System Identification And Simulation Of Bicycle Robot

As the basis of designing kinds of control system, the dynamic model of the controlled object is meaningful for its performance. However, when set up these dynamic model with those traditional methods, there are always such problems as low efficiency or even impossible to set up useful model since time-delay or non-inertial of these object, which constrict the application of kinds of controlling theory and improvement of controlling performance. In this paper, we introduce the theory of system identification to get a more exact system dynamic than traditional methods. Research result shows that this theory is feasible to our bicycle robot controlling system.This paper focus on the identification of bicycle robot system based on ARX model and ANFIS model using MATLAB. Firstly, basic theory and common methods are introduced, then the system identification based on neutral network and fuzzy system; Then the basic concepts and algorithm of fuzzy theory and neutral network theory.Based on which, we also introduce the detailed identification methods of fuzzy model system; at last, identification the bicycle robot system with ARX and ANFIS model, using MATLAB identification box tools. We also compare the result of these two identification methods. At last, the summary of this paper and future picture of our research work is outlined.