5G-Enabled Teleoperated Driving: An Experimental Evaluation

Testouri, Mehdi; Elghazaly, Gamal; Hawlader, Faisal; Frank, Raphaël · 2025 · OpenAlex-citations

DOI: 10.1109/mt-its68460.2025.11223534

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 an experimental evaluation of teleoperated driving over a commercial non-standalone (NSA) 5G network, addressing the critical need for real-world validation of remote human intervention in autonomous vehicles. While teleoperated driving offers a solution for complex scenarios where fully autonomous systems fail, its effectiveness relies heavily on ultra-low latency and high-reliability communication. The authors aim to bridge the gap between simulation-based studies and real-world deployments by assessing the feasibility, performance metrics, and safety limitations of 5G-enabled teleoperation. The study utilizes a real-world testbed consisting of a modified Kia Soul EV equipped with a drive-by-wire system, an onboard computer running the RoboCar autonomous driving software, and a 5G smartphone modem. The remote operator controls the vehicle via a high-fidelity simulator featuring a curved monitor, force-feedback steering wheel, and motion platform. The system architecture employs a VPN for secure connectivity, with video streams transmitted via RTSP/H.264 and control commands exchanged using JSON over ZeroMQ sockets. Experiments were conducted in Luxembourg City, measuring key performance indicators including glass-to-glass (G2G) latency, round-trip time (RTT), network jitter, packet loss, and steering command delay. The results indicate that while the network connection was stable with zero packet loss and low jitter (0.709ms), G2G latency remains a significant bottleneck. The average G2G latency was measured at 202.41ms (standard deviation 31.56ms), with values ranging from 150ms to over 300ms. This latency is attributed primarily to video encoding and transmission rather than network transmission alone. In contrast, the RTT for control commands was significantly lower, averaging 46.63ms (standard deviation 10.05ms). Steering control analysis revealed that the vehicle response was generally accurate, though delays up to 750ms were observed due to the combined effects of network latency and steering system dynamics. The authors conclude that the achieved G2G latency restricts safe teleoperated driving to low speeds (below 20 km/h), whereas the lower RTT allows for safe operation at moderate to high speeds. The significance of this work lies in its empirical demonstration that commercial NSA 5G networks can support teleoperated driving, but current video streaming pipelines introduce unacceptable latency for high-speed operation. The findings highlight that reducing G2G latency requires improvements in video encoding and streaming strategies rather than just network upgrades. The paper recommends future research focus on adaptive network optimization, haptic feedback, and predictive control mechanisms to mitigate latency effects and enhance the reliability of teleoperated systems for broader autonomous mobility applications.

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-prismaquant on 2026-06-26; verification: verified.

Topics

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

• teleoperation remote driving