Sensor Recycling via AI-Generated 3D Digital Twins for Automated Trucks
| Reference number | |
| Coordinator | Traton AB |
| Funding from Vinnova | SEK 750 000 |
| Project duration | May 2026 - April 2027 |
| Status | Ongoing |
| Venture | FFI ASP - Accelerated Startup Partnership: increased innovation collaboration between startups and need owner |
| Call | FFI ASP - Accelerated Startup Partnership: increased innovation collaboration between startups and needs owners - 2026 |
Purpose and goal
The project aims to implement AI-based “sensor recycling” for automated trucks. By reconstructing historical sensor data into photorealistic 3D digital twins, new sensor configurations can be evaluated virtually without physical prototyping. The goal is to reduce costs and development time, future-proof validation data, and accelerate the development of safe ADAS and autonomous driving functions.
Expected effects and result
The project is expected to enable virtual evaluation of future sensor configurations through AI-generated 3D digital twins based on historical sensor data. The result is reduced R&D costs, shorter development cycles, and future-proof validation data. In the long term, the project contributes to faster development and safer deployment of ADAS and automated trucks, while reducing the need for physical testing and new data collection.
Planned approach and implementation
The project is carried out through four work packages: (1) ingestion and preparation of historical sensor data, (2) creation of AI-generated 3D digital twins of real road environments, (3) virtual evaluation of new sensor configurations, and (4) integration and validation in Traton’s operational R&D environment. The work is conducted in an agile manner through frequent technical syncs and joint evaluation sprints between Traton and Recohere.