A Hybrid Microstructure Piezoresistive Sensor with Machine Learning Approach for Gesture Recognition

Developments in flexible electronics have adopted various approaches which enhanced the applicability of human–machine interface fields. Recently, microstructural integration and hybrid functional materials were designed for realizing human somatosensory. Nonetheless, designing tactile sensors with smart structures using facile low-cost fabrication processes remains challenging. Furthermore, recognizing stimuli feedback applications poorly validated. In this study, a highly piezoresistive sensor was developed by homogeneously dispersing carbon black (CB) microstructure porous sugar/PDMS-based sponge. Owning to its high flexibility softness, can be mounted on or robotic systems different clinical applications. We validated proposed applying it grasp release forces an open setting classifying hand motions that surgeons apply master system during intravascular catheterization. For purpose, we implemented long short-term memory (LSTM)-dense classification model five traditional machine learning methods, namely, support vector machine, multilayer perceptron, decision tree, k-nearest neighbor. The models used classify gestures obtained open-setting experiment. Amongst all, LSTM-dense method yielded highest overall recognition accuracy (87.38%). Nevertheless, performance other similar range, showing our structure applied intelligence sensing systems. Secondly, prototype analyze made while manipulating interventional robot. analyzed displacement velocity typical axial (push/pull) radial operations results show is capable recording unique patterns operations. Thus, combination wearable could yield future generation artificial things (AIoT) devices.