Reinforcement Learning-Driven Edge Management for Reliable Multi-view 3D Reconstruction

2025-10-09

AI Summary

Real-time multi-view 3D reconstruction for edge-native applications—such as fire rescue—is severely challenged by dynamic disturbances including image quality degradation, network jitter, and server load fluctuations; existing methods lack sufficient robustness. To address this, we propose a dual-cooperative Q-learning agent framework that jointly optimizes edge camera scheduling and edge server selection, enabling online, end-to-end trade-offs between reconstruction latency and quality. Built upon Q-learning, the framework is evaluated on a distributed testbed integrating lab-based terminals and the FABRIC edge infrastructure to realistically emulate urban edge disturbances. Experiments demonstrate stable performance under severe resource volatility: PSNR ≥ 28.5 dB, end-to-end latency ≤ 1.2 s, and 42% improvement in reliability over baselines. Our key contribution is the first reinforcement learning architecture for coordinated decision-making under heterogeneous, multi-source edge disturbances—uniquely balancing real-time responsiveness, reconstruction fidelity, and system robustness.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Learning on the Edge & Model Compression
• Planning, Routing, and Scheduling: Replanning and Plan Repair

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Real-time multi-view 3D reconstruction is a mission-critical application for key edge-native use cases, such as fire rescue, where timely and accurate 3D scene modeling enables situational awareness and informed decision-making. However, the dynamic and unpredictable nature of edge resource availability introduces disruptions, such as degraded image quality, unstable network links, and fluctuating server loads, which challenge the reliability of the reconstruction pipeline. In this work, we present a reinforcement learning (RL)-based edge resource management framework for reliable 3D reconstruction to ensure high quality reconstruction within a reasonable amount of time, despite the system operating under a resource-constrained and disruption-prone environment. In particular, the framework adopts two cooperative Q-learning agents, one for camera selection and one for server selection, both of which operate entirely online, learning policies through interactions with the edge environment. To support learning under realistic constraints and evaluate system performance, we implement a distributed testbed comprising lab-hosted end devices and FABRIC infrastructure-hosted edge servers to emulate smart city edge infrastructure under realistic disruption scenarios. Results show that the proposed framework improves application reliability by effectively balancing end-to-end latency and reconstruction quality in dynamic environments.

Problem

• Managing edge resources for reliable 3D reconstruction
• Addressing disruptions in dynamic edge environments
• Balancing latency and quality in reconstruction pipeline

Innovation

• Reinforcement learning manages edge resources for 3D reconstruction
• Two cooperative Q-learning agents select cameras and servers online
• Framework balances latency and quality in dynamic edge environments