Effect of pressure insole sampling frequency on insole-measured peak force accuracy during running.

Pressure sensing insoles enable us to estimate forces under the feet during activities such as running, which can provide valuable insight into human movement. Pressure insoles also afford the opportunity to collect more data in more representative environments than can be achieved in laboratory studies. One key challenge with real-world use of pressure insoles is limited battery life which restricts the amount of data that can be collected on a single charge. Reducing sampling frequency is one way to prolong battery life, at the cost of decreased measurement accuracy, but this trade-off has not been quantified, which hinders decision-making by researchers and developers. Therefore, we characterized the effect of decreasing sampling frequency on peak force estimates from pressure insoles (Novel Pedar, 100 Hz) across a range of running speeds and slopes. Data were downsampled to 50, 33, 25, 20, 16 and 10 Hz. Force peaks were extracted due to their importance in biomechanical algorithms trained to estimate musculoskeletal forces and were compared with the reference sampling frequency of 100 Hz to compute relative errors. Peak force errors increased exponentially from 0.7% (50 Hz) to 9% (10 Hz). However, peak force errors were < 3% for all sampling frequencies down to 20 Hz. For some pressure insoles, sampling rate is inversely proportional to battery life. Therefore, these findings suggest that battery life could be increased up to 5x at the expense of 3% errors. These results are encouraging for researchers aiming to deploy pressure insoles for remote monitoring or in longitudinal studies.

Tibial bone forces can be monitored using shoe-worn wearable sensors during running.

Tibial bone stress injury is a common overuse injury experienced by runners, which results from repetitive tissue forces. Wearable sensor systems (wearables) that monitor tibial forces could help understand and reduce injury incidence. However, there are currently no validated wearables that monitor tibial bone forces. Previous work using simulated wearables demonstrated accurate tibial force estimates by combining a shoe-worn inertial measurement unit (IMU) and pressure insole with a trained algorithm. This study aimed assessed how accurately tibial bone forces could be estimated with existing wearables. Nine recreational runners ran at a series of different speeds and slopes, and with various stride patterns. Shoe-worn IMU and insole data were input into a trained algorithm to estimate peak tibial force. We found an average error of 5.7% in peak tibial force estimates compared with lab-based estimates calculated using motion capture and a force instrumented treadmill. Insole calibration procedures were essential to achieving accurate tibial force estimates. We concluded that a shoe-worn, multi-sensor system is a promising approach to monitoring tibial bone forces in running. This study adds to the literature demonstrating the potential of wearables to monitor musculoskeletal forces, which could positively impact injury prevention, and scientific understanding.