Hybridization of 3d Stacked Piezoelectric Nanofibers and Biomimetic-Textured Triboelectric Layers as Self-Powered Sensors for Developing the Large Scale Data Deep Learning Algorithms(2/3)

This project proposes a novel hybridization of 3D stacked piezoelectric nanofibers and biomimetic-textured triboelectric layers as self-powered sensors for developing the large scale data deep learning algorithms. Three specific applications include the biomedical engineering applications as the in-situ monitoring devices of stroke-patient rehabilitation and smart mask of Tourette’s patients. Additionally, the personalized human-machine interfact smart keyboard will also be applied based on the developed devices and algorithms. Respective aims and project goal can be summarized as follows:Year 1: In the first year study, the 3D piezoelectric PVDF fibers were sequentially constructed on the flexible substrates such as PCB or paper to fabricate the highly flexible and structurally durable piezoelectric sound-sensing elements with a simple processing and low cost strategy. In addition, the ultrathin intelligent self-powered sound-sensing elements can tightly cohere on the human moveable joint, can not only detected the vibration cause by sound but also detect the human motion without external energy source.Year 2:In the second year project, the smart patch was further integrated with the micro/nanotextured surfaces as the triboelectric friction layers to significantly enhanced the electrical output of seld-powered sensors. The developed sensors will be applied to the smart mask for the Tourette;s patients, as a mean to quantatively monitored the physical behaviors for the medical doctors to estimate the progress of symptoms.Year 3: The third year will continue the advancement of biomimeticnanotextured surfaces (the characteristic length will be 10-100 nm) as the triboelectric friction layers and therefore, the enhaced hybrid self-powered sensors can be effectively utilized as the basis for the large scale analysis in the era of artificial intelligence. he specific application will be the intelligent keybaord (IKB) such that indivisual typewriters’ keystroke dynamics can be identified via the deep-learning-based algorithm for increased keyboard-based information security. Multilayer long short term memory neural network (LSTM) will be established to mine the useful information from raw electronic signals as recorded from the developed hybrid sensors and output the keystroke dynamics identification result.