High-performance intelligent biorobotics and assistive locomotion

点击翻译按钮获取中文摘要

The development of intelligent machines that can mimic biological locomotion or enhance human mobility, or biorobotics in general, is undergoing a rapid growth as the underlying mechatronic technologies are evolving. However, these machines are far less optimal than biological systems in terms of their stability, robustness, intelligence, and energy efficiency. This research, inspired by biological systems, aims at optimizing the performance of legged robots through optimal motion planning, intelligent learning controllers and exploiting passive dynamic motions of mechanical systems. Bipedal locomotion, due to its significance to biomedical and human kinetic sciences, has been selected as the main focus. Energy-efficient and stable gaits will be obtained from bipedal walking, through off-line an off-line multi-criteria optimization process or through modulating simple dynamics. A new learning control strategy will be devised that will be applicable to biped locomotion on uneven terrains and postural stability on moving grounds such as ship boards. This research also aims at using the passive dynamic principles to develop new classes of efficient bipeds and wearable devices with intelligent controllers that can be used as walking aids. A single-degree-of-freedom compliant assistive device for knee joint will be realized to investigate intelligent learning controllers and improvements in energetics. The EMG (Electromyogram) signals and user interaction forces are used in a feedback controller to determine the required torques to augment the knee power in a safe fashion for different users. Collaborations with Ottawa Hospital's Rehabilitation Centre will ensure the applicability of the results which can improve the quality of life of many Canadians with walking difficulties.