As an example of human-machine interaction, research on musical instrument ensembles between humans and machines has been conducted. In this study, we propose a robot system that can play an acoustic guitar. There have been several previous studies of guitar robots, but they have been limited by (1) the heat generation problem of the string pressing mechanism, (2) the difficulty of independently pressing the strings, and (3) the difficulty of controlling the strings. (2) The strings are independently plucked, and the stroking technique is not the same as that used by humans. (3) It is not practical because of these problems. In this study, we solved these problems and developed a guitar robot system that can play in concert with humans. In an evaluation experiment using subjects who could distinguish between the performances of humans and robots, the correct answer rate was about 70%, and we confirmed that even humans could not distinguish between the two performances. The effectiveness of the proposed system was demonstrated.

Although many factories adopt machining centers for fully-automated processing, humans still need arrange works to be processed on pallets with appropriate attachments, in which humans' operations have large effects for efficient plans. Such problems can be regarded as variants of job-shop scheduling problems. It is not easy to obtain efficient schedules for such problems since we need consider combination of attachments, human's operation time, and so on. In this study, we propose a method to obtain schedules for machining centers using real-time model checking. We also do experiments in an actual factory to show the usefulness of our method.