We are proud to share the achievements of the AI4CSM project, part of the Horizon 2020 ECSEL-JU program, where the EDI – Institute of Electronics and Computer Science team demonstrated outstanding innovation and dedication. Our experts – Dr. Kaspars Ozols, Dr. Rihards Novickis, Edgars Lielāmurs, Ibrahim Sayed, Anatolijs Zencovs, Andis Bizuns, Maksis Celitans, and Andrejs Cvetkovs – successfully prepared the final demonstrator, which you can see in the video below.
As part of the project, the EDI team developed and validated a custom hardware and software solution for an AI-based near-field, high-resolution 360-degree perception system using Time-of-Flight (ToF) sensors. Dr. Kaspars Ozols and Dr. Rihards Novickis provided project leadership, working in collaboration with key technology partners.
Key technical outcomes of the project:
- Development of a 360-degree perception system using a 3D ToF camera network;
- Algorithms for 3D data registration and continuous fault monitoring;
- ToF camera synchronization solutions;
- Spatio-temporal 3D data encoding;
- 3D object detection with conventional Convolutional Neural Networks (CNNs);
- 3D object detection using modern Spiking Neural Networks (SNNs);
- Integration of the system into the EDI Drive by Wire (DbW) vehicle and testing in a controlled environment;
- Distributed system architecture with real-time monitoring.
We especially thank our collaborators: Infineon Technologies (Jochen Koszescha and Jochen Reisinger), Mercedes-Benz AG, and AVL. Their support enabled our solution’s successful integration and testing in a Mercedes-Benz vehicle at AVL’s facilities in Graz, Austria.
Our success has also been highlighted in the AI4CSM project’s latest newsletter.