Design for Accuracy and Repeatability for Robots Using Taguchi Methods

In this paper, we apply Taguchi Methods in the framework of concurrent engineering design for robot-based manufacturing sys tems to obtain higher accuracy and repeatability for robot end-effectors. It is well known that the configurations of robot end-effectors in workspace will affect the accuracy and repeatability Taguchi Methods are employed to find the configuration and location within which the accuracy and repeatability are the best. The optimal position with higher accuracy and repeatability is determined by using the signal-to- noise ratios (S/N ratios), which are defined according to the "smaller-the-better" characteristics of Taguchi Methods to measure the accu racy and repeatability The results match well with the theoretical analysis of the repeatability and accuracy of the end-effector based on the kinematic relationship. In addition, we also study the random and systematic errors due to joint uncertainties and examine their effects on the accuracy and repeatability. The analysis using S/N ratios yields results that are consistent with expectation. The results demonstrate that Taguchi Methods and S/N ratios are very useful tools in design for accuracy and repeatability of robotic end-effectors. The sensitivity and consistency of accuracy with respect to the tolerance of joint angles are represented by 3D surfaces. The tolerance design is per formed, using the sensitivity analysis, to satisfy requirements stipulated by design criteria One of the significant results of consistency analysis is the "threshold" value of tolerance for concurrent engineering design considerations. Simulation results show that the consistency of accuracy and repeatability will be increased dramatically once the tolerance is below such a threshold Both sensitivity and consistency analysis are shown to be important in tolerance design for achieving the desired accuracy and repeatability and for enhancing quality.