The radial basis function neural network (RBFNN) can be employed as an approximator to compensate the system uncertainties after effective learning.