Most haptic interactions involve a very large number of widespread skin sensors that provide information mutually modulated by both the skin mechanics and the interaction of the skin with objects. Predictive coding provides a principled computational approachable to extract perceptual information of the raw sensory signals. Conversely, it will inform the designers of haptic interfaces of methods require to synthesize and transduce the signals that can elicit desired percepts that can be experienced without training or habituation. The specific work-packages are to:
(1) Develop a predictive coding model to extract stable haptic percepts despite variability in the sensory signals, and demonstrate it in a robotic implementation.
(2) Design and fabricate a soft artificial eSKIN based on bendable electronics integrating a variety of micro-/nano-sensors.
(3) Demonstrate the transmission of ultra-rich haptic information to a remote user via novel active haptic interfaces. Carried out by leading neuroscience, nanotechnology and robotics groups and an SME at the forefront of haptics technology, this ground-breaking research will usher a new generation of passive and active devices enabling human-like, robust and rich interaction with objects, devices, and robots.
Project in collaboration with :