Two separate but related intriguing new touch technologies for robots from Georgia Tech has some intriguing implications.
The first being Charlie Kemp’s development of a form of tactile sensing for robotic arms (Kemp was supervised by the world-famous Aussie roboticist Rodney Brookes). Reported in the New York Times (here) and elsewhere, searching in cluttered places for objects poses a real challenge for robotics. The research is published in The International Journal of Robotics Research (paper here), the ability to relate ‘touch’ data with visual search has been productive. And this form of search in cluttered environments will be invaluable outside of the automated and controlled sites where most robots work. To work alongside humans in healthcare or disaster recovery settings, say, cluttered environments are the norm. Increasingly, robots are having to work alongside humans and so the safety issues of human-robotic interactions (HRI) are considerable. Here is a video of the robotic arm wiping a hospital patient’s mouth:
And as the New York Times says:
The robot also has an artificial “skin” that can sense pressure or touch.
Made all the more exciting by the fact that this Georgia Tech project is also open source, so this skinning could proliferate for all kinds of projects. It is developments like these that are going to drive robots increasingly into mixed spaces of human and robotic interactions.
In a separate but related development, as reported in MIT Technology Review (here), Zhong Lin Wang and others at Georgia Tech have developed an ‘artificial skin’ as a flexible, bendable layer that could be used for object detection but also, with its fine spatial and pressure discrimination, other applications will arise. Such as artificial touch for prostheses. Now, this idea of artificial skin is not itself new (e.g. Someya 2005’s ‘e-skin’ amongst others), in this instance it’s pretty startling how
The density, resolution, and sensitivity of the sensors, says Wang, is comparable to that of the skin of a human finger.
Not only that, but being truly flexible might speed its uptake across a number of different contexts, such as smart clothing/ wearable computing applications.
Leave a Reply