Advanced Technology for Quantifying and Optimizing Human Physical Performance

Human performance is defined as physical performance, cognitive performance, and social/organizational performance. This project specifically focused on physical performance. The research team spent a great deal of time with practitioners within the tactical athlete (i.e., U.S. Special Operations) and professional sports physical performance communities, and discovered that there is a strong desire to use 3-D motion capture analysis for biomechanical assessment; however, current practices and technologies are not able to meet this need (either too labor intensive [marker-based] or measurement accuracies are too low [markerless-based]). Thus, the objective of this research effort was to design, integrate, and evaluate the performance of a novel model-based markerless motion measurement system.

SwRI developed a new markerless 3-D motion capture system that provides measurement accuracies comparable to traditional marker-based 3-D motion capture systems, but requires a minimal setup time that aligns with the operational needs of the human performance community. The new SwRI markerless biomechanics system was developed by combining biomechanical modeling, deep neural networks, and sensor fusion techniques. Some advantages of this new system are: Commercial-off-the-shelf video components are used compared to specialized infrared cameras used with traditional marker-based systems A minimal number of cameras are required to capture 3-D motion compared to traditional systems (two cameras vs. eight or more) No external markers are required to be placed on the subject compared to up to 60 markers when using a traditional system The system can be extended to capture multiple subjects The methodology will allow for a complete biomechanical analysis using a single system compared to multiple systems (e.g. marker-based motion capture, force plates, separate data analysis and modeling codes) using traditional methods.

This research project was successful in accomplishing the goal of researching and developing an SwRI-developed markerless 3-D motion capture system that provides measurement accuracies comparable to marker-based 3-D motion capture systems, while requiring a minimal setup time to align with the operational needs of the human performance community. The results of this project have defined a framework for a paradigm-shifting foundational technology for biomechanical human performance measurement. We anticipate that with further development, this technology will be widely adopted across multiple disciplines that can benefit from the accurate quantification of human motion.