Recent progress in wearable tactile sensors combined with algorithms based on machine learning and signal processing

Abstract

The development of nanomaterials facilitates fabrication of personalized wearable tactile sensors, which are essential components for next generation sophisticated electrical devices, such as smart robotics, robot-assisted surgery, artificial skin, and biomedical devices. Wearable tactile sensors detect various physiologically relevant information from the human body, including mechano-acoustic signatures and precision kinematics. In contemplation to analyze complex superposition of signals with high dimensionality and high frequency, new requirements are put forward for data processing algorithms. The applications of advanced algorithms from machine learning and signal processing greatly boost the performance of the whole tactile sensing system and help redesign the sensor system. Especially, multimodal identification, performed on dataset incorporating different data sources, will be a breakthrough direction in the future. This Perspective highlights the benefits of utilizing advanced algorithms in wearable tactile sensors, summarizes the guidelines for the integration of the algorithm-sensor framework, and proposes potential applications in human body science.