Online Response-Time Analysis for Early Detection of Deadline Violations in 5G NR Applications

Abstract

Wireless communication via Fifth Generation (5G) New Radio (NR) networks plays a crucial role in supporting safety-critical systems, such as mobile industrial robots and autonomous connected vehicles. These systems depend on real-time information to respond rapidly to dynamically changing environments. To ensure reliability and timely data transmission, 5G NR introduces concepts such as network slicing and Ultra-Reliable Low-Latency Communication (uRLLC). A critical issue arises when the available channel resources are exceeded due to high data volume, degraded spectral efficiency, or both. Due to the dynamic nature of 5G NR, real-time guarantees cannot be determined statically at design time but must instead be evaluated continuously at runtime. In this work, we take a first step toward extending classical Response-Time Analysis (RTA) to dynamic 5G NR environments by proposing an online analysis framework for real-time applications with multiple concurrent data streams. The approach combines a safety-oriented timing model with a lightweight prediction of channel-quality degradation, allowing the computation of prediction intervals over which application deadlines remain guaranteed. Our results demonstrate that, under the evaluated assumptions, the proposed method can detect upcoming deadline violations sufficiently early to provide the timing margin required for a potential fallback reaction.