Position and elasticity control for biomimetic robot finger

In this paper, a robot finger with a small mass and elasticity has been proposed to develop a human-friendly robot that can coexist in the same environment as the human area. The tendon-driven mechanism has an advantage of being able to control the fingertip elasticity by adjusting the tensions of each tendon. However, the tendon-driven mechanism can not achieve a precise position control of the fingertip. This is because the posture of finger is stable at the equilibrium point of each tendon tension that mainly depends on the friction force. In addition to this, the relation between the elasticity and the tendon tensions is non-linear one. Therefore, we use a fuzzy controller to control the fingertip position and the elasticity using fuzzy map that is tuned by, a local learning algorithm. The obtained experimental results have been analyzed and discussed front a viewpoint of further improvement.