Closed Loop Identification of the Physical Parameters of an Industrial Robot

Abstract

This paper discusses experimental identication of a flexible robot under strong feedback. Black-box models are compared with physically parameterized models. The parameters in the physical models and the blackbox models are identified from the input-output data using prediction error methods. We also investigate the importance of noise models when performing identication under strong feedback. The robot that we use is a commercial ABB robot, IRB 1400.