Autonomous Wireless Networks for Smart Factories: An Experimental Study of Latency, Reliability, and Self-Healing Performance in Industrial Environments

Abstract

In this paper, a detailed experimental study of autonomous wireless network technologies deployed in a live smart factory setup at an automotive component manufacturing facility in Nagpur, India, is presented. A coexisting multi-technology wireless infrastructure, including a 5G NR private network, Wi-Fi 6E access points, WirelessHART sensor mesh, and Bluetooth Low Energy (BLE 5.3) asset tracking, was deployed and monitored over a 45-day continuous production period across a 4,800 m2 factory floor hosting a total of 337 wireless devices (comprising 48 5G UEs, 2 gNBs, 112 Wi-Fi STAs, 5 APs, 124 WirelessHART sensor nodes, 2 gateways, 28 BLE tags, 4 BLE anchors, 12 UWB anchors). We experimentally characterized key performance metrics, including end-to-end latency distributions, packet delivery ratio (PDR) vs. signal-to-noise ratio (SNR), aggregate throughput degradation in high node density, autonomous AP handover and self-healing response under simulated link failures, RF coverage distribution under factory EMI conditions, and node energy consumption under three radio duty-cycling strategies. The proposed reinforcement learning (RL)-based adaptive duty-cycling strategy reduced average node power consumption by 76.7 % compared to an always-on baseline, while maintaining PDR above 99.95 %. 5G NR URLLC links achieved a median end-to-end latency of 1.8 ms with a 99.9th-percentile latency of 4.3 ms, satisfying IEC 61784 industrial control requirements under the measured backhaul and EMI conditions at this deployment site. The self-healing handover was completed within 1.2 seconds of primary link failure detection. To the best of our knowledge, this constitutes the first experimental validation of 5G NR URLLC performance in an operational Indian automotive factory environment under measured industrial EMI conditions. The results offer a validated performance baseline for the deployment of autonomous wireless networks in Industry 4.0 smart factory scenarios.