Rethinking Urban Traffic with a Global Scheduling Model for Autonomous Vehicles
Automated driving technology has advanced at lightning speed in recent years, with research pouring into sensors, vehicle-to-vehicle communication, artificial intelligence, and even public acceptance.
Yet one crucial piece of the puzzle still lags behind: how to control and schedule traffic when autonomous vehicles flood into busy city streets. Traditional local decision-making, where each car relies on its own data and immediate surroundings, only scratches the surface of what’s possible. It fails to exploit the full promise of automation in easing congestion, cutting travel times, and optimising road networks.
Cities are already wrestling with clogged roads, rising emissions, and safety concerns. Add autonomous cars into the mix and the complexity multiplies. That’s why experts argue it’s no longer enough to let vehicles make isolated decisions. What’s needed is a global perspective, one that treats the road network as an interconnected ecosystem rather than a collection of independent journeys.
The research team leading the charge
Addressing this gap, researchers Kunpeng Li and Xuefang Han from the School of Management at Huazhong University of Science and Technology, alongside Xianfei Jin from Sabre Inc., have put forward a bold new framework. Their study, published in Frontiers of Engineering Management (2024), lays the groundwork for what they call the global scheduling mode of urban automated driving traffic.
At its core, the model shifts away from localised decision-making and moves towards centralised control, achieved through vehicle-infrastructure cooperation (VIC). In other words, instead of every car deciding for itself, a central system orchestrates traffic flows across the entire network. According to the researchers: “The key is to divide large-scale road networks into manageable subregions, enabling more efficient scheduling of autonomous vehicles while balancing complexity and practicality.”
The Autonomous Vehicle Global Scheduling Problem (AVGSP)
The team conceptualised the idea as the Autonomous Vehicle Global Scheduling Problem (AVGSP). The approach recognises that modern cities are too complex to manage as a single unit. By breaking the network into subregions, the model can ensure traffic flows are coordinated without being overwhelmed by scale.
To tackle this problem, the researchers built a mixed-integer linear programming (MILP) model. They then introduced optimisation cuts that slashed the average solution time by 61%. That’s a major breakthrough, considering computational efficiency often makes or breaks large-scale traffic management systems.
Algorithms at work
The AVGSP doesn’t stop at MILP. Recognising the problem’s similarity to the classic shortest-path challenge, the team went further and developed the Modified A-star Algorithm (MASA). Unlike traditional A-star, which prioritises shortest routes, MASA accounts for the realities of traffic congestion and resource allocation.
The results speak volumes. Experiments revealed that MASA consistently produced high-quality solutions within short timeframes. More importantly, the global scheduling mode outperformed local decision-making models by reducing congestion, spreading road usage more evenly, and boosting overall efficiency. It also held up under different conditions, from varying autonomous vehicle penetration rates to scenarios where vehicles crossed regional boundaries.
Why global scheduling matters for cities
The implications of this research are far-reaching. With traffic congestion costing billions in lost productivity and environmental damage, cities are desperate for smarter solutions. Global scheduling offers a blueprint for:
- Reducing bottlenecks in key corridors.
- Ensuring fairer use of road capacity across districts.
- Improving reliability of travel times.
- Supporting future mobility systems such as robo-taxis and shared autonomous shuttles.
As urban populations swell and sustainable mobility becomes a global priority, such systems could make the difference between chaos and order on tomorrow’s roads.
The global race for smart mobility
The Chinese-led study fits into a wider global race. Across the world, governments and private firms are experimenting with intelligent transport systems (ITS) and connected vehicle frameworks. In Europe, projects under the Horizon Europe programme are testing AI-driven traffic coordination. The United States is investing heavily in vehicle-to-everything (V2X) technologies, while Japan continues to lead in advanced ITS deployments. All share the same goal: managing the surge of connected and autonomous vehicles.
But this study’s global scheduling concept sets itself apart by framing the challenge as a comprehensive optimisation problem. It’s not about marginal gains; it’s about rethinking the entire control structure of automated traffic.
Support and recognition
The research was supported by the National Natural Science Foundation of China (Grant No. 71821001), and published here: Framework, Model and Algorithm for the Global Control of Urban Automated Driving Traffic.
The findings are already attracting attention in academic and professional circles, marking a potential milestone in the future of urban traffic systems.
Towards smarter, safer roads
The transition to fully automated urban traffic won’t happen overnight. Policymakers will need to grapple with regulatory frameworks, data-sharing agreements, and cybersecurity. Technology providers must ensure interoperability between infrastructure and vehicles. Public acceptance will also play a role, as commuters come to trust a system that dictates their routes in real time.
Yet the promise is undeniable. Global scheduling could transform the daily commute, turning snarled intersections into smooth-flowing corridors. It could help cities hit climate targets by cutting idle times and fuel use. Most of all, it could put us on the road towards a safer, cleaner, and more efficient future.
Driving the future forward
Urban mobility is at a crossroads. Autonomous vehicles will either add to the chaos or become part of the solution. The work of Li, Han, and Jin suggests the latter is well within reach, if we’re bold enough to rethink how traffic is managed.
By shifting the focus from individual decision-making to coordinated global scheduling, cities can unlock the real potential of automation and reclaim control of their streets.
