Development of Artificial Muscle Actuator for Physical Joint Simulator

While great advances have been made into understanding the biomechanical environment within human joints, large gaps in our knowledge of the basic mechanics still exist. These gaps are a direct result of limitations in our ability to experimentally characterize detailed internal joint mechanics during in vivo human motion. This lack of fundamental knowledge underlies a multitude of basic and applied research areas in musculoskeletal biomechanics. For example, a comprehensive understanding of internal human joint mechanics is critical to deciphering the underlying mechanisms leading to the onset and progression of osteoarthritis, determining the optimal course of treatment for repairing torn ligaments, developing artificial ligaments that restore normal joint function, designing joint braces, developing optimal rehabilitation strategies, developing better prosthetics, improving treatment strategies for muscle injury, etc. Studying the mechanics of joints presents a series of technical challenges that severely limits the type of experiments that can be performed and the data that can be collected. First and foremost is the challenge of experimentally collecting data from inside the joint while replicating natural boundary conditions and realistic muscle forces. Muscle forces are a significant driver of the mechanics within a joint, and current joint simulators are unable to replicate the complex multicomponent force generation of natural muscle.

The objective of this project was to develop and characterize a biomimetic artificial muscle utilizing thermally actuated polymer artificial muscle (TPAM) fibers. In pursuit of this objective a methodology for consistently and accurately producing the individual TPAM fibers was developed. These TPAM fibers were then tested in different mechanical conditions (various loads, strains, activation currents, etc.) to determine their performance, repeatability, and limitations. Next, control strategies were developed to allow for the scalable, coordinated control of multiple TPAM fibers working together. These strategies were inspired by the methodologies that natural muscles use to coordinate the activation of their individual muscle fibers, allowing groups of actuators to accomplish tasks that they could not do alone. Lastly, this control strategy was incrementally scaled up to a system of 100 individual TPAM fibers working together to lift 32.5 lbs. This corresponds to the amount of force the bicep has to exert to perform a curl exercise with a 5-lb weight.

A semi-automated methodology for TPAM fiber production was developed and used to produce all the TPAMs used over the course of the project. During the characterization of the individual actuators several significant limitations of TPAMs were discovered. Most significant of these limitations is the variability in the supply of conductive nylon sewing thread used as raw material in the production of TPAMs. Using material from different supply lots produced performance variations that far exceeded the inherent variation that was characterized between fibers with material from the same lot. Despite the limitations that were discovered, we were able to successfully create and control systems of individual TPAM actuators whose capabilities far exceeded that of the individual components. We were able to successfully control 100 individual TPAM actuators working together to lift 32.5 lbs, compared to an individual TPAM actuator, which can lift approximately 1/3 lb. Even with the limitations that were discovered over the course of this project, TPAMs have the potential to have a major impact on biomechanics research. With the knowledge gained we believe that we can take this technology combined with our other work in 3-D printing of biomimetic human surrogates and make functional artificial muscles relevant for biomechanics research.