Feedforward learning control of nonlinear plant: Introduction of offset terms to multi inverse models

This paper proposes a design method for feedforward (FF) learning control. For an unknown nonlinear plant which is free of zero dynamics, we construct a FF controller consisting of a bank of approximation models that estimate unknown parameters. A conventional scheme used linear approximation however the accuracy of response shaping was deteriorated because every operating point is not an equilibrium point. In view of this, we propose to introduce offset terms in the approximation around operating points, thereby improving accuracy of response shaping. Numerical simulation is carried out to verify the effectiveness of the proposed scheme over existing one.