Collaborative Research: HCC: Medium: Learning to coordinate between human and a robotic prosthesis for symbiotic locomotion

点击翻译按钮获取中文摘要

Individuals with lower limb amputation depend on assistive devices such as prostheses to restore basic mobility in daily living, and emerging robotic lower limb prostheses are powered and programmable, showing great potential for improved functionality. Yet, the expected benefit of these modern devices has not yet been fully realized, partly because of a lack of seamless coordination between computer controlled artificial joints and human joints in walking. This project will develop novel technology to address this “human-prosthesis symbiotic locomotion problem” which is to say that a human and a wearable robotic lower limb prosthesis interact as collaborating agents and function in unison to achieve a common locomotion performance goal. This research will improve amputees’ walking performance and efficiency while enhancing their sense of prosthesis embodiment and device acceptance, thus dramatically improving their quality of life. This is the first study of human-prosthesis coordination and co-adaptation in locomotion, and will contributes important knowledge in the fields of physical human-robot interaction and human-centered computing. Additional broad impacts will derive from integration of the research into interdisciplinary training of undergraduate and graduate students, in areas including machine learning, human motor control and learning, and rehabilitation engineering. The objective of this project is to create and evaluate a Human-Centered Computing and Control (HC3) framework for improved human-prosthesis walking performance (in terms of symmetrical gait, postural stability, and reduced sense of effort) and prosthesis embodiment, the research to include three thrusts: (1) an innovative formulation of a coordinated human-robot control problem and a learning-based design solution to achieve human-prosthesis symbiotic locomotion, a problem that poses great and new challenges to classical control and existing multiagent reinforcement learning methods; (2) a novel augmented biofeedback interface to enable active human locomotion adaptation to coordinate with the robotic prosthesis; and (3) a new measure of embodiment of intelligent lower limb prostheses in walking based on the concept of biological motion. Furthermore, the HC3 framework will be evaluated on transfemoral amputees walking with a robotic knee prosthesis.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.