Integration of CMAC-GBF and Support Vector Regression Techniques

In this paper, we integrate the techniques of cerebellar model articulation controller with general basis function (CMAC-GBF) and support vector regression (SVR) to develop a more efficient scheme. The advantages of CMAC-GBF include: fast learning speed, guarantee learning convergence, capability of derivative, etc. On the other hand, a SVR is a novel method for tackling the problems of function approximation and regression estimation based on the statistical learning theory and has robust properties that against noise. In this paper, we propose the SVR-based CMAC-GBF systems that combined SVR with CMAC-GBF systems. From the results of simulation, the proposed structure has high accuracy and noise against. Besides, the experimental testing results demonstrate that the SVR-based CMAC-GBF systems outperform the original CMAC-GBF systems.