Mamdani-type fuzzy controllers are universal fuzzy controllers (Q5951779)

The authors address a problem about universal fuzzy controllers along with a proof that Mamdani-type fuzzy controllers are universal fuzzy controllers. The system under control is described by a state-space model of the form \(dx/dt=f(x(t),u(t))\) with a general form of the state feedback law \(u(t)=g(x(t))\), ``\(f\)'' and ``\(g\)'' being functions with first-order continuous derivatives with respect to the corresponding variables, \(x(t)\in \mathbb{R}^n\), \(u(t)\in \mathbb{R}^p\). It is also assumed that an equilibrium \(x_0=0\) exists and \(f(0,0) =0\). The Mamdani-type fuzzy controller is composed of control rules of the form ``if \(x_1\) is \(A_{i1}\) and \(x_2\) is \(A_{i2}\) and\dots and \(x_n\) is \(A_{in}\) then \(u(t)=g_i\)'' (\(i\)-th control rule with \(A_{i1}\), \(A_{i2}\), etc. being fuzzy sets). Along with the property of universal approximation, it is demonstrated how to design this controller for a given reference fuzzy model.