Advancing Inertial Measurement Unit Technology for Human Biomechanics and Engineering Education

Abstract

Inertial measurement units (IMUs) are a ubiquitous technology found in navigation systems, mobile devices, and multiple products related to the Internet of Things. In its simplest form, an IMU contains a triaxial accelerometer and a triaxial angular rate gyroscope needed to deduce the six degrees of freedom of a rigid body. While their history traces back to navigation systems of aircraft, spacecraft, and satellites, IMUs now support new and innovative applications made possible through miniaturization via microelectromechanical systems (MEMS) fabrication methods. This thesis specifically advances the use of IMU technology within two important fields, namely, 1) human biomechanics and 2) engineering education.

Within the field of human biomechanics, this thesis makes two major contributions for using IMUs to quantify and understand human performance. The first deploys a pair of thigh- and shank-mounted IMUs to estimate the three-dimensional rotations across the human knee. This significant challenge requires a sequence of estimations that define: 1) the orientation of the IMU frames relative to their independent world frames, 2) the orientation of their independent world frames relative to each other, and 3) the orientation of the IMU frames relative to their respective body segment anatomical frames. Importantly, this thesis contributes a measurement theory to correct for the inevitable integration drift error arising in this sequence of estimates without reliance on magnetometer data. The theory exploits an anatomical kinematic constraint that the knee acts (predominantly) as a hinge. The resulting theory is first validated against data from high precision optical encoders embedded within a mechanical linkage and yields RMS differences of less than 5 degrees. The theory is further validated against data from conventional optical motion capture on human subjects (and across increasingly dynamic tasks) and yields overall RMS differences of less than 5 degrees. The second contribution leverages thigh- and upper arm-mounted IMUs to define novel metrics of human crawling performance and technique to support the evaluation of warfighters. Performance metrics derived from the raw IMU data successfully distinguish superior from inferior crawling performance and the degradations in performance from added body-borne loads.

Within the field of engineering education research, this thesis contributes a thorough investigation of an active learning intervention that employs IMUs to explore concepts in an introductory engineering dynamic course (ME240 at the University of Michigan). The intervention takes three forms that elicit increasing cognitive engagement per Chi's ICAP framework, namely: 1) Demonstrations, 2) Prescribed Experiments, and 3) Student Projects. Building from a foundation of supporting literature and learning theories, this research tests the hypothesis that students who engage with the active learning IMU intervention will demonstrate positive responses in 1) conceptual understanding, 2) self-efficacy, and 3) intention to persist relative to students who do not (control). As measured solely by the Dynamics Concept Inventory, the active learning IMU intervention elicits little change in conceptual understanding relative to the control. By contrast, as measured by a modified version of the Longitudinal Assessment of Engineering Self-Efficacy, the active learning IMU intervention elicits significantly higher course-specific self-efficacy and intention to persist in the field relative to the control.