Abstract
Clinically oriented Inertial Measurement Unit sensor systems consisting of several sensors tethered to a mobile device are emerging as a powerful tool for continuous monitoring of gait in the community setting. Inertial Measurement Units provide indirect information regarding movement of the lower limbs and require gait parameter estimation algorithms to transform their raw kinematic output into physiologically meaning gait parameters. Mobile IMU applications operate under battery life, sensor number, and bandwidth limitations that constrain the information available to use as inputs. This research investigated whether there is sufficient information in lower limb angular velocity signals acquired at low sampling rates to track human motion. Novel algorithms that use gyroscope data from the shanks and thighs and biomechanical models of gait to determine single and double limb support times, gait cycle time, step length, stride length, gait speed, and knee angle were proposed. Concurrent validity of these algorithms with widely accepted criterion measure systems was assessed through Bland Altman analysis and found to be accurate and precise, demonstrating that clinically relevant gait parameters can be captured outside the constraints of the traditional gait lab using IMUs tethered to mobile systems.