Artificial Intelligence enhancing vehicle vision in low visibility conditions

Press Release

One of the biggest remaining challenges towards the commercialization of automated vehicles is the ability to drive also autonomously under any relevant weather and lighting conditions. To date, the testing and validation of automated vehicles is mostly taking place in a small, designated area under good weather conditions and even the best in class prototypes still struggle or even fail completely in adverse weather situations. The key for market introduction and large proliferation is a robust and fail-safe perception system able to operate in all relevant weather conditions experienced 365 days of the year.

AI-SEE, a PENTA EURIPIDES² funded research project, will try to close this gap by developing a novel sensing technology with associated AI enabling automated driving in all relevant weather and lighting conditions (e.g. snow, heavy rain or fog) in a 24h/365-day mode. Thereby, AI-SEE will extend the Operational Design Domain (ODD) of automated vehicles and extend the technology from today’s SAE level 3 (conditional automation) to SAE level 4 (high automation) where vehicles can drive themselves with no human interaction in almost any circumstances. Thus, AI-SEE has the potential to set a new industry benchmark for automated driving that can operate flawless in all visibility and weather conditions.

The project will set up on the results of the EU-funded predecessor project DENSE. The environment perception system developed in DENSE will be improved by realistical sensor simulation runs that are conducted under adverse weather scenarios and is adapted to and enriched with artificial data. The result will be a robust, fault-tolerant, high resolution and adaptive sensor suite operating in all-weather and daylight conditions as experienced 365 days of the year on the road.

Specifically, the innovative sensor suite will comprise:

For the first time the data obtained from the high resolution adaptive all-weather sensor suite will be fused and simulated by means of sophisticated AI-algorithms tailored to adverse weather perception needs.

To test the viability of the components as a means for safe automated driving, testing campaigns in real traffic scenarios will be carried out. In addition, extensive tests for sensor and perception validation will be conducted in the partners’ fog and rain chambers.

With its innovative sensor system AI-SEE will take automated driving to the next level, enabling the market introduction of SAE level 4 automation by 2030 and strengthening Europe’s competitiveness in this highly competitive automotive sector.

AI-SEE Structure

Project Budget
€21.57 million

€9.51 million

Project Timeframe
36 Months (01/06/2021 – 01/06/2024)

Project Leader
Mercedes Benz AG

21 Partners
OEMs, Automotive Suppliers, Research Institutes, Engineering Companies

AI-SEE Partners

Austria, Canada, Finland, Germany, Israel, Sweden

AI-SEE is a co-labelled PENTA and EURIPIDES² project endorsed by EUREKA.

National Funding Authorities
Austrian Research Promotion Agency (FFG), Business Finland, Federal Ministry of Education and Research (BMBF), National Research Council of Canada Industrial Research Assistance Program (NRC-IRAP)  

Project Coordinator
Dr. Werner Ritter
Manager Vision Enhancement Technology
Environment Perception
Mercedes-Benz AG
D-71059 Sindelfingen, Germany

Project Management
Dr. Hristiyan Stoyanov & Majid Vesalimahmood
European Center for Information and Communication Technologies – EICT GmbH
EUREF Campus Haus 13
Torgauer Straße 12-15
10829 Berlin, Germany;

Communications Manager
Myriam Ben Ammar
Phone: +49 30 3670 235-129

Data Privacy