Intelligent control of flexible joint based on cooperative learning theory

Abstract It is complex and time-consuming for obtaining accurate dynamics of industrial robots. The flexibility the joint not only increases adverse effects nonlinear factors, but also makes controller design robot significantly difficult. In order to improve performance robots containing flexibility, cooperative learning (COL) theory proposed in this paper. Based on theory, a model-free intelligent trained successfully applied flexible joints. Compared with conventional PID RBF controller, cumulative tracking error produced by reduced 38.25% 31.08%, respectively, robustness improved trajectory experiment experiment. experimental results demonstrate feasibility effectiveness theory.