Decision making in autonomous driving
Busy intersections are hard enough for people to cope with. So how do we ensure that self-driving cars can safely make their way around towns and cities in the future? TNO is currently experimenting with ‘hybrid AI’ as part of an EU collaborative project.
A constant stream of pedestrians and cyclists. The driver behind you banging impatiently on his horn. For car users, negotiating such chaotic traffic situations can be trying enough, but for self-driving cars, it is even more difficult. After all, there are always sensors that notice an approaching object.
Write the software by hand? forget that!
In other words, self-driving cars are a long way off being able to deal with the complicated traffic situations that exist in urban settings. In the meantime, though, more and more driver-related tasks are being automated in cars. That requires complex software – so complex that it is not possible to write it by hand.
An autonomous car that takes the right decisions
Working closely with NXP, Infineon, DAT. Mobility, AnyWi, and Eindhoven University of Technology, TNO is now part of an EU project dedicated to preparing self-driving cars for urban traffic. Together, we are attempting to demonstrate that applying AI algorithms will make decision making possible. This is an important point, as it could swiftly accelerate the development of a safe autonomous car.
No standard AI, but ‘hybrid’ AI
The AI algorithm works closely with a knowledge graph. This is an information system that AI requires in order to be able to properly assess dangerous traffic situations, and therefore make the right decisions. This ‘hybrid AI’ is a combination of data-learning AI and domain knowledge.
Motorways first, and then more difficult situations
A car fitted with radar, lidar, GPS, and cameras drove around for several days for the purpose of collecting data. TNO is now using the information that it collected to train the AI algorithm. At this first stage, the scope does not extend beyond situations on motorways. This because the conditions on motorways are more clear-cut. The next step will be a good deal more complicated: this will involve the AI algorithm having to decide when a car is able to cross an intersection where there are cyclists. Safely, obviously. In other words, the self-driving car will have to constantly monitor the behaviour of cyclists and respond accordingly.
Handy in harbours
In due course, this type of AI algorithm could also have an important part to play in many other applications, in addition to self-driving cars. One such possible example is that of automated guided vehicles (AGVs) in ports. Another is a care robot, moving around in a hospital. This is another setting that would benefit from an effectively functioning vehicle.
Christopher BrewsterFunctie:Senior scientist
Christopher Brewster is a Senior Scientist in the Data Science group and Professor of the Application of Emerging Tecnologies in the Institute of Data Science, Maastricht University. His research has focussed on the application of Semantic Technologies, Open and Linked Data, interoperability architectures and Data Governance, mostly to the food and agriculture domains.
Daniël WormFunctie:Senior consultant
Joris SijsFunction not known
Judith DijkFunctie:senior research scientist
Judith is specialised in extracting information from camera images. She now applies the subject of her PhD thesis in Physics, which she obtained 18 years ago, to her work as a research scientist at TNO, including in a research programme on camera systems for the Dutch Ministry of Defence.
Looking for another expert?View all experts
Responsible decision-making between people and machines
Bias in facial recognition and accidents with self-driving cars. AI must be developed further. The fastest way to do this is in close cooperation with people.
Knowledge representation and reasoning
Correct and unambiguous information is needed when making a decision. That is why we use AI technology called "knowledge representation & reasoning".
Natural language processing
What is natural language processing (NLP) and how do we use it intelligently? Find out how we use this AI technique to gather information from textual data.
Robotics and autonomous agents
Robotics brings a future-proof industry a big step closer. For example, we are working on automatic path planning with AI techniques.
Fair machine learning
Fair machine learning is relevant to all kinds of discrimination and bias arising from the use of biased data. Read more!