This CAREER award seeks to develop an integrated research and educational foundation in nonlinear control methods for artificial electrical stimulation of human skeletal muscle. Current clinical practices for treating medical disorders through the application of electrical stimulation are based on patient independent protocols for the open-loop adjustment of stimulation parameters (e.g., amplitude, phase, and pulse duration). Research focused on closed-loop electrical stimulation is based on either pure feedback control or feedback control with the addition of a feedforward component composed of soft computing elements (e.g., artificial neural networks, fuzzy logic sets). Unfortunately, these methods do not fully exploit the potential benefits of on-going research focused at more effectively modeling muscle response. Research efforts in this project will focus on the development and experimental validation of nonlinear control methodologies that incorporate muscle models in the design and analysis as a means to achieve a predictable response despite intra- and inter-subject variability and fatigable force production capabilities of the muscle. Outcomes of this research will advance knowledge in artificial electrical stimulation research through the development of new mathematical models that can be used to alter stimulation parameters to reduce fatigue. These models will be incorporated in the first ever use of Lyapunov-based methods as a means to encapsulate muscle response phenomena in a neuromuscular electrical stimulation controller. Lyapunov-based methods will be used to design controllers that exhibit input saturation and are adaptive to time-varying inter- and intra-subject variations, yielding a customizable neuroprosthesis. There are direct outcomes of the research efforts that impact society through the clinical applications and the scientific advancement of modeling and control research, but further outcomes will result by the exposure of students to human-machine interaction technologies and the impacts of science and engineering for the treatment of disease and disability.
Neuromuscular electrical stimulation (i.e., the application of an electrical current via internal or external electrodes which results in a muscle contraction) offers an enormous promise to treat or alleviate the debilitating morbidity associated with certain diseases and dysfunctional conditions. Neuromuscular electrical stimulation is currently prescribed to treat a wide range of disorders and has rapidly grown because of the potential improvement in the activities of daily living for individuals with movement disorders such as stroke and spinal cord injuries that affect over one million Americans annually. In addition to serving as a treatment, neuromuscular electrical stimulation can provide an artificial extension to the body (i.e., a neural prosthetic) to restore or supplement function lost due to disease or injury with the goal of reducing the resulting implications on society and improving the quality of life of individuals. An attraction of such a neuroprosthesis is that the bodys own muscles are used to restore movement, leading to significant secondary benefits such as reducing muscle atrophy and the associated changes in metabolism and reduced risk for heart failure, the leading cause of death among individuals with spinal cord injuries. Unfortunately, current commercial electrical stimulation products have yielded limited functional outcomes for patients because ad hoc stimulation strategies are used that are not patient specific and lead to rapid muscle fatigue. The goals of this project are to develop new electrical stimulation controllers that (1) would enable long periods of physical activity that (2) could be customized and adapt to an individuals ever changing musculoskeletal system. To achieve the goals, efforts will focus on bridging the current gap between biological physiology research and control engineering research by incorporating neuromuscular models in novel adaptive control designs that elicit a desired muscle response. Given the potential benefits to the quality of life of an individual and the impacts on society of such emerging research, the worldwide market for neurotechnology products including motor system neuroprostheses, neuromodulation devices, and therapeutic muscle stimulators is predicted to grow from $2.4 billion in 2004 to $7.2 billion by 2008.
This project was motivated by the desire to enable computer controlled motion of a person's limb for rehabilitation potential purposes. There is a desire to increase the intensity and duration of rehabilitative treatments to more rapidly return individuals with injury or other neurological impairments or disease to activities of daily living and thereby reduce secondary complications and healthcare costs. A fundamental barrier present in current electrical stimulation treatments is that muscle response to external electrical stimulation is nonlinear and uncertain and leads to a rapid onset of muscle fatigue during repeated contractions. The onset of muscle fatigue during electrical stimulation is strongly correlated with stimulation parameters such as intensity, frequency, and pattern of stimulation. Intellectual merits in this project were inspired by issues resulting from trying to yield accurate control of human muscle dynamics in the presence of uncertainty in the muscle response to stimulation and muscle fatigue. Software was developed and implemented to enable a computer to adaptively adjust the intensity and the electrical pulse shape that was applied through external electrodes to the legs of groups of volunteers. The outcome was accurate limb motion for a longer duration. To achieve this result, feedback from a sensor positioned at the volunteer’s knee was provided to new algorithms that were developed and encoded in software. The family of algorithms that were developed and tested on volunteer subjects yielded breakthroughs in terms of how the electrical stimulation was adjusted by the computer interfaced stimulation unit to compensate for uncertainty and disturbances in the muscle response and the uncertain reaction time to the stimulation. Broader outcomes resulted in this project in new control engineering principles, clinical rehabilitation practice, and student development. The development of more effective electrical stimulation treatments has a large societal impact in terms of the activities of daily living and morbidity of millions of people with neurological disorders/injury and disease. The developed algorithms and experimental tests illustrated new stimulation protocols that were demonstrated to yield a longer duration of stimulation while still maintaining accurate limb motion. Such results were published in clinically focused journals. New control algorithms were required due to the unique dynamics of the muscle response and the fact that the muscle response changes with fatigue. The development of these algorithms required new breakthroughs in robust and adaptive control design which were disseminated to the control engineering discipline through journal and conference publications. To achieve these results a number of undergraduate and graduate students gained new training in control algorithm development, design and construction of a new computer controlled stimulation machine, and the methods to safely implement experiments on volunteer individuals in a manner that leads to quantifiable clinical outcomes. As an outreach effort, this project also engaged graduate students (in particular, minority engineering students) to teach elementary and middle school children a series of robotics camps. The camps exposed the children to robotics and automation, engineering, and provided an exposure that a broad set of individuals are involved in engineering (i.e., underrepresented groups).
