Researchers at the Technical University of Vienna (TU Wien) have developed a self-learning robot that is able to mimic a human being for simple tasks, such as cleaning a washbasin.
This seemingly simple task is a huge challenge for the robot. Traditional robot programming requires precise mathematical modelling of the complex curves of the washbasin, a time-consuming and complex task. Instead, the team simplified the process by combining human demonstration and haptic data to train the robot to mimic human movements.
This learning approach is not only applicable to household cleaning, but can also be applied to a wide range of tasks in industrial production, such as polishing, painting, sanding and gluing.
'Capturing the geometry of a washbasin with a camera is relatively simple,' says Professor Andreas Kugi of the Institute of Automation and Control at the Technical University of Vienna, 'but the key is to teach the robot how to choose the right movement for different surfaces - how fast should it move? What is the right angle? What is the right amount of force?'
IT House notes that this type of learning is similar to how humans learn new skills, especially those that require hands-on work.
'In the workshop, a master would instruct an apprentice to say, 'You need to press a little harder on that narrow edge,'' said Christian Hartl-Nesic, industrial robotics group leader on Andreas Kugi's team. 'We wanted to find a way for the robot to learn in a similar way.'
To do this, the team developed a special cleaning tool - a sensor-impregnated sponge. The human 'teacher' uses the sponge to repeatedly clean the front edge of the sink using force sensors and tracking markers.
'We generated a lot of data in just a few demonstrations and then processed it to help the robot understand the right way to clean,' explains Christian Hartl-Nesic.
The research team then used an innovative data-processing strategy to enable the robot to efficiently clean an entire sink or other object with a complex surface, even if it had only been shown how to clean one edge of the sink.
Christoph Unger, a PhD student in the Industrial Robotics group, explains, 'The robot learnt to adapt the way it holds the sponge to the shape of the surface, as well as applying different forces in tightly curved areas and on flat surfaces.'
In the future, the robots could be mounted on mobile platforms, making them powerful assistants in any workshop environment. And just like humans, these robots could pass on their experience to other robots.