary

This work addresses the challenge of jointly optimizing visual model updates, semantic fidelity, and resource constraints in unmanned aerial vehicle (UAV) edge networks under label-free conditions. To this end, the authors propose a unified framework integrating model update scheduling and resource allocation, featuring three key innovations: an original unsupervised semantic distortion rating (OSDR) mechanism, a sensitivity-aware structural synchronization strategy (SASS), and a Lyapunov-guided discrete reinforcement learning algorithm that transforms long-term energy constraints into virtual queue stability problems to reduce decision complexity. Experimental results on real-world traffic trajectory data demonstrate that the proposed approach reduces risk backlog by up to 33.3% compared to baseline methods, while significantly improving semantic recovery efficiency and update triggering accuracy—all within strict adherence to long-term energy budgets.

Abstract

While deploying hierarchical vision models to process mission-critical tasks, UAV edge systems must adaptively update the models to sustain inference reliability under low-level environmental corruption. However, existing work has overlooked the optimal timing for model updates, the impracticality of relying on real-time expert labels, and the significant bandwidth and energy constraints of UAVs. This paper proposes a joint model update scheduling and resource allocation framework, aiming to maximize long-term semantic fidelity and resource efficiency of UAV edge intelligence systems. To address the challenge of label-free semantic evaluation, we formulate the Online Semantic Disagreement Rate (OSDR) as a proxy for timely update triggering, thereby enabling fine-grained Sensitivity-Aware Structural Synchronization (SASS). Furthermore, to overcome the curse of dimensionality in hybrid action spaces and effectively bound long-term energy budgets, we propose a Lyapunov-guided discrete reinforcement learning algorithm that performs action space dimensionality reduction and transforms constraints into virtual queue stability problems. The reported experimental results, based on real traffic traces, demonstrate that the proposed framework consistently outperforms representative baselines in semantic recovery efficiency and update triggering precision, by satisfying long-term energy budget and by reducing average risk backlog by up to 33.3\% in the dynamic environmental corruption scenario.

Problem

• UAV edge networks
• label-free
• model update scheduling
• resource allocation
• semantic fidelity

Innovation

• Label-Free Learning
• Structural Synchronization
• Lyapunov Optimization
• Reinforcement Learning
• UAV Edge Intelligence