Self-learning safety systems in EU project
A system that teaches itself by observing human action and reaction. That is the goal for the EU-backed project DIPLECS (Dynamic-Interactive Perception-Action LEarning Systems) which is coordinated by Linköping University.
A technologically advanced safety system for automobiles is an intrinsic element of the project.
Many of today's new cars are equipped with several kinds of driver-assisting automation: anti-lock braking systems (ABS), traction control, adaptive cruise control (ACC), and collision detection systems.
"We need much smarter solutions than these, future systems will capture a total overview of the road and traffic situation and analyze it," says researcher Michael Felsberg.
He is associate professor in computer vision at the Computer Vision Laboratory and project coordinator. The total funding is 2.6 million Euro financed by the EU Seventh Research Framework Programme (FP7). LiUs share is around 45 %.
The expertise of LiU computer vision researchers has long been recognized by the programmes. LiU received grants even before Sweden was an EU member. A current EU-LiU project, COSPAL, created an industrial robot system capable of self-learning in a real, non-mobile environment. The new challenge is to create one which can self-learn in a dynamic and interactive environment where unanticipated events occur.
"One of our industrial partners tested the COSPAL system concept in a small radio-controlled model," Michael Felsberg continues, "and we became interested in the cognitive-car concept."
The industrial partner in the DIPLECS project is the Swedish company, Autoliv, a global leader in automotive safety.
Thirteen other projects with LiU participation have been awarded funds from FP7:
• SHIELDS, Nahid Shahmehri, Department of Computer and Information Science (coordinator)
• LiveDiverse, Geoffrey Gooch, Department of Management and Engineering (coordinator)
• Polynet, Magnus Berggren, Department of Science and Technology
• ArtistDesign, Petru Eles, Department of Computer and Information Science
• ONE-P, Mats Fahlman, Department of Physics, Chemistry and Biology
• FNIR, Rego Granlund, Department of Computer and Information Science
• Sendora, Erik G. Larsson, Department of Electrical Engineering
• Multibase, Dake Liu, Department of Electrical Engineering
• TheraEDGE, Sigvard Mölstad, Department of Medical and Health Sciences
• ELIXIR, Bengt Persson, Department of Physics, Chemistry and Biology
• AHEAD III, Stefan Stenfelt, Department of Clinical and Experimental Medicine
• IAPP@ranplan, Di Yuan, Department of Science and Technology
• DEBUGIT, Hans Åhlfeldt, Department of Biomedical Engineering
Last updated: 2009-06-03