Researchers at TU Graz announced an on-demand, autonomous mobility service designed to operate 24 hours a day without a fixed timetable, aiming to bridge the gap between private car convenience and public transit efficiency.
Autonomous Shuttles Tailored for Demand
The transport sector remains responsible for roughly one-third of greenhouse gas emissions. Shifting individual travel, particularly journeys currently made in combustion-engine cars, toward electric public transport presents a direct route to cutting those totals. To tackle this, researchers at TU Graz collaborated with partners in China to build a battery-electric, self-driving shuttle bus capable of carrying up to 15 passengers.
Rather than running along fixed lines on a rigid schedule, the vehicle responds dynamically to user requests within an expanded routing network. A test vehicle is already operating successfully on a real-world road network in China, while demonstrator vehicles are planned for deployment in Austria.
Collaborative Perception and High-Fidelity Simulation
Safe navigation relies heavily on collaborative perception. The shuttle combines onboard lidar, radar, and camera systems with 5G connectivity linked to intelligent infrastructure, such as cameras positioned at challenging or obstructed intersections. This setup allows the vehicle to detect pedestrians and cyclists in real time, even when they remain hidden from the shuttle’s native sensors. In the event of a network failure, an independent safety system brings the vehicle to a controlled stop.
Before hitting physical roads, the team tested the safety architecture using high-fidelity simulations. Researchers built virtual replicas of real streetscapes, including buildings and surrounding infrastructure, to validate algorithms against everyday and extreme scenarios. Because validating autonomous systems requires countless test variations and millions of virtual kilometers, simulation remains the primary method for securing safety approvals.
Dynamic Optimization for Fleets
The research partners in China developed dynamic optimization tools for routing and vehicle charging to minimize passenger wait times while maximizing fleet utilization.
“Good utilization in particular helps to shrink the ecological footprint,” Lex noted. “When several people reach their destination using an electrically powered shuttle instead of everyone sitting alone in a car, a great deal is gained in terms of air quality and space on the street.”
The project brought together TU Graz, Tongji University, Wenzhou University, and Seisys, with financial backing from the Austrian Research Promotion Agency (FFG).
