This dissertation develops, implements and analyzes a dynamic control system for a pneumatic muscle actuator (PMA) utilizing an augmented orthosis application. The application of PMAs are limited due to poor control capabilities resulting from dynamic nonlinearities. An adequate control system applying an appropriate dynamic pneumatic muscle actuator model increases the potential utility of PMAs in high-force applications including augmented orthotic applications.
The research conducts an initial analysis evaluating the feasibility of PMAs in high-force applications (force assistance with minimal displacement). A computational simulated control system (CSCS) is developed to analyze three different control schemes. The three PMA control schemes (position feedback, moment/force feedback and adaptive control) are theoretically developed and compared using MATLAB software code. The biomimetic/biomechanical phenomenological model is utilized in the CSCS to characterize the pneumatic muscle actuator. The augmented orthotic application of the physical therapy knee extension task represents the human operator within the CSCS. By implementing the PMA model variations and human operator perturbations, the CSCS is evaluated for each control scheme. The moment/force feedback control outperformed the other schemes by providing accuracy less than ±0.5 degrees.
Finally, the dissertation implements the moment/force feedback control scheme on a physical dynamic test system. The dynamic test system contains a commercially available pneumatic muscle actuator. A comparison between open loop control utilizing strictly the phenomenological PMA model and the closed loop control implementing the moment/force feedback is conducted. Statistical analysis concludes that the closed loop method better controls the PMA dynamic nonlinearities associated with displacement. The closed loop method provides significantly lower root mean square error values for all cases analyzed.
This research develops and implements a PMA control system utilizing the phenomenological model. It provides an adequate control scheme that responds and compensates for PMA nonlinearities. Additionally, this research provides a unique high-force augmented orthotic application compared to conventional low-force applications. It introduces the use of a commercially available PMA allowing the results to be reproduced and compared. Finally, the research implements a dynamic test system providing time-dependent responses. The PMA dynamic control system presented in this research enhances the potential of PMA applications especially in the rehabilitation, assistive, and aerospace fields.