Adaptive control using radial basis function networks

This paper proposes the use of a radial basis function network (RBFN) for the online adaptive identification of inverse plant dynamics. The RBFN is incorporated into a model reference adaptive controller which tracks an idealised linear model of the plant under stable optimal control. Since online training is maintained during plant operation the effects of environmental changes, disturbances, and parametric uncertainty, are accommodated by the inverse model and system performance is maintained. This leads to a high degree of fault tolerance within a given class of nonlinear plants which is demonstrated by the performance of a helicopter model under fault conditions.