Adaptive inverse control based on online SVR

At present, the control of a dynamic system is generally done by means of feedback. This paper proposes a new method that use online SVR to achieve feedforward control for both linear and nonlinear plants. Online SVR is a new method when a new sample is added to (or removed from) the training set, it didn't need retraining from scratch for each new data point. Therefore, it is an efficient algorithm. It has some advantages such as low computation, good approximation properties and so on. The Simulation results given in this paper shows that the algorithm has good control performance.