The LiU researchers use hybrid AI, an approach that combines different methods from the AI field such as machine learning, reasoning and planning. In this way, they create practically useful systems for specific operations.
“We figure out how to implement safe and robust AI systems that work in real life and are reliable,” says postdoc Mattias Tiger.
Many stakeholders want to collaborate with the research group. For example, the researchers have looked at how new AI systems can support several of Region Östergötland’s operations, including health care. The group is also collaborating with the public transport provider Östgötatrafiken and has improved the punctuality of their buses. Current estimated arrival times originate from previous projects where they made major improvements using machine learning. The researchers also have ongoing collaboration projects with the Swedish Transport Administration to improve arrival and departure time estimates for all rail traffic in Sweden. The results have been good, and the research group is now supporting the Swedish Transport Administration in moving the forecast models to production.
A research project with Saab and Ericsson is about coordinating a large number of low-flying objects in society and the safety aspects this would entail. The task of such an object can be to deliver packages, collect inspection data on things such as heat loss in a house or make a three-dimensional map for self-driving vehicles.
The article is also published in LiU magasin.
