Skip to content
Conference

LLM-assisted edge-coordinated scheduling and fault recovery for roadside IoT systems: a field deployment study in urban intersection clusters

Oct 2026 · 3rd International Conference on Electronic Materials and Information Engineering (EMIE 2023) · Vol 14382, pp. 143822H - 143822H-6 · 0 citations · 9 references
Engineering

TL;DR

Results show that the proposed LLM-assisted edge-coordinated operation framework improves latency, reporting reliability, and recovery efficiency within the verified roadside IoT deployment boundary, while safety-critical signalphase changes remain outside automatic execution.

Abstract

Urban intersection clusters place roadside IoT systems under a mix of heterogeneous sensing, unstable communication links, peak-hour bursts, and slow manual fault recovery. This study develops an engineering-oriented cloud-edge-end system that combines edge state summarization, cache-assisted task scheduling, multi-gateway load balancing, tiered fault recovery, and a retrieval-augmented large language model (LLM) advisory module guarded by rule-based safety validation. The system was evaluated in a continuous field deployment from 2026-03-09 to 2026-06-08 across 12 intersections and 149 roadside sensing, communication, computing, and execution nodes. The anonymized paper data package contains 326,999 structured measurement records, 90 fault/stress events, 45 LLM recommendation records, and 30 safety-check records. Compared with the cloud fixed-rule baseline, the full semi-automatic non-safety configuration reduces mean network latency from 56.2 ms to 33.8 ms and P95 latency from 143.4 ms to 83.6 ms, increases the data reporting success rate from 95.39% to 98.30%, and reduces mean time to recovery (MTTR) from 117.4 s to 52.8 s. Across 45 LLM recommendations, expert consistency reaches 88.9% and mean generation latency is 1.76 s; safety validation blocks 66.7% of high-risk suggestions in the checked human-in-the-loop and semi-automatic modes. These results show that the proposed LLM-assisted edge-coordinated operation framework improves latency, reporting reliability, and recovery efficiency within the verified roadside IoT deployment boundary, while safety-critical signalphase changes remain outside automatic execution.

View source

Similar papers

Conference Oct 2026

LHPA-ECR: link-health and priority-aware edge collaborative recovery for industrial IoT measurement-control networks

Continuous field data are a basic operating condition for Industrial Internet of Things (IIoT) measurement-control networks. In deployed systems, however, packet loss, latency spikes, gateway outages, node interruptions, duplicated messages, timestamp skew, and edge overload can interrupt the path from field measuremen...

Bo-Hai Tang, Xin Du, Dan-Dan Gu et al. · 0 citations
Open access Oct 2026

Dynamic Workload Offloading Framework for IoT-Edge-Cloud Architectures in Smart Cities

Smart city infrastructures generate massive streams of time-sensitive data from distributed Internet of Things (IoT) devices, ranging from traffic management sensors and autonomous transit to public safety surveillance. Processing these workloads exclusively in centralized cloud data centers leads to unacceptably high...

Nimish Arvind · 0 citations
Open access Oct 2026

Machine learning-driven SDN for reliable emergency response in industrial IoT networks

Industrial Internet of Things (IIoT) applications operating in safety-critical environments—such as chemical plants, oil refineries, and disaster response systems—demand ultra-low latency, deterministic communication, and absolute message reliability, particularly during emergency scenarios. Traditional Software-Define...

Q. Ali, Sohaib R. Awad · 0 citations
#edge computing Open access Nov 2026

An IoT based three-tier fog computing architecture for policy-aware Smart Waste Management using fog and edge

SWMS is presented, a three-tier edge/fog/cloud system in which smart bins sense fill level, methane, temperature, and weight, district fog nodes evaluate thresholds and compute routes locally, and a cloud tier runs graduated compliance enforcement and an Eco-Score layer.

A. Prakash, M. Sambath, A. Subramanian et al. · 0 citations
Open access Oct 2026

A Queueing-Predictive Task Offloading Framework for Joint Latency and Energy Optimization in Smart City Edge–Cloud IoT Networks

Smart-city Internet of Things (IoT) platforms run safety-critical alarms, near-real-time video analytics and delay-tolerant batch jobs on a shared edge–cloud infrastructure. Keeping enough edge capacity powered for peak demand wastes energy for most of the day, whereas aggressive consolidation pushes work to distant cl...

Uma Bennett, Serena Zhou, D. Jackson · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.