Investigation of Impact Hammering Task of Musical Saw With a Humanoid Robot for Industrial Automation

This study aimed to automate the hammering action and dual arm coordination of a humanoid robot using percussion instruments such as musical saw. In this report, we examined the hammering motion of the robot by using its elastic stiffness, the vibration reduction method of the hitting tool caused by the movement of the robot arm, and dual arm cooperative motion based on hammering sound feedback. The experimental results show that the humanoid robot can play the musical saw without making an S-shaped curve, as performed by humans. By considering these nonlinearities through the mathematical expression of the relationship among frequency, deflection, length of beam, and fulcrum deviation, an arbitrary scale can be generated in the saw. Based on the hammering task and acoustic recognition system, we developed an impact sound feedback system. It was possible to converge to the target frequency of the musical saw through the hammering sound feedback operation.