For safe deployment of automated vehicles on the public road, the AI systems in the vehicle should be aware of their own competence in the current situation, and act more cautiously or hand over to the driver in situations where the AI is uncertain.


The current generation of driver assistance systems in vehicles are designed to operate in a limited domain (for example only on highways) and in a limited set of circumstances (like no road works). It is the driver’s responsibility to judge whether the automated system is able to perform its tasks in the current situation. However, in practice it has turned out that drivers have too much confidence in the automated system, which has led to accidents.

Research approach

TNO is designing a system that combines knowledge about traffic (in the form of a knowledge graph) with data-driven AI into a hybrid-AI system that can reason about the competence of the system in the current situation. This is demonstrated in an AI system that predicts the intention of other road users. The newly developed hybrid-AI solution can reason about the competence of the prediction and can advise the vehicle to hand control over to the driver.


  • Jan-Pieter Paardekooper, Scientist, TNO, e-mail: