Nonlinear Controller Design via Inferenced-Augmentation of Equivalent Linearized System

Abstract

Output compensation of nonlinear system response is studied in this work with an inference-based design method used to control the nonlinear dynamics in the presence of non-linear effects. This design which is driven by practical considerations is informed by the comparisons made between nonlinear models and their linearized derivatives. While a detailed mathematical route has not been followed here, the results seem to show the workability of the proposed method and at its core, the design is driven by its intuitiveness. The proposed technique utilizes as primary parameter for comparison, the steady state error term which is derived from the behavior of the linear dynamics. Possible application areas for this design is in reduced energy control of nonlinear systems. The method was tested in simulation on a generic nonlinear system and a cart-driven inverted pendulum benchmark system.