Demo: A Hands-on Teleoperated Driving over 5G

Testouri, Mehdi; Elghazaly, Gamal; Hawlader, Faisal; Frank, Raphael · 2025 · Crossref

DOI: 10.1109/vnc64509.2025.11054256

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper presents a demonstration of a 5G-enabled teleoperated driving system designed to bridge the capabilities of autonomous vehicles in complex or unpredictable scenarios, such as construction zones or adverse weather. The primary motivation is to enable remote human intervention without requiring a safety driver, thereby enhancing the safety and accessibility of autonomous mobility. The authors highlight that while teleoperation alleviates limitations of fully autonomous systems, it is critically dependent on network performance, specifically requiring ultra-low latency and high reliability. The study aims to showcase the potential and current shortcomings of using commercial 5G networks for this application through a real-world testbed in Luxembourg. The system architecture consists of a modified KIA Soul EV equipped with a drive-by-wire system for steering, acceleration, and braking, controlled via a Polysync DriveKit. The vehicle runs the RoboCar autonomous driving software and utilizes a Trimble BX992 GNSS-INS for localization and an Insta360 X3 camera for video feedback. The remote vehicle connects to a VPN via a 5G smartphone tether. The teleoperation interface features a high-fidelity setup including a 49-inch ultra-wide monitor, a force-feedback racing wheel, pedals, and a motion platform to simulate driving forces. Control commands and telemetry are transmitted via ZeroMQ TCP sockets at 100Hz, while video is streamed at 30 FPS using H.264 compression. The authors conducted an initial evaluation of the system’s performance, measuring round-trip time (RTT), glass-to-glass (G2G) latency, and steering delay. Over 1,000 samples, the mean RTT was 46.63ms with a standard deviation of 10.05ms. However, the mean G2G latency, measured via video analysis, was significantly higher at 202.41ms with a standard deviation of 31.56ms. Additionally, steering response latency reached up to 750ms due to vehicle actuation delays. Based on these metrics, the authors determined that safe teleoperation is feasible at speeds below 20km/h. At 30km/h, the vehicle would travel approximately 0.2 meters during the latency period, posing safety risks. The demonstration confirms that 5G networks can support teleoperated driving at low speeds, providing a viable solution for handling edge cases in autonomous driving. However, the significant G2G and actuation latencies highlight current limitations for higher-speed operation. The authors conclude that future work must focus on adaptive network optimization, improved streaming strategies, and predictive control mechanisms to reduce latency and enhance the reliability and scalability of teleoperated driving systems for broader deployment.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• teleoperation remote driving

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Methodological Resource: tool software