Finally! The project ILIAS has its Kick-Off Meeting scheduled for August 5th – August 7th, 2019!
The project is a cooperation between Germany and South Korea and funded by the BMBF (Federal Ministry of Education and Research). The BMBF aims at supporting R&D-projects (Research & Development) between the two countries that involve the participation of two industrial partners and research institutions from both countries (also referred to as “2+2”) with a focus on robotics.
ILIAS is an acronym for (German): Imitationslernen von menschlichen Demonstrationen in virtueller Realität für physikalische Mensch-Roboter-Interaktion bei Assistenzaufgaben or (English): Imitation Learning from Human Demonstrations in Virtual Reality for Physical Human-Robot-Interaction in Assistance Tasks.
The project proposes a novel way of programming robotic assistance tasks that will scale better towards open task domains. “Task domains” describe different categories of tasks: mundane, formal, and expert, ranging from tasks that do not require special training (mundane/everyday activities) to tasks that require special expert knowledge. Assistance tasks are used, for example, for people with motor impairments performing everyday care tasks.
In ILIAS, the programming approach includes humans demonstrating how to accomplish assistance tasks in virtual environments. The demonstrations performed by humans are transformed into knowledge (data)bases. This system then uses the data from the demonstrations in virtual environments in order to learn complex motion primitives, i.e. data types, using deep learning (a method of information processing using neural networks similar to the human brain). The learned complex motion primitives are then transferred to motion planning methods, because one of the crucial tasks for robots is to move collision-free in an environment. The methods adapt motion plan patterns to new tasks and situations.
This approach is of great value to robotic companies that develop robots used, for example, to provide assistance to elderly people. This is of special importance for societies that struggle with changing demographics and an aging population.
EASE will closely collaborate with ILIAS. We are excited to see how the project will develop.