# Mantis Leg-Inspired Smart Insole Integrating Closed-Loop Power Supply for Advanced Wearable Gait Diagnostics

## TL;DR

A smart insole inspired by mantis legs improves gait monitoring with precise sensors, sustainable power, and AI for diagnosing foot issues.

## Contribution

A fully integrated smart insole with bioinspired sensors, solar-powered energy system, and AI for gait diagnostics.

## Key findings

- Dual-microstructure capacitive sensors detect pressures from 0.10 Pa to 1.4 MPa with high stability over 12,000 cycles.
- Nano-perovskite solar cells and lithium–sulfur nanobatteries achieve 11.21% photocharging and 72.15% energy storage efficiency.
- AI algorithms detect foot arch abnormalities with 96.0% accuracy and classify 12 gait patterns with 97.6% accuracy.

## Abstract

Precise diagnosis and management of lower extremity dysfunction disorders hinge on continuous gait monitoring. Nevertheless, the existing wearable devices fall short as they grapple with insufficient sensing precision, inadequate energy endurance, and ineffective intelligent data analysis. Here, we report a fully integrated, biomimetic smart insole that incorporates 3 synergistic innovations to overcome these challenges. First, inspired by the hierarchical mechanosensory apparatus of mantis legs, we design dual-microstructure capacitive sensors with a detection limit of 0.10 Pa and a maximum detection range of 1.4 MPa. This sensor can distinguish pressures across a wide range from subtle to substantial and exhibits robust mechanical stability over 12,000 cycles, making it highly suitable for insole applications and outperforming current flexible pressure sensors. Second, we realize energy-autonomous operation by integrating nano-perovskite solar cells with high-capacity lithium–sulfur nanobatteries, achieving an average photocharging efficiency of 11.21% and energy storage efficiency of 72.15%. Third, embedded artificial intelligence algorithms interpret the spatiotemporal pressure data transmitted via a 16-channel wireless module. These models achieve 96.0% accuracy in detecting foot arch abnormalities and 97.6% accuracy in classifying 12 pathological gait patterns. Collectively, these 3 advances, including bioinspired high-resolution sensing, sustainable energy interfacing, and intelligent mechanodiagnosis, establish a closed-loop wearable platform validated in clinical studies. This system offers promising applications in early disease screening, personalized rehabilitation, and remote healthcare.