From Pilot Project to Public System

Why Autonomous Mobility in Public Transit Is Too Often Underrated

When discussing autonomous mobility in public transit, we must not think in terms of individual vehicles or small test fleets. The real question is: How does a pilot project become a system relevant to the public?

The series “From Pilot Project to Public System” does not first ask whether autonomous vehicles are technically capable of driving. The crucial question is whether they can become robust, scalable, and effective as part of public transportation. And it is precisely at this point that many current projects are still conceived too narrowly and/or in isolation.

Too many initiatives remain so narrowly defined in terms of geography, operations, and politics that while they generate visibility, they provide hardly any reliable answers to the central systemic question: How can autonomous mobility in public transport be translated into widespread implementation, operational reality, and everyday use? Bitkom therefore calls for larger pilot regions, broader service areas, and higher numbers of vehicles. Source: https://www.bitkom.org/sites/main/files/2024-10/241001-bitkom-thesenpapier-autonomes-fahren-in-deutschland.pdf

An international perspective: In Singapore, the scale is already being set significantly higher. The Ministry of Transport states: “AVs will add to our public transport network without running up against fundamental manpower constraints” and further: “These shuttle services can directly connect residents to key transport nodes and amenities such as the market and polyclinic.” That is precisely the difference: The focus is not on the individual test vehicle, but on its role within the public transport system. Source: https://www.mot.gov.sg/news-resources/newsroom/oral-reply-by-senior-minister-of-state-for-transport-to-parliamentary-questions-on-avs/

A vehicle operating in a limited test environment does not yet constitute a reliable public transit service. A demonstrator shows that the technology can work. However, it does not automatically demonstrate how availability, integration, demand, operational stability, and public benefit are organized in everyday life. This is precisely the difference between pilot logic and system logic.

The federal government also aptly describes the situation as a phase between completed testing and the lack of scaling. At the same time, there is a lack of production-ready services as well as scalable operator and business models for autonomous shuttle services. Source: https://www.bmv.de/SharedDocs/DE/Publikationen/StV/die-zukunft-faehrt-autonom.pdf?__blob=publicationFile

What is often overlooked here is that the step from demonstrator to operation is not merely a question of fleet size, business model, or organization. It is also a question of the technical controllability of vehicle movement in real-world operation.

In pilot operations, many aspects can still be safeguarded: through limited spaces, low complexity, clearly defined scenarios, or additional intervention options. In regular operations, the requirements shift fundamentally. Vehicles must remain safely controllable at all times, under varying conditions, and without an implicit human fallback.

This is precisely where the actual system gap arises. Autonomous systems today can detect, calculate, and decide on many things. The real test begins where these decisions must be controlled and safely translated into motion under everyday conditions.

This brings a central question to the fore that is underestimated in many projects: How does vehicle movement remain controllable at all times—even when conditions vary, systems degrade, or human intervention is no longer provided for?

If autonomous mobility is to become part of public transportation, it must therefore be conceived on the scale of real-world operations: in terms of fleets, service areas, interfaces, operator structures, and real user needs. And above all, in systems that keep vehicle movement safely controllable at all times. Because scaling in public transportation means more than just more vehicles. Above all, it means: reproducible, controllable movement in real-world operations.

The Handbook on Autonomous Driving in Public Transport therefore describes autonomous services not merely as a technical issue, but as an integrated planning and operational task. Source: https://www.bmv.de/SharedDocs/DE/Anlage/K/autonomes-fahren-oeffentlicher-verkehr.pdf?__blob=publicationFile

The gap between pilot projects and widespread public deployment is therefore not just a matter of funding, regulation, or operator logic. It is also a matter of system responsibility. Anyone who wants to bring autonomous mobility into regular operation must design in-vehicle control from the outset in such a way that it remains viable under real-world conditions.

Before we discuss operator logic, trust, financing, or accessibility, we must clarify the fundamental question: Why do so many approaches fail before they can be scaled up?

Not because vehicles are fundamentally incapable of driving. But because a functioning demonstration has not yet become a manageable public system.

Anyone who seriously wants to advance autonomous mobility in public transit must stop treating pilot projects as endpoints. They must be planned as a precursor to a public system—technically, operationally, and architecturally.