TL;DR:
- V2X (vehicle-to-everything) lets vehicles communicate with each other (V2V), with infrastructure (V2I), with pedestrians (V2P), and with networks (V2N) to share position, speed, and hazard data in real time
- C-V2X (cellular V2X) uses 5G New Radio sidelink for direct device-to-device communication without going through a base station, achieving 5-10ms latency suitable for safety-critical applications
- Roadside units (RSUs) with embedded edge compute are the deployment model — processing locally rather than round-tripping to the cloud keeps latency deterministic even when cellular coverage varies
- The EU Intelligent Transport Systems Directive now mandates C-V2X capability in new road infrastructure projects, accelerating European deployment; US adoption is fragmented between DSRC and C-V2X standards
The core problem V2X is solving is reaction time. A driver’s visual reaction time to a hazard is roughly 1.5 seconds. That’s enough time to stop at 30mph but not at motorway speeds. A vehicle that knows about a hazard 300 metres ahead — because another vehicle or a roadside sensor broadcast it — can begin braking or alerting the driver before the hazard is visible. That’s the value proposition, and it requires the data to arrive in under 100ms, ideally much less.
Cloud-based solutions can’t reliably deliver that. A round-trip from a vehicle to a cloud data center and back adds 50-200ms under normal conditions — and degrades unpredictably under cellular congestion. Edge compute at the roadside, combined with direct V2V communication for the shortest distances, is how the industry is solving the latency constraint.
The V2X Communication Stack
V2X encompasses several distinct communication types that operate at different ranges and for different purposes:
V2V (Vehicle to Vehicle) — direct communication between vehicles in proximity. Uses 5.9GHz DSRC (Dedicated Short Range Communications) or C-V2X PC5 (direct sidelink) radio. Range is typically 300-1000 metres depending on conditions. Applications: collision avoidance, emergency braking alerts, lane change assistance, platooning coordination.
V2I (Vehicle to Infrastructure) — communication between vehicles and roadside equipment: traffic signals, toll gantries, road sensors, variable message signs. Infrastructure-side compute at RSUs processes vehicle data and pushes back signal phase and timing (SPaT) information, roadwork warnings, speed advice.
V2P (Vehicle to Pedestrian) — alerts between vehicles and pedestrians or cyclists with V2X-capable devices. Still early stage in deployment but included in current standards. Primarily implemented via smartphone apps that broadcast position using the same protocols as vehicles.
V2N (Vehicle to Network) — communication between vehicles and cloud/edge services via cellular networks. Used for non-safety-critical applications: HD map updates, traffic flow data, software updates. Higher latency is acceptable because these aren’t time-sensitive.
C-V2X vs DSRC
The V2X standard has been contested. The original approach — DSRC/WAVE (Wireless Access in Vehicular Environments), standardised as IEEE 802.11p — operates on a dedicated 5.9GHz spectrum band and provides direct communication between vehicles and infrastructure without cellular infrastructure. It’s been in deployment in the US since the mid-2010s, particularly for truck platooning and US DOT pilot programmes.
C-V2X (cellular V2X), standardised in 3GPP Release 14 and enhanced in Release 16, offers two modes:
- PC5 mode (sidelink): Direct device-to-device communication on the 5.9GHz band, similar to DSRC but using LTE/5G radio technology. Works without network coverage.
- Uu mode (cellular uplink): Communication via the cellular network to RSUs or cloud infrastructure. Requires network coverage but supports wider-range communication.
The key advantage of C-V2X PC5 over DSRC is that it’s part of the 5G ecosystem — it benefits from 5G chipset scale, integrates with cellular network management, and provides a path to network-assisted V2X that DSRC doesn’t offer. Most new automotive programmes are standardising on C-V2X, and the EU’s ITS Directive mandates C-V2X as the primary technology for new infrastructure deployments.
The RSU Architecture
Roadside units are where the edge computing happens in V2X deployments. An RSU is a weatherproof compute node installed at intersections, motorway gantries, and other critical points. It combines:
- V2X radio (C-V2X and/or DSRC) for direct communication with vehicles in range
- 5G cellular uplink for connectivity to traffic management centres and cloud infrastructure
- Edge compute for local processing: sensor fusion, SPaT generation, event detection, local caching of HD map data
- Sensor integration — RSUs can aggregate data from roadside cameras, LIDAR, radar, and inductive loop detectors, fusing them into a local environment model
The processing that happens at the RSU edge:
Signal Phase and Timing (SPaT): The RSU reads the traffic controller signal state and broadcasts it to approaching vehicles over V2X. Vehicles can display “green wave” speed advice: drive at X km/h to catch the next green. This reduces stop-start driving and fuel consumption while improving flow.
Collective perception: RSUs receive individual vehicle position and sensor data and build a shared environment model. A vehicle blocked from seeing a cyclist by a bus can receive that cyclist’s position from the RSU, which triangulated it from multiple other vehicle feeds.
Hazard warning aggregation: When one vehicle broadcasts an emergency brake event, the RSU amplifies it to all approaching vehicles in range, including those that aren’t within direct V2V range of the original vehicle.
Local caching: RSUs cache current HD map segments, weather data, and construction zone information for their geographic coverage area. Vehicles in range receive updates without needing cellular connectivity.
Latency Numbers
The latency requirements for V2X safety applications are stringent:
| Application | Max End-to-End Latency |
|---|---|
| Intersection collision warning | 20ms |
| Emergency vehicle alert | 50ms |
| Road hazard warning | 100ms |
| Traffic information | 500ms |
| HD map update | No hard requirement |
C-V2X PC5 achieves 3-5ms for direct vehicle-to-vehicle communication. V2I communication via RSU edge processing adds the RSU processing time (typically 1-5ms) plus the radio propagation, giving 5-15ms end-to-end for safety-critical alerts. Network-assisted V2X adds the RSU-to-cloud round-trip, which varies significantly by cellular network architecture but typically runs 20-50ms for edge-deployed multi-access edge computing nodes at base stations.
The 20ms requirement for intersection collision warning is achievable with PC5 direct communication and local RSU processing. It’s not reliably achievable with cloud round-trips.
Current Deployments
Europe: The EU ITS Directive (2023/1462) mandated that all new and significantly upgraded major road infrastructure include C-V2X capability from 2025. Germany’s Autobahn programme has deployed RSUs at 500+ locations with SPaT data and hazard warning. The Netherlands’ A2/A4/A10 smart motorway corridors are fully C-V2X enabled. UK deployment is behind EU-mandated timelines (Brexit removed the regulatory driver) but the National Highways V2X programme is deploying along the M1 and M25.
United States: Fragmented between DSRC deployments (New York, Tampa, and various DOT pilots from the 2010s programmes) and newer C-V2X deployments. Michigan’s US-23 corridor is a notable C-V2X deployment with real vehicle data. Federal progress has been slow, with spectrum policy debates over 5.9GHz allocation delaying commercial rollout.
China: The most aggressive deployment globally. China’s “Vehicle-Road Cloud Integration” national programme mandates C-V2X in new road construction and has deployed at scale in multiple cities. Shanghai, Chongqing, and Wuhan have large-scale deployments with active vehicle fleets participating.
What’s Needed for Full Deployment
The infrastructure is only half the problem. V2X requires vehicles with compatible radio hardware. The penetration rate of C-V2X-capable vehicles in the UK fleet is currently low — the technology has been mandated for EU new vehicles from 2024 but the existing fleet will take a decade to cycle through.
This creates a chicken-and-egg problem that V2X advocates have struggled with for years. Infrastructure won’t be built at scale until there are enough equipped vehicles to justify it. Vehicles won’t be equipped until there’s enough infrastructure to use. The regulatory mandate approach being used in the EU is the most direct solution: mandate infrastructure first, then mandate vehicle equipment on a timed schedule.
For applications that don’t require dense vehicle penetration — SPaT delivery, hazard alerts from infrastructure sensors, HD map distribution — the value is available now with current vehicle penetration rates. The collective perception and V2V safety applications that require many equipped vehicles nearby are later-stage benefits that emerge as fleet penetration grows.
The infrastructure being deployed now will enable those applications when the vehicles are there. Building it correctly — with edge compute at RSUs, proper C-V2X radio, and cellular backhaul — is what makes that possible.