Cyber-Enabled Repetitive Motions in Rehabilitation
Abstract:
The project seeks to develop cyber-enabled exercise machines (CEEMs), which will be characterized by: i. intrinsic safety, ii. an extended use of sensing and estimation of biomechanical data, iii. real-time adaptation and guidance to the user to achieve optimal exercise. The project will develop optimality criteria on the basis of collective activation of target muscle groups. The CEEM hierarchical control system will contain machine and biomechanical sensing, estimation, online optimization and impedance modulation subsystems. The optimizer uses model information and sensed data to generate:
1. updates to the machine port impedance,
2. reference trajectories for the machine and
3. cues to the user to modify her mechanical output.
Optimality is obtained whenever the user is able to follow the system-generated cues. Otherwise, the system adopts a safe subpotimal state. Training adaptations are expected to be superior with CEEMs in comparison with fixed-impedance machines. This will be confirmed with human subject tests and two prototypes to be developed as part of the project. The project has the following broad goals:
1. Development of foundational cyber-physical science and technology in the field of human-machine systems
2. Development of new approaches to modeling, design, control and optimization of advanced exercise machines
3. Application of the above results to develop two custom-built machines: a rowing machine and a 2- degree-of-freedom planar machine
4. Dissemination and outreach The following research tasks have been planned to accomplish the above goals:
A. Design concepts for CEEMs and offline optimization of machine parameters
B. Modeling and simulation of coupled biomechanics-machine systems
C. Extended and Unscented H∞ estimation and stability
D. Real-time musculoskeletal state estimation
E. Exercise machine control and adaptation
F. Online micro evolutionary optimization
G. Validation, human subject testing, dissemination and outreach with two custom-built CEEM prototypes.
Intellectual Merit: Recent innovations in exercise machines are characterized by an increased use of sensing and computation. The project will provide a solid foundation to invent machines which simultaneously guide the user and themselves adapt towards optimal exercise. Optimality is sought relative to muscular activation of target muscle groups. This approach is entirely new and contrasts with the use of power as an optimization objective, which has been used in two decades of exercise machine optimization and control research. The generation of optimal cues for the user and consideration of its associated feedback in the analysis is unique to the proposed approach. The development of CEEMs will spur new cyber-physical foundational research encompassing biomechanical modeling and sensing, estimation theory, human performance, control theory and optimization. In addition, human subject tests will be conducted at the CSU Human Performance Laboratory using two CEEM prototypes: a two degrees-of-freedom planar machine and a rowing machine.
Broader Impacts: The project will advance cyber-physical systems at both foundational and applied levels. The invention of CEEM systems will bring benefits to the research community, will support and enhance the university’s engaged learning mission and will have an impact on casual and athletic physical conditioning methods and results. CEEMs can be used by athletes to maximize training for optimal performance. CEEMbased training can also help eliminate training injuries by programming custom loads. CEEM systems could also be utilized by beginners and older populations wanting to improve strength, and for rehabilitation of patients with musculoskeletal problems. Due to its high degree of reconfigurability, CEEM technology could also be used in microgravity environments. Cleveland State University is a leader in enrolling and graduating minorities, and has received the Higher Education Excellence in Diversity (HEED) Award in 2014 and 2015. This will enable the PIs to be successful in recruiting students to participate in the project. Research products will be disseminated through scholarly publications and through publicly-available software codes.