Learning Sensorimotor Control of Balance and Locomotion

点击翻译按钮获取中文摘要

9812591SpongThe goal of this project is to investigate new computational methods for learning sensorimotor control in bipedal locomotion. The overarching theme of the project is the integration of cross-fertilization of ideas from engineering, psychology, and kinesiology. The PI's will utilize techniques from control theory and artificial intelligence to improve their understanding of the dynamics and control of human movement and the mechanisms by which humans learn sensorimotor control, while at the same time utilize studies of human movement to aid the development of improved learning and control techniques for multi-degree-of-freedom mechanical systems. Further applications of their computational methods might include more dextrous and useful robots and more effective diagnostic and physical therapy approaches for disabled humans, as well as better balance training and falls prevention programs for elderly and individuals with balance deficits.They have assembled a team of researchers with expertise in robotics, nonlinear and hybrid control theory, discrete event dynamical systems, machine learning and artificial intelligence, computer vision, developmental psychology and perceptual-motor coordination to explore these research issues from both the robotic and the human side and to integrate concepts from these diverse disciplines into a coherent theory of learning control in multi-degree-of-freedom anthroopmorphic systems.They will study the problems of postural control, balance, gain initiation, and gait transition. The goal will be to develop models of the feedback control and learning mechanisms involved first in the control of balance and then in locomotion. The models will be developed using data generated from human movement studies and will be based on recent techniques of hybrid control theory, such as supervisory control and logic-based switching control. Machine learning techniques of artificial intelligence will be used to learn switching control strategies from observed data. The novelty of the research lies in:- The integration of human movement studies with analytical studies based on the most recent concepts of nonlinear dynamics and control theory, hybrid control theory, and discrete event systems theory.- The pre-eminent role attached to sensing for balance control and control of walking.- The use of machine learning methods of artificial intelligence to learn hybrid and switching control strategies.- The combination of experimental and analytical research. ***