Experimental evaluation of parameter identification schemes on a direct-drive robot

The dynamic and friction parameters of a robot are used in advanced control schemes, and their accuracy significantly affects their performance. These parameters can also be used for a realistic simulation. In principle, the numerical value of the parameters could be obtained via computer-aided design analysis but inevitable assembly and manufacturing errors exist. Direct measurement is not a realistic option because the complex nature of the system would involve an intense time-consuming effort. Alternatively, we can deduce the values of the parameters by observing the natural response of the system under appropriate experimental conditions, that is, by using identification schemes. This article presents the experimental evaluation of five identification schemes used to obtain the dynamic and friction parameters of a two-degree-of-freedom, direct-drive robot. We assume that the dynamic and friction parameters are totally unknown but, by design, the dynamic model is fully known. We consider the schemes based on the dynamic regression model, filtered-dynamic regression model, supplied-energy regression model, power regression model, and filtered-power regression model. The article presents a comparison between experimental and simulated robot responses, which enable us to verify the accuracy of each regression model.