Researchers from the National University of Singapore (NUS) announced on Wednesday that they are conducting work aimed at giving robots a sense of touch through artificial skin.
The two researchers, who are also members of the Intel Neuromorphic Research Community (INRC), presented research that demonstrates the promise of visual and tactile sensing combined with Intel neuromorphic processing for robotics.
The majority of today's robots operate solely based on visual processing and lack the sense of touch that humans have.
Researchers hope to change that using artificial skin, which NUS believes will be able to detect touch more than 1,000 times faster than the human sensory nervous system. The artificial skin, NUS said, will also be able to recognize the shape, texture and hardness of objects “10 times faster than the eye.”
Intel Neuromorphic Computing Lab director Mike Davies said the research provides a glimpse into the future of robotics where information is sensed and processed in a fact-based way.
NUS said that enabling a human sense of touch in robotics could significantly improve current functionality, offering the example of robotic arms fitted with artificial skin that could easily adapt to changes in products being manufactured in a factory, using the sense of touch to recognise and grip unfamiliar objects with the right pressure to prevent slipping.
Intel is helping researchers by providing a chip that deploys inside the robot to draw accurate conclusions based on real-time skin sensory data.
They also need an artificial brain that can ultimately achieve perception and learning, which is another very critical piece of the puzzle. Our unique demonstration of a cutaneous AI system with neuromorphic chips like Intel Loihi is a significant step forward in power efficiency and scalability.
Using Intel's Loihi neuromorphic research chip, in their initial experiment, the researchers used a robotic hand with artificial skin to read words written in Braille, transferring the data to Loihi via the cloud to convert the microdevices that sensed the hand into "semantic meaning."
Intel said Loihi achieved over 92% accuracy in classifying Braille letters, while using 20 times less power than a typical Von Neumann processor.
Building on this work, the NUS team further improved the robot's perception capabilities by combining vision and touch data with a neural network. To do this, they tasked a robot with sorting various opaque containers containing different amounts of liquid, using sensory inputs from the artificial skin and a camera.
The received sensory data was then sent to GPUs and Loihi to compare their processing capabilities. The researchers recorded that the combination of vision and touch resulted in 10% greater accuracy in object classification compared to a system using vision alone.
