INTENTION in Motion: How Autonomous Vehicles Learn to Anticipate Cyclists

A Quick Reminder: What INTENTION Is About

As announced when we launched the project, INTENTION is one of our SAB research projects, funded by the Sächsische Aufbaubank (SAB) and driven forward by our Inno-Team.

Our goal is to recognize the intentions of cyclists early and to feed this understanding directly into the trajectory calculation of autonomous vehicles. In other words, we want the vehicle not just to see a cyclist, but to anticipate what that cyclist is likely to do next – and to plan its path accordingly.

It is precisely this anticipation that makes the difference between a vehicle that merely reacts and one that interacts safely with the most vulnerable road users.

From Concept to Cockpit: The PixKit Demo

At our most recent project meeting, things got especially exciting. Martin Haase presented our internal development kit, the PixKit from PIX Moving and the vehicle showed just what it is capable of:

• Fully autonomous navigation within a predefined corridor
• Successful emergency braking maneuvers
• Autonomous parking procedures

A big thank you to the Pix project team for making this possible!

The live demonstration prompted plenty of questions among our project partners – around path planning, path following, map creation, and obstacle and traffic-light detection, and above all their interplay with cyclist intention recognition. This is exactly where the different building blocks of autonomous driving have to come together into one coherent system.

Looking Ahead While Building Today

Our partners at Envisible captured the essence of the challenge well: on that Friday, we took a genuine look into the future and tested the interaction between cyclists and autonomous vehicles.

For now, this still takes place at walking speed on closed test tracks. And that defines the real challenge of INTENTION: we have to anticipate and think ahead about the parallel developments in autonomous mobility, while we work intensively on recognizing and predicting the behavior of cyclists in real traffic.

In other words, we are developing for a reality that is still emerging – designing systems today that will need to perform in the more complex, faster, and open traffic environments of tomorrow.

What Comes Next

The demo didn't mark the end of the meeting – it was the springboard for what followed. We moved straight into a brainstorming session on new project ideas and into the detailed specification of our upcoming study on the cyclist's field of view.

Because understanding cyclists means understanding both sides of the encounter: how the vehicle perceives the cyclist, and how the cyclist perceives and moves through the traffic environment.

Strong Partners Behind INTENTION

None of this would be possible alone. A big thank you to our strong project partners:

• TU Chemnitz – SportsEngineering SGT
• Professorship of Ergonomics and Innovation Management
• Professorship of Production Systems and Processes, TU Chemnitz
• FDTech GmbH

Together, we are shaping the future of safe, autonomous mobility – while keeping the perspective of cyclists firmly in view.