Abstract
In the next 50 years, the main transportation evolution will be the shift of traffic management from individual vehicles to centralized solutions. Such a centralized application could unlock the full potential of transport evolution across safety, speed, efficiency, and programmability.
Autonomous vehicles are a must as soon as possible — but because they act independently, they may not be the most efficient solution for future transportation. What if we could organize and coordinate them together?
The Challenge: Traffic Incidents
3,180 people died in traffic incidents on the roads in Germany in 2017. Another 390 thousand were injured, according to the Bundesanstalt für Straßenwesen (BASt). In the E28 countries, more than 25,000 people died in traffic incidents in recent years. The 390 thousand injured represent about 0.5% of Germany's population every year — statistically, one of every three persons closest to you will be injured in a traffic incident over a lifetime.
Where Is the Industry Today?
The Automotive Industry
The Automotive industry has implemented many passive and active safety solutions over recent decades. Advanced safety features such as blind-spot detection and collision warning appeared around 20 years ago. More recently, partially automated features — lane-keeping assists, adaptive cruise control, pedestrian recognition, automated emergency braking — have become standard. Safety features are evolving toward full automation.
Today, we are at SAE Levels 2–3 of driving automation. Most vehicles can trigger actions like emergency braking but require the driver to remain engaged. Some vehicles run in autopilot mode (like Tesla), and rare projects operate at SAE Level 4, such as driverless taxis in San Francisco (Waymo, 2023).
Telecommunication & IT Industries
Telecommunication and IT contribute significantly to traffic safety and efficiency. Standardization bodies (ETSI, IEEE, 3GPP) have defined standards enabling both short-range DSRC (<1km) and long-range V2X (>1km) communication. However, Service Providers haven't yet rolled out the physical networks required for ultra-low-latency, distributed application hosting, and edge computing at the scale needed.
The Case for Central Traffic Management
Autonomous vehicles have better reaction times than humans — machine drivers deliver shorter stopping distances (up to 10m shorter from 50 km/h in dry conditions). But autonomous vehicles are still individual participants. The critical question: could we combine their individual knowledge into a shared, common knowledge?
Success Criteria for Future Transportation
- Ultra-high level of safety — no more casualties on the roads
- Increase the speed of vehicles transporting people or goods from A to B
- Minimize the total sum of individual trip durations, not just each trip individually
- Shorter travelling time reduces road utilization and need for new infrastructure
- Reduce overall environmental impact through better traffic management
- Provide prioritizations for all vehicles — enabling programmable transportation
3-Step Evolution Path
The following three-step path describes the transition from individual autonomous vehicles to fully coordinated centralized control:
Autonomous Drive
Vehicles rely on onboard sensors and map/traffic information. Decision-making is entirely in the vehicle. Vehicles are not connected to each other.
Coordinated Vehicles
Some coordination among smaller groups (V2V). Vehicles connect directly or via mobile network. Use cases include platooning and coordinated lane changes. Decision still with individual vehicles.
Central Application Control
A central application manages all objects in a traffic situation — cars, trucks, trams, trains, pedestrians. Decision-making centralized; vehicles only override in emergency.
A centralized application provides not just the highest safety level, but unlocks further benefits: reducing travel time, increasing road capacity, and minimizing environmental impact. Connecting vehicles in a group of 5 requires 10 connections; a group of 100 requires ~4,950 — making a central topology far more scalable.
Traffic System Definition
A traffic system is the minimum traffic unit containing all objects that can physically interact with each other. Objects include moving items (cars, trucks, trams, trains, bicycles, pedestrians) and steady objects in the path of moving objects (parked cars, roadblocks, potholes).
Objects belong to the same traffic system if they can physically interact. The maximum number depends on speed: at 200 km/h, a vehicle needs up to 150 meters to stop, creating a significant interaction area. At 50 km/h (urban), a traffic system covers up to ~139 meters; at 250 km/h (Autobahn), up to ~347 meters (per 3GPP TR 22.886).
Application High-Level Design
The central traffic system management application has four main functional components:
Traffic Management Policies
Creates, stores, and enforces hierarchical traffic policies — from universal speed limits to temporary local amendments and optimization opportunities. Enables traffic programmability.
Active/Passive Object Inventory
Stores all known active objects (vehicles, traffic signals) and passive objects (potholes, obstacles). Active objects share sensor data and can be remotely controlled.
Real-Time Instance Inventory
Manages live traffic system instances, their geographic coverage, and communication between parallel instances. Handles domino-effect propagation across regions.
Continuous Situation Engine
Collects and processes data, continuously assesses traffic situations against policies, detects violations, calculates optimal corrective actions, and communicates change orders to vehicles.
Traffic System Management Application Instance
One instance per geographically dependent set of active/passive objects. Handles connectivity & identity management, data processing, situation assessment, decision-making, and acknowledgement management.
Required Network Capabilities (3GPP KPIs)
Network Development Priorities
Service Providers must take the following actions to support future transportation use cases:
- Connect mobile base stations to the application location with meshed architecture, ideally directly or via ring topology
- Replace existing physical and logical topologies with more meshed topology to enable low latency and high mobility
- Enable application hosting at the access, backhaul, and backbone networks — roll out virtual infrastructure from backhaul router locations toward core sites
- Enable fast application session handovers among neighbouring application locations
- Implement mobile network resource orchestration for dynamic resource allocation per application requirements
- Enable high upload speed up to 50–60 Mbps everywhere
In the Upcoming Decades
There may be 50–100 years until total safety and fully automated transportation becomes a reality. The probable achievements for the next ten years include: full autonomous driving vehicles becoming a reality; continued automotive safety feature rollout; evolution of connectivity options; and initial implementation of central traffic management use cases that go beyond individual autonomous driving.
The efficiency level of central traffic management depends on the ratio of managed vehicles — a dependency that is more exponential than linear. Technology adoption is therefore critical. Society's acceptance will likely begin with services where people already use external providers: taxis, public transport, car sharing.
Download Full PDF