Result description
AI is currently used in automated driving for supporting tasks like image recognition. For safe autonomous driving, it will be impossible to program all systems with hand-written code especially when the world around the autonomous vehicle becomes more complex, such as in urban situations where there are much more different actors around the vehicle than in highway situations. Therefore we studied the use of (hybrid) AI for the complex task of Decision Making and in PRYSTINE we were able to demonstrate a proof of concept of a hybrid AI set-up for Decision Making in autonomous vehicles
Proof of concept for a scalable set-up for AI-based Decision Making for highly automated vehicles. We demonstrated the results in defined highway and urban scenarios using FUSION-AI based algorithms in dedicated HW, in combination with a traffic state prediction system. This traffice state prediction functions as an additional external sensor (looking at the traffic state ahead, so beyond the field of view of the vehicle.
Addressing target audiences and expressing needs
- Business partners – SMEs, Entrepreneurs, Large Corporations
- Expanding to more markets /finding new customers
- Collaboration
TNO would like to continue the research in the field of Connected Cooperative Automated Mobility. Continuing with further research and development of using a hybrid AI set-up for Decision Making is one of these topics.
- Other Actors who can help us fulfil our market potential
- Research and Technology Organisations
- Academia/ Universities
R&D, Technology and Innovation aspects
Current stage is that we managed to show proof of concept for a scalable AI set-up using Inverse Reinforcement Learning in defined highway and urban use cases. More research is needed. We have chosen Inverse Reinforcement Learning (IRL) for the Decision Making task. This is at a low TRL-level, mainly due to its complexity and required computation power. We learned that it is a promising way forward but that it requires a lot of data and a lot of computation power in order to develop this into higher TRL levels.
Result submitted to Horizon Results Platform by ANYWI TECHNOLOGY BV
Please 
Get instant access to tailored funding opportunities that perfectly match your needs. This powerful feature is exclusively available to our premium members—helping you save time, stay ahead of the competition, and secure the right funding faster.
Curious about who’s viewing your listings? Unlock insights into your last 20 visitors and discover potential leads instantly! 