Engineering Fellowship for Growth - Neuromorphic Printed Tactile Skin (NeuPRINTSKIN) (Ext)

点击翻译按钮获取中文摘要

This project is an extension of the Engineering Fellowship for Growth: Printable Tactile Skin (PRINTSKIN). PRINTSKIN focused on developing a robust ultra-flexible tactile skin, endowing state-of-the-art robotic hand with the tactile skin and validating it by using tactile information from large areas of robot hands to handle daily object with different curvatures. The tactile skin is critical for autonomy of robots and for the safe human-robot interaction need to meet societal needs such as helping elderly. The tactile feedback is critical in such tasks as the robots often use incomplete environmental model which are insufficient to deal with external changes. The touch sensing is also needed to augment other sensory modalities (e.g. vision) in robotics. Inspired by nature, numerous works including PRINTSKIN project, have harnessed the technological advances to develop e-skin with some features mimicking human skin - particularly the contact parameters and morphological features. However, just morphology of skin or capturing few parameters that we experience as touch is not enough. To develop an effective tactile skin, there is also a need to understand the perceptual mechanism and to find the ways to extract the information from large tactile data (especially in the case of large area tactile skin). Research suggests that distributed computing takes place in the biological tactile sensory system. For example, the ensemble of tactile data from peripheral neurons is considered to indicate both the contact force and its direction. This means raw tactile data is not sent to brain and that some distributed computing takes place in our skin. This is in sharp contrast with current e-skin approaches which transmit the as acquired tactile data to higher perceptual levels. The research proposed here will break this trend by introducing neuron like processing and bring a step change in the tactile sensing research by developing the first neuromorphic tactile skin or the brainy skin. A new neural layer, developed using the printed silicon nanowire methodology developed in PRINSKIN, will be integrated under the e-skin to enable fast, energy efficient and distributed tactile data processing. This groundbreaking research will lead to the first hardware implementation of neuromorphic tactile skin. Innovative schematic, with novel neural nanowire field effect transistors and memory devices as building blocks, will be used to develop the neurons which will eventually lead to the neural layer. The advanced tactile skin will be benchmarked against available semi-rigid skins and the skin developed through PRINTSKIN. The skin will be validated on at least three different robotic hands (Shadow Hand, i-Limb, and custom 3D printed hand) used for dexterous manipulation and prosthetics. By adding neural layer underneath the current tactile skin, this extension project will add significant new perspective to the fellowship achievements and trigger transformations in strategic areas such as robotics, prosthetics, neurotechnology, wearable systems, next-generation computing and flexible and printable electronics.