II-EN: Acquisition of Sensors and Displays for Research on Motion Synthesis and Rehabilitation

University Of California - Merced, Merced CA

Investigators

Abstract

Real-time motion-capture and interactive 3D computer-generated environments are rapidly emerging as an integrated and powerful human-computer interaction context with the ability to revolutionize applications in many areas. Recent advances in sensor and display technologies can now be seen in both low-cost consumer-oriented and in high-end industrial-oriented human-centered computer systems. This proposal aims to support and enhance research and educational activities in new research and application areas by strengthening and expanding the available research infrastructure at the University of California - Merced, specifically by enhancing an existing visualization and motion capture facility with the acquisition of (a) a high-end data glove, (b) an occlusion-free motion capture suit, and (c) portable interactive displays integrated with Kinect sensors. The proposed equipment enhancement will allow the development of new research projects on the following topics: (a) gesture synthesis models with coordinated hand-arm motions, (b) remote virtual collaborative sessions for upper body motion rehabilitation assessment and remote therapy delivery, (c) automated progress monitoring of hand motor rehabilitation in respect to given motion protocols, and (d) human-like full-body motion planning in tight spaces for training applications. The projects will target interactive training and therapy applications that can benefit from new human-computer interfaces that are situated in virtual environments. Intellectual Merit An interdisciplinary team from computer science and cognitive science will explore novel human-centered computing research with a focus on the channels of communication that are used in parallel during physical training and therapy. The proposed portable and occlusion-free high-end motion sensing equipment will allow the development of human-like motion models for synthesis of gestures as well as full-body task-oriented motions. These models will enable, during immersive remote interactions, the discovery of new interaction paradigms, which will involve autonomously synthesized human-like motions and human motions that will be captured and streamed back to the system. A gesture analysis tool will be developed to provide algorithmic infrastructure to automatically analyze and extract the components of the gestures. The infrastructure will permit rehabilitation and therapy to be evaluated in actual practice. The scientific and practical research projects will advance the effectiveness of virtual humans in diverse educational, training, and therapeutic applications, and will help make such 3D computer-generated environments more accessible to the common user. Broader Impact The infrastructure will support research projects that relate to motion rehabilitation, which has the potential to benefit the many Americans who are in need of physical therapy such as for ailments in the back, neck, and shoulders. The research will advance an understanding of a principled approach for using virtual humans for a broad range of practical training and therapy applications. The instrumentation and research will provide unique learning opportunities for graduate and undergraduate students. The research outcomes will be disseminated broadly in cognitive science, computer science, and rehabilitation therapy publications. The infrastructure and research will be demonstrated to local high school and community college students to motivate interest in science and engineering.