Proposal of Method “Motion Space” to Express Movement of Robot
Kentarou Kurashige, Naoki Kitayama, Masafumi Kiyohashi · Journal of Advanced Computational Intelligence and Intelligent Informatics · 2012
In recent years, the use of robots has been spreading to various fields. Further, requirements for the use of robots are increasing. A method is therefore necessary for allowing persons who are not experts in using robots to actually operate a robot. We intend to develop a method for robot operation whereby a user need not have technical knowledge. In this paper, we focus on methods whereby a user of a robot assigns movement to the robot and the robot reproduces movement. One of the most widely used techniques that is used nowadays involves teaching playback. Teaching playback is a method in which a teacher moves a robot using controllers and lets the robot record movement and then play it back. Robots functioning via teaching playback cannot however adapt to a changing environment. The environment in which human beings live generally change. Teaching playback is therefore not usable in variable environments. Methods for generating movement robustly in environments have been studied. Designing the movement of a robot by using these methods cannot be done, however without understanding complicated formulas. Only movement designers having technical knowledge can use these techniques. We propose new knowledge of movement to solve the problems present in these methods. Knowledge of movement is information involving the generation of movement in a robot. In conventional methods, knowledge ofmovement was a complicated formula. By using our method, a robot incorporates knowledge based on information obtained by moving a robot just like in the teaching playback method. We expect that by using our methods a user can move a robot in the desired manner.