Proceedings of the 7th World Congress of Biomechanics, July 6-11, 2014, Boston, Massachusetts. Eugene, OR: American Society of Biomechanics, 2014 Jul; :438
Many musculoskeletal injuries to the lower extremity are caused by over-exertion during occupational activities such as lifting and carrying a load. Limited accessibility to the workplace prevents practical motion capture systems from studying these injuries in settings where they normally occur. Currently, portable kinematic sensors include inertial measurement units (IMUs) and markerless motion capture (MMC), but each has limitations. Two-dimensional (2D) IMUs provide segment kinematics about the pitch and roll axes; 3D IMUs allow for the measurement of kinematics about the yaw axis, though these can be adversely affected by ferromagnetic objects. MMC utilizes observational cameras, but requires anticipated segment kinematics. We hypothesized that 2D IMUs could provide MMC with the data necessary to return 3D kinematics, and that combining these kinematics with plantar pressures would create a Portable Biomechanical Assessment Suite (PBAS). We constructed a prototype PBAS with 2D IMUs, sparse motion capture, and pressure insoles to serve as a platform for the development of MMC systems. The purpose for this feasibility study was to compare the accuracy of PBAS lumbar (L5/S1) and lower extremity biomechanics to a traditional motion capture system (TMCS). Five subjects (3 male, 2 female; mean age 27 yr, height 1.7 m, mass 66 kg) had full body, six-degree-of-freedom data collected while they performed an occupational task involving lifting and lowering a load of 6.7 kg between the floor and a position level with the sternum. Data for the TMCS were collected using a 14-camera motion capture system and two force plates. The PBAS used pelvis and heel marker trajectories, unilateral (left) segment kinematics collected via 2D IMU's, plantar pressures, and a sensor fusion based upon probabilistic inference methods within Visual3D. Both methodologies calculated joint kinetics using inverse dynamics in Visual3D. For the lumbar (L5/S1), hip, and knee joints, PBAS peak positions and moments, measured during lifting initiation and termination, were similar to TMCS values. Our results suggest lower extremity biomechanics can be quantified using data sets limited to sparse motion capture, 2D segment kinematics, and plantar pressures. Additionally, this prototype platform may lead to future development of workplace MMC systems.
Proceedings of the 7th World Congress of Biomechanics, July 6-11, 2014, Boston, Massachusetts