Reduced nonlinear unknown inputs observer using mean value theorem and patternsearch algorithm

This paper is devoted to reduced order of unknown inputs observer design. A new structure is proposed, and it includes the dynamics of the estimation error and the uncertain parameters beta integrated into the mean value theorem. The pattern search algorithm will be used for the determination of beta parameters and the observer design. The stability study relies on the use of the quadratic Lyapunov functions. The observer gain is systematically determined by using an analytical method. A second method will be proposed for the gain determination, and it is based on the LMI technical and the polytopic transformation. The case of outputs affected by the disturbances is also studied. Finally, a numerical example is provided to illustrate the performance of the proposed approach. (C) 2019 Elsevier Ltd. All rights reserved.