1、Background：
With the increasing aging population, high incidence of chronic diseases, and the growing number of congenital or acquired foot deformities, lower limb dysfunction and abnormal gait problems are becoming increasingly common, posing a significant threat to public health and quality of life. Gait analysis is widely considered a sensitive biomechanical indicator for evaluating lower limb function, disease progression, and rehabilitation effectiveness. However, existing clinical gait assessment primarily relies on laboratory equipment such as optical motion capture systems and force platforms, which are not only expensive and spatially constrained but also failing to reflect natural movement in real-life scenarios.
Wearable pressure-sensing insoles offer a decentralized and continuous new approach to gait monitoring, but existing technologies still face three major bottlenecks in clinical translation: firstly, sensors struggle to simultaneously achieve ultra-low pressure resolution and high load tolerance, making it difficult to cover the full biomechanical range of the sole, from subtle postural adjustments to violent impacts; secondly, energy supply relies on traditional batteries, resulting in insufficient battery life and frequent charging, which hinder the continuity of long-term monitoring; thirdly, the large-scale spatiotemporal pressure data collected lack effective intelligent analysis and real-time feedback, limiting its application in disease screening and clinical decision-making. Therefore, developing a wearable gait monitoring system that integrates high-precision sensing, autonomous power supply, and intelligent diagnosis is of great scientific significance and clinical value.
2、Research Progress：
This study reports a biomimetic smart insole system that, through multidisciplinary collaborative design, achieves high-resolution plantar pressure sensing, energy self-sufficiency, and artificial intelligence-assisted gait intelligent diagnosis. Inspired by the hierarchical mechanosensory structure of the mantis leg, the research team designed a dual-microstructure capacitive pressure sensor, combining microstructured PDMS with compressible elastic foam. This achieves an ultra-low detection limit of 0.10 Pa, a wide detection range up to 1.4 MPa, and maintaines excellent mechanical stability over 12,000 loading cycles, significantly outperforming existing flexible pressure sensors and fully meeting the requirements for insole applications.
In terms of the energy system, the smart insole integrates a perovskite solar cell and a high-energy-density lithium-sulfur nanobattery, constructing a closed-loop, adaptive energy supply system. It can operate stably under various indoor and outdoor lighting conditions, with an average light charging efficiency of 11.21% and an energy storage efficiency of 72.15%, effectively addressing the energy bottleneck for long-term continuous operation of wearable devices.
At the data processing level, the system collects plantar spatiotemporal pressure distribution through a 16-channel wireless module and embeds artificial intelligence algorithms for real-time analysis. Based on a random forest model, the system can achieve 96.0% accuracy in identifying arch abnormalities; based on a one-dimensional convolutional neural network (1D-CNN), it can classify 12 pathological gait patterns with an accuracy of 97.6%. The accompanying mobile app intuitively presents the dynamic force field distribution through color maps, providing interpretable and real-time decision support for clinicians and rehabilitation personnel.
3、Future Prospects
By deeply integrating biomimetic high-precision sensing, sustainable energy interfaces, and intelligent mechanical diagnostics, this research has constructed a clinically validated closed-loop wearable platform, providing a novel technological pathway for early screening of lower limb diseases, personalized rehabilitation training, and remote medical monitoring. This demonstrates the broad prospects for the transformation of intelligent wearable devices into clinical-grade diagnostic tools.
The complete study is accessible via DOI: 10.34133/research.1063