SSVEKI

Pedestrians, cyclists and other vulnerable road users have hardly any protection when a vehicle hits them. SSVEKI develops a system that foresees such collisions in real time and warns in time, using artificial intelligence that runs directly on embedded hardware rather than in a distant cloud.

It starts from the position, speed and heading of vehicles and road users, as modern sensors and connected vehicles provide them. From these the system is to recognise dangerous situations before it is too late, so that accidents are avoided and traffic as a whole becomes safer.

Our contribution

We are responsible for integrating the AI into the embedded system: designing, shrinking and mapping the collision detection models onto the target hardware so that they meet their real-time requirements with limited computing power and energy. We bring in our work on AI on RISC-V processors and microcontrollers and on virtual platforms, on which the software can be tried out before the hardware exists.

Partners and funding

Our partner is consider it GmbH of Hamburg, which develops technology for connected mobility (V2X). The project is funded by the German Federal Ministry for Economic Affairs and Energy through the Central Innovation Programme for SMEs (ZIM).