Adaptive inverse control using support vector regression

This paper explores application of support vector regression to adaptive inverse control problems. Support vector regression (SVR) has been proven to generate global solutions contrary to neural networks, because SVR basically solves quadratic programming (QP) problems. With this advantage, a plant model is identified and its inverse model is learned. In addition, adaptive algorithms for compensating the errors between the actual model and identified model are proposed and their convergence property is discussed. Finally, numerical simulation is performed for the validation of the proposed approach using the longitudinal dynamics of unmanned aerial vehicle (UAV).