Comparing robot controllers through system identification

In the mobile robotics field, it is very common to find different control programs designed to achieve a particular robot task. Although there are many ways to evaluate these controllers qualitatively, there is a lack of formal methodology to compare them from a mathematical point of view. In this paper we present a novel approach to compare robot control codes quantitatively based on system identification: Initially the transparent mathematical models of the controllers are obtained using the NARMAX system identification process. Then we use these models to analyse the general characteristics of the cotrollers from a mathematical point of view. In this way, we are able to compare different control programs objectively based on quantitative measures. We demonstrate our approach by comparing two different robot control programs, which were designed to drive the robot through door-like openings.