Self-Powered Smart Insole for Monitoring Human Gait Signals

Wei Wang, Junyi Cao, Jian Yu, Rong Liu, Chris R. Bowen, Wei-Hsin Liao2019 · DOI

TL;DR

Researchers have developed self-powered smart insoles that harvest energy from human motion using piezoelectric nanogenerators. These insoles can simultaneously monitor gait signals, distinguishing between normal walking and limping with high accuracy using multiscale entropy analysis. This work paves the way for more sustainable and informative wearable health monitoring systems.

What Problem Does This Paper Address?

The proliferation of low-power wearable electronics for health monitoring is hampered by the need for frequent battery charging or replacement. While energy harvesting technologies offer a solution, integrating them with flexible and comfortable wearable devices that can also perform sensing functions remains a challenge. Specifically, monitoring human gait signals, which are crucial for assessing health conditions, requires a self-powered system that can reliably capture subtle variations in movement. Existing approaches often focus on either energy harvesting or gait monitoring, but rarely combine both effectively in a practical, wearable form factor.

Methods

The core of this work lies in the fabrication of a novel self-powered smart insole. The energy harvesting and sensing capabilities are achieved through piezoelectric nanogenerators (NGs) made from poly(vinylidene fluoride) (PVDF) films. Aluminum electrodes are sputtered onto the PVDF to create the NGs.

The smart insoles themselves are constructed using advanced 3D seamless flat-bed knitting technology. This allows for the integration of the NGs into specific locations (forefoot and hindfoot) within a comfortable and durable insole structure. The knitting process utilizes a combination of polyamide elastomers and polyester fibers to ensure good tactile perception, shape retention, and air permeability.

For gait monitoring, the voltage signals generated by the NGs during walking are analyzed. The key technique employed is multiscale entropy (MSE) analysis. This method quantifies the complexity and regularity of time series data. By applying MSE to the stride intervals extracted from the voltage signals, the researchers aimed to differentiate various walking states. The specific MSE implementation used an overlapping multiscale method with parameters m=2 (embedding dimension), τ=1 (time delay), and a threshold value of r×SD (where r=0.15 and SD is the standard deviation of the time series).

Datasets

Not discussed in the paper. The experiments involved two healthy human subjects walking on a treadmill under controlled conditions.

Key Insights and Ideas

  • Seamless Integration of Energy Harvesting and Sensing: The paper demonstrates a practical approach to creating a single device that both harvests energy from human motion and uses that energy to power its own sensing capabilities, eliminating the need for external power sources.
  • Comfortable and Durable Wearable Design: The use of 3D knitting technology results in a smart insole that is likely to be comfortable and durable, crucial for long-term wear and continuous monitoring.
  • Gait Discrimination via Multiscale Entropy: The application of MSE to stride intervals proves to be a powerful tool for distinguishing subtle differences in gait, particularly between normal walking and limping. This highlights the potential for non-invasive health diagnostics.
  • Influence of Load Resistance on Output: The study clearly shows how external load resistance significantly impacts the output voltage of the piezoelectric nanogenerators, a critical consideration for designing efficient energy harvesting circuits.

Key Results

  • Energy Harvesting Performance: The piezoelectric nanogenerators achieved a maximum open-circuit voltage of 41 V and a maximum output power of 168.1 µW.
  • Charging Characteristics: The charged voltage of a capacitor was found to decrease with increasing load resistance and capacitance. Conversely, increasing motion speed led to higher charged voltages.
  • Gait Monitoring Accuracy: Multiscale entropy analysis of stride intervals successfully differentiated between slow, normal, and fast walking and walking with a limp. Specifically, walking with a limp showed lower complexity (smaller MSE values at larger scales) compared to normal walking states.
  • Durability: The output performance of the nanogenerators remained consistent even after multiple measurements, demonstrating their durability.

Limitations and Future Work

  • Limited Subject Pool: The study involved only two healthy subjects. Future work should include a larger and more diverse group of participants, including individuals with various gait abnormalities and conditions.
  • Controlled Environment: Experiments were conducted on a treadmill. Testing in real-world, unconstrained environments would be a valuable next step to assess the system's performance under more dynamic conditions.
  • Further Gait Analysis: While limping was successfully identified, exploring the detection of other specific gait disorders or subtle changes in walking patterns could further enhance the clinical utility of the smart insoles.
  • Optimization of Energy Harvesting Circuitry: Further research could focus on optimizing the power management and energy storage components to maximize the harvested energy for powering more complex sensing and communication systems within the insole.

Summary updated

BibTeX
@misc{wang2019self,
  author = {Wei Wang and Junyi Cao and Jian Yu and Rong Liu and Chris R. Bowen and Wei-Hsin Liao},
  title = {{Self-Powered Smart Insole for Monitoring Human Gait Signals}},
  year = {2019},
  doi = {10.3390/s19245336},
}
Figure