Toward further innovation in human-machine systems

Abstract

A human-machine system (HMS) is an intelligent system in which a human operator cooperates and/or shares the functions of the system with a partially- or fully-automated machine through an operational interface with the goal of completing a target task. Examples of an HMS include an automobile, an airplane, and a wearable power-assist robot. The most important point related to the development of such an operational HMS is how the machine could move and cooperate with a human operator while executing a target task without adversely affecting not only the performance of the task but also the human operator's comfort. Despite the fact that the hardware systems of recent machines are well-developed, there remain many issues affecting the realization of an ideal HMS that could fully harmonize its behavior with a human because of the difficulty in treating uncertain human properties and individuals. Therefore, recent studies on HMS have addressed many related aspects, including control theory, robotics, biomechanics, human factors/ergonomics, and neuroscience as the need arises. Recent trends in the field of HMS research include a human-centric methodology and a shared control methodology, for which better computational models for task-related human motor function and/or perceptual functions are formulated based on experimental data and embodied into the control systems of automated machines. This talk provides an outline of an HMS, describes some state-of-the-art research, and invites open questions with the aim of identifying what we should/could do to achieve further innovation in the HMS field.