User-friendly IMU-based Human Motion Measurement

William Marsh Rice University, Houston TX

Investigators

Abstract

PI: Fregly, Benjamin J. Proposal: 1805896 Stroke, Parkinson's disease, and osteoarthritis affect roughly 15% of the U.S. adult population and often impair walking ability, leading to an increased risk of serious health conditions and a decreased quality of life. Traditional approaches for assessing patient function and delivering rehabilitation prescriptions require that the patient make repeated visits to the clinic, which is expensive and time consuming for the patient, the clinician, and the health care system. Ideally, clinicians could gather quantitative information about patient walking function, as well as deliver rehabilitation prescriptions, when the patient is at home or in the community. Wireless inertial measurement units (IMUs), which are electronic devices that use accelerometers to measure linear acceleration (along a straight line) and gyroscopes to measure angular velocity (rotation), are already being explored to provide such capabilities. However, existing computational algorithms (problem solving instructions) that convert IMU measurements into joint angle measurements are inadequate to support an IMU-based motion measurement system that provides accurate joint motion measurements AND is so simple that patients, family members, or caregivers can use it reliably with no technical expertise. This project seeks to develop a novel computational algorithm that achieves these goals. The algorithm will achieve accuracy and simplicity by combining several unique computational approaches. The proposed project will 1) DEVELOP the necessary computational algorithm using preexisting motion capture data collected simultaneously, 2) EVALUATE the algorithm's ease of use and accuracy by having 10 healthy subjects collect their own IMU-based motion capture data, and 3) DEPLOY the algorithm on a Windows tablet or Mac laptop to demonstrate that it can be used for real-time feedback applications. If successful, this project could facilitate the development of effective remote monitoring and telerehabilitation methods that reduce the need for clinician time and patient office visits, decrease healthcare costs, and increase treatment effectiveness. For outreach, "at risk" students in a Houston middle school will interact with the technology firsthand by competing in an annual "Jump Off" competition held via the internet (Skype) with students in a sister middle school in Auckland, New Zealand. The goal of this project is to develop a wireless user-friendly IMU-based motion measurement system that is so simple and easy to use that virtually any patient or caregiver can use it to measure full-body walking movements at home or in the community with little training. Such a system could be clinically beneficial for both remote assessment of patient function using logged data and remote delivery of rehabilitation prescriptions using real-time feedback. The core of the stem is a novel motion estimation algorithm that will work for data logging and real-time applications. The algorithm will require only 6 IMUs-- each foot, pelvis, torso and each wrist--to measure full-body walking kinematics easily and repeatably. The project will progress in three phases. Phase 1 will involve algorithm development. The algorithm will be an expanded version of an unscented filter (UF) algorithm developed for studying an existing planar 6 DOF walking model. The first task will be to make that model more general and flexible by adding more coordinates and extending it to 3D. The algorithm will be tested using experimental walking data and data obtained from an existing 29 DOF walking model representative of real-life conditions. Phase 2 will involve algorithm evaluation for off-line applications. To demonstrate that, with minimum training, individuals in any environment can collect their own kinematic walking data using the algorithm, 10 healthy subjects will collect their own IMU-based walking data while also wearing markers to enable simultaneous marker-based motion capture to validate results. Phase 3 will involve algorithm deployment for real-time applications. The computational speed of the algorithm when deployed on hardware for real-time applications will be evaluated. A wireless IMU based motion feedback system, consisting of 6 wireless IMUs and a Windows (or Mac) tablet control unit with software for real-time calculation and display of joint angles during walking, will be developed. To verify that the system meets the project's goals --ease of use, rapid calibration, and collection and display of calculated joint positions in real time, 3 additional subjects will perform real-time data collection on themselves without wearing markers. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.