Study on Data Processing of a Teaching Robot with Redundant Degrees-of-Freedom

Abstract

Shape data input of an industrial art object can be easily performed by using a teaching robot. The objective of this research is to reduce the random errors in the teaching data and to obtain smooth curved lines and surfaces. By this, 3D graphic expression and machining of an industrial art object can be made easy. For characteristic lines, the least square method is used to fit a polynomial expression. For curved surfaces, two parameters (u_i, v_i) for expressing surface are used, and the teaching data are sorted with respect to these parameters. The least square method is then used to fit polynomial expressions on the curved surfaces. Finally, the fitted polynomial expressions are used to calculate corrected value of teaching data for generating smooth curved lines and surfaces.