Wearable Sensor Detection of Treadmill-Induced Slip Perturbations (2023)

Undergraduate: Jordan Feldman

Faculty Advisor: Jason Franz
Department: Joint Department of Biomedical Engineering

---

Monitoring instability is critical for the prevention of dangerous falls. Recent developments in the use of wearable sensors have shown promise in their application to monitor physical activity. However, it is unclear whether and the extent to which wearable sensor thresholds for detecting locomotor instability resemble the biological threshold for perceiving that instability, even in otherwise healthy younger adults. This study in young adults aimed to build on previous studies using small-amplitude treadmill-induced slip perturbations to establish and objectively compare a conscious threshold for perceiving instability and wearable-sensor based thresholds for detecting that instability. We hypothesized that wearable sensors have the capacity to detect the instability elicited by treadmill-induced slip perturbations, with thresholds for conscious perception that will be indistinguishable from thresholds for wearable sensor detection. 15 young adults were fit with seven wearable inertial measurement unit (IMU) sensors. While walking, a treadmill controller rapidly decelerated one foot at initial contact during the stance phase at random. Eight perturbation magnitudes delivered in randomized order five times each ranged from 0.02 m/s to 0.3 m/s. After each perturbation, subjects were asked to respond “Yes” or “No” if they felt a balance disturbance. Ultimately, acceleration signals alone perform worse than conscious perception in detecting the instability elicited by treadmill-induced slip perturbations in healthy younger adults. However, IMU-derived ankle joint kinematics are highly sensitive to even the smallest amplitude perturbation. Continued work in this area seeks to investigate the generalization of these conclusions to individuals more representative of those at risk of falls.

Link to Poster