Uncertainty-aware and safety-enhanced management of CAVs for safer mixed traffic
| Reference number | |
| Coordinator | Chalmers Tekniska Högskola AB - Inst f Arkitektur & samhällsbyggnadsteknik |
| Funding from Vinnova | SEK 500 000 |
| Project duration | August 2024 - April 2025 |
| Status | Completed |
| Venture | Safe automated driving – FFI |
| Call | Traffic-safe automation - FFI - spring 2024 |
Important results from the project
The project fully achieved its goals, completing all five deliverables with robust technical and practical outcomes. Key advances include an interpretable trajectory prediction framework, uncertainty-aware control designs, and a modular CarSim-MATLAB simulation. New additions include risk perception via vision-based labeling and SHAP-based explainability, enhancing human-centered design and AI transparency in safety-critical systems.
Expected long term effects
The project addressed the need for safe and resilient CAV operation in complex traffic by developing a framework focused on uncertainty perception, quantification, and adaptive control. Key outcomes include trajectory prediction with driver heterogeneity, multi-source uncertainty handling, and robust platoon control. Validated in CarSim, the work supports FFI goals and provides tools for future CAV development in non-ideal conditions.
Approach and implementation
The project has implemented as planned without delays and negative deviation. Combining another pre-study short-term project, we also extended the research contents about risk perception modeling using computer vision-based labeling.