We have looked at several approaches to improving millimetre-wave (mmWave) connectivity on this blog, including transparent antennas, metasurfaces and, more recently, KDDI and JR East's demonstration of bringing 28 GHz signals inside railway carriages. While these technologies address challenges such as signal penetration and coverage, another important problem is maintaining stable connectivity when vehicles are moving between different antenna coverage areas. This is something that NTT DOCOMO, NEC and NTT have been investigating through their work on distributed MIMO (Multiple-Input Multiple-Output) technology.
In May 2026, NTT DOCOMO and NEC, in collaboration with NTT, announced the successful demonstration of a technique that allows multiple vehicles travelling at speed to maintain more stable, high-capacity wireless communications using the 40 GHz mmWave band. The work builds on their earlier research and is being positioned as an important step towards future 6G applications, particularly connected vehicles and railways.
The challenge with mmWave is not just its relatively limited coverage range or sensitivity to obstructions. When vehicles travel at speed, the radio environment changes continuously. Signals may become blocked, different antennas may offer better coverage at different locations, and switching between them can introduce sudden changes in the frequency and timing of received signals.
One possible solution is distributed MIMO, where multiple antennas are deployed at different locations but coordinated through a common base station. Instead of depending on a single antenna site, a moving device can communicate through different distributed antennas as it travels through an area.
However, this introduces another challenge. As a vehicle moves between antennas, differences in its relative movement and the distance travelled by radio signals can cause changes in Doppler frequency and propagation delay. The Doppler effect is the same principle that makes the pitch of a passing ambulance siren appear to change, although in wireless communications it affects the frequency of the received radio signal.
These sudden frequency and timing changes can temporarily degrade the quality of the connection, even when sufficient radio coverage is available.
To address this, NEC and its partners developed an approach that compensates for these changes at the base station before signals are transmitted, rather than relying entirely on the receiving device to correct them afterwards.
The technique uses reference signals transmitted by each vehicle to estimate the appropriate frequency and timing adjustments for each distributed antenna. The base station then applies these corrections separately to the signals intended for each vehicle before combining them for transmission. This helps reduce abrupt frequency and timing changes experienced by the receivers as the serving antennas change.
Importantly, the technique is designed to support multiple vehicles simultaneously, with different adjustments applied according to their individual movements and radio conditions.
The companies had previously demonstrated related distributed MIMO techniques in March 2025. That earlier work included faster antenna and beam selection, as well as frequency and timing pre-compensation for a single moving vehicle. The latter technique was tested at speeds of up to 100 km/h.
The latest demonstration builds on this foundation by extending the pre-compensation approach to multiple moving vehicles, creating a more challenging situation because each vehicle may experience different frequency shifts and propagation delays.
The new trials were conducted on 26–27 March 2026 at a full-scale tunnel test facility belonging to Japan's National Institute for Land and Infrastructure Management (NILIM). Three distributed antennas connected to a base station were installed approximately 150 metres apart along one side of a road. Two vehicles equipped with mobile terminals travelled at 60 km/h in opposite directions.
A tunnel is an interesting environment for this type of experiment because radio signals can reflect from surrounding surfaces, creating a more complex propagation environment. As the vehicles move, the system also needs to deal with frequent switching between the distributed antennas.
The researchers compared a conventional approach, where the receiving terminals compensate for changes in frequency and timing, with the proposed technique that additionally performs per-vehicle pre-compensation at the base station. The difference was significant.
With the conventional approach, the combined downlink throughput of the two vehicles sometimes dropped from around 550 Mbps to approximately 110 Mbps during antenna switching. Average throughput over the 30-second measurement period was approximately 430 Mbps.
With the new technique, these sudden reductions were substantially suppressed. Combined throughput remained at or above 380 Mbps during the reported test, while average throughput increased to approximately 560 Mbps, an improvement of around 30%.
Perhaps even more interesting than the improvement in average throughput was the change in fifth-percentile throughput. This measures performance towards the lower end of the observed throughput distribution, helping to show how the system behaves during less favourable conditions.
The fifth-percentile combined throughput increased from 270 Mbps with the conventional approach to 480 Mbps with the proposed technique, an improvement of approximately 80%.
This matters because a network delivering very high speeds most of the time, but occasionally suffering severe performance drops, may be less useful for some applications than one delivering slightly lower but more consistent performance. The distinction becomes particularly important for moving vehicles, where radio conditions can change rapidly.
It is worth highlighting that these throughput figures represent the combined performance of both vehicles, rather than the individual data rate available to each one. The results also demonstrate improvements in throughput stability, rather than independently establishing performance for other requirements such as ultra-low latency or mission-critical reliability.
NEC has published a short video explaining the technology and showing the experimental setup and results.
Although the companies describe this as technology for the 6G era, the underlying problem is much broader. Maintaining high-capacity, reliable wireless connectivity for moving vehicles is already an important challenge for existing mobile networks.
Railways are an obvious potential application. As discussed in our earlier post on improving 5G connectivity on Deutsche Bahn, providing good connectivity along railway corridors involves several challenges, including continuous trackside coverage, signal penetration into carriages and maintaining connections as trains move between coverage areas.
Distributed MIMO combined with improved antenna coordination and signal compensation could eventually help address some of these challenges, particularly where very high wireless capacity is required.
There are also possibilities for connected and autonomous vehicles, where large amounts of information may need to be exchanged with roadside infrastructure. Potential applications highlighted by the companies include sharing sensor data for cooperative driving, immersive in-vehicle entertainment, real-time translation and AI-assisted guidance services.
However, the latest demonstration should still be viewed as an experimental milestone rather than a deployment-ready solution. The trials involved just two vehicles travelling at 60 km/h inside a controlled tunnel test facility. Further testing will be necessary to understand how the technology performs with more vehicles, at higher speeds and across different real-world environments.
There are also practical considerations around the cost and complexity of deploying distributed antennas, the infrastructure required to connect them to base stations, and the coordination needed to support multiple moving users. Higher-frequency coverage can require relatively dense infrastructure, which makes deployment economics particularly important.
NTT DOCOMO, NEC and NTT plan to extend their trials to environments including high-speed railways, conventional railways and major roads.
Japan has been particularly active in investigating ways to make mmWave more practical, something we have previously covered in our report on XGMF's conference to advance millimetre-wave technology. This latest research adds another dimension to those efforts by addressing mobility and throughput consistency rather than simply extending coverage or increasing peak data rates.
What makes this demonstration interesting is the shift in emphasis from achieving the highest possible throughput to maintaining good performance under challenging radio conditions. For future transport and mobility applications, the ability to deliver a consistent wireless connection may ultimately prove just as important as headline data rates.
- Connectivity Technology Blog: KDDI and JR East Bring 5G mmWave Inside Train Cars Using Glass Antennas
- Connectivity Technology Blog: Seamless 5G Connectivity Across Germany on Deutsche Bahn
- Connectivity Technology Blog: Highlights from XGMF's Conference to Advance Millimetre Wave Technology
Comments
Post a Comment