Human adaptive assist planning without operator awareness

A purpose of our research is to realize teleoperation assistance to improve the performance with operator´s initiative. One of advantages of teleoperation system is utilized human abilities which are global recognition, planning, prediction and so on. However, the operator is required to get skills for suitable operation, and assistance from the system is important. Although effects of the operation assistance are generally contemplated by system designers, it is not always available because of depending on variable operator´s characteristics and environment condition. Solving this problem, adaptive assistance is required, however, autonomous robot motion may disturb to give an initiative to human operator, that is, human abilities cannot be brought out. Therefore, assistance by modification of robot dynamics parameters has been proposed. Because human operators tries to adapt and learn the robot dynamics instinctively, such assistance often brings about discomfort and less maneuverability. Consequently, human awareness characteristics are considered to improve this problem an assistance technique without human perception using a cognitive science approach is addressed. In this paper, new assist technique without discomfort feelings of operators is proposed.