Human-Robot Analogy − How Physiology Shapes Human and Robot Motion

摘要

The fast grow of expectations from robots and the technical obstacles the robot developers face when trying to meet the requests, force an orientation towards designing and controlling robots by following biological paragons. This tendency increases interest in human-robot analogy, and the present work is a part of this stream. The paper questions the voluntariness of human motion by relating it to physiological processes. We concentrate on the important question of redundancy. One notes that humans do not resolve the redundancy on the level of consciousness, except in some specific examples (like obstacle avoidance), but rather on a lower level of decision-making – the human reaction to the faced problem is somehow "automatic". We suggest that this automatism is closely related to physiological processes, particularly to the progress of fatigue. Then we try to mathematically model these processes and their influence to human motion. The mathematical model of fatigue progress is derived as well as an algorithm for human-like redundancy resolution. We finally consider the implications of the obtained results to anthropomimetic robotics. The concept is verified by simulating the system behavior and comparing it qualitatively with the behavior of a human control group.