Nonlinear, adaptive model-following control using numerical inverse of a fuzzy model

A novel, adaptive, nonlinear model-following control using an on-line, numerical inversion of a fuzzy model is proposed. We consider a nonlinear plant (slave system) with unknown parameters and implicit, nonlinear dependance of control and state variables. The plant is supposed to follow a nonlinear master system. The appropriate function of control and state of the plant is represented by a fuzzy model, not necessarily being a one-to-one map. The model is numerically inverted on-line to get the desired control value. Consequent functions of the model are changed by robust adaptive laws, to cope with any modelling inaccuracies, while the other adaptive parameters are changed to follow the unknown parameters of the plant. This approach achieves a short execution time and can be applied in standard DSP-based controllers.