Model identification of a 3 finger adaptive robot gripper by using MATLAB SIT

Abstract

This paper presents the method of finding the estimated plant transfer function of the 3 Finger Adaptive Robot Gripper by using the MATLAB System Identification Toolbox (SIT). The robot gripper consists of 3 under actuated fingers, where the active joint is driven by a DC motor and the passive joint is driven by the underactuated mechanism (elastic tendons). To simplify the study, the model identification only considers the single angular joint of each finger. The approach of fast variable step input (i.e. stairs input) and slow variable step input (i.e. slope input) was introduced while the output of the robot gripper is referring to the motor encoder position of each finger. The best estimated modelling of the gripper is obtained by selecting the transfer function that has the most similar performance (in term of position control) compared to the actual system. Moreover, the result shows that the transfer function obtained by using fast variable step input is sufficient to represent the 3 Finger Adaptive Robot Gripper. Additionally, the PID position control was employed and the result shows that the gripping performance is satisfactorily achieved in simulation and experiment.