Estimating Internal Biomechanical Forces During Running Using Consumer-grade Wearable Sensors

Running injuries are caused by cumulative damage induced by internal biomechanical forces. Estimating these forces usually requires musculoskeletal modeling, which is not feasible for monitoring training loads in the real world. PURPOSE: To identify sensor-measured gait metrics associated with internal forces during running that may be used in future work to predict internal forces using consumer-grade wearable sensors. METHODS: Six runners to date completed a 38 min treadmill run at 70%-125% preferred speed (3.09-4.02 m/s) equipped with consumer-grade chest- and shoe-worn sensors. We recorded 3D-synchronized force, motion, and sensor data. Lab estimates of Achilles and patellar tendon force were calculated using a musculoskeletal model driven by kinematics/kinetics from ten gait cycles across 12 speeds. We used functional data mixed models to estimate the independent effects of the chest and shoe sensor-measured speed, cadence, and vertical oscillation on Achilles and patellar tendon forces across the gait cycle. RESULTS: Maximum Achilles and patellar tendon forces increased by 0.132 and 0.096 BW per 0.1 m/s increase in speed, respectively (both p < 0.0001). After adjusting for running speed, decreased cadence and increased vertical oscillation were associated with greater internal tendon forces (all p < 0.0001). Similar results were observed for both the chest and shoe sensors. CONCLUSION: Gait metrics measured by consumer-grade chest and shoe-worn sensors track well with internal loads, showing promise for predicting these loads during real-world training.Figure 1: Functional regression model estimates of the variation in internal forces across the gait cycle in the Achilles (top) and patellar tendons (bottom) at a range of sensor-measured speeds, cadences, and vertical oscillations. Cadence and vertical oscillation plots show estimates with speed at a constant 3 m/s.