Crack-Across-Pore Enabled High-Performance Flexible Pressure Sensors for Deep Neural Network Enhanced Sensing and Human Action Recognition

Citations Over TimeTop 10% of 2022 papers

Abstract

Flexible pressure sensors with high sensitivity over a broad pressure range are highly desired, yet challenging to build to meet the requirements of practical applications in daily activities and more significant in some extreme environments. This work demonstrates a thin, lightweight, and high-performance pressure sensor based on flexible porous phenyl-silicone/functionalized carbon nanotube (PS/FCNT) film. The formed crack-across-pore endows the pressure sensor with high sensitivity of 19.77 kPa-1 and 1.6 kPa-1 in the linear range of 0-33 kPa and 0.2-2 MPa, respectively, as well as ultralow detection limit (∼1.3 Pa). Furthermore, the resulting pressure sensor possesses a low fatigue over 4000 loading/unloading cycles even under a high pressure of 2 MPa and excellent durability (>6000 cycles) after heating at high temperature (200 °C), attributed to the strong chemical bonding between PS and FCNT, excellent mechanical stability, and high temperature resistance of PS/FCNT film. These superior properties set a foundation for applying the single sensor device in detecting diverse stimuli from the very low to high pressure range, including weak airflow, sway, vibrations, biophysical signal monitoring, and even car pressure. Besides, a deep neural network based on transformer (TRM) has been engaged for human action recognition with an overall classification rate of 94.96% on six human actions, offering high accuracy in real-time practical scenarios.

Related Papers

• → Modeling Confinement Efficiency of Reinforced Concrete Columns with Rectilinear Transverse Steel Using Artificial Neural Networks(2003)32 cited
• → Study on Rock Bolt Support of Roadway of Coal Mine Using Neural Network(2013)1 cited
• Artificial Neural Network (ANN) Method for Predicting of Hydrochemical Types of Salt Lakes(2005)
• Prediction Study of Farmer Income Based on the Artificial Neural Network(2005)
• Analysis and Prediction of Steel Price Based on Artificial Neural Network(2010)