Researcher(s)
- Kate Menzer, Biomedical Engineering, University of Delaware
Faculty Mentor(s)
- Jocelyn Hafer, Kinesiology and Applied Physiology, University of Delaware
Abstract
Changes in environment often cause changes in gait that are unseen in typical lab settings. Researchers study out-of-lab gait to understand how gait truly changes. Wearable inertial measurement units (IMUs) are useful for evaluating out-of-lab gait because of their portability and ability to measure gait-related signals. Gait analyses usually require an IMU on both sides of a joint of interest and on the foot for identifying gait events. Minimizing this considerable number of sensors needed would reduce the burden on participants. Therefore, the purpose of this study was to investigate various IMU placements to minimize sensors while maintaining accurate and meaningful gait measurements. Ten healthy young adults (5 female, 25.5±2.8 years) walked a self-selected speed down a ~85-meter hallway. Six IMUs were placed on the sacrum, lateral aspect of the midpoint of both the right thigh and lower leg (typical placements), medial and lateral right lower leg near the ankle (to test against typical placements), and right dorsal foot. The outcome variables of interest were number of gait events detected, stride length, stride velocity, and knee range of motion (ROM) calculated using each of the lower leg sensors. One-way ANOVAs determined if sensor placement affected outcome variables. Post-hoc t-tests assessed differences between sensor placements. The typical lower leg sensor identified fewer gait events than the foot sensor (p<0.05). The foot and lower leg sensors differed significantly for stride velocity and length (both p<0.001). The foot sensor had a higher stride velocity and length compared to the typical and medial ankle sensors (all p<0.05). For knee ROM the typical sensor was significantly greater than both the lateral (p=0.02) and medial (p=0.03) ankle sensors. Overall, the lateral ankle sensor most closely compared to the foot sensor for all outcomes. These findings demonstrate the possibility for minimizing the required sensors, increasing participant comfort.