Hierarchical Synergistic Engineering for Machine Learning-Assisted Gesture Recognition and Integrated Thermal Management

Abstract

Flexible strain sensors are revolutionizing human–machine interactions and next-generation health care by enabling real-time monitoring of human motion and precision medical treatment. However, developing lightweight flexible strain sensors that combine high sensitivity with a broad monitoring range remains a significant challenge. To address this challenge, an advanced structural engineering strategy based on the sodium chloride (NaCl) template sacrificial method is employed to simultaneously increase sensitivity and mechanical robustness. By leveraging a NaCl template sacrificial method, a hierarchical synergistic conductive network is constructed within the thermoplastic polyurethane (TPU) matrix formed via in situ growth. This design enables ultra-high sensitivity across a broad strain range, offering promising potential for wearable sensing applications. The resulting sensor exhibits exceptional performance characteristics, including a low detection limit (0.176%), high sensitivity (gage factor, GF = 331.7), wide sensing range (up to 230.1%), rapid response/recovery times (133 ms/133 ms), and remarkable durability exceeding 4000 cycles. Furthermore, the sensor demonstrated excellent electrothermal conversion performance with a positive temperature coefficient of 0.00207 °C⁻¹ and an achievable saturation temperature of 54.2 °C (1.0 A). Finally, the sensor was successfully integrated into a smart wearable system, enabling precise recognition and classification of multiple gestures through machine learning algorithms while also exhibiting significant potential for inflammation hyperthermia therapy.

Graphical Abstract