This paper proposes an adaptive and learning control scheme for unknown nonlinear systems. Plant dynamics are approximated by finite power series models around several operating points. The controller performs estimaion of parameters, prediction of parameter changes and storage of parameters.
