A Distributed Nonlinear Observer Design Method for Output Estimation in Nonlinear Systems
Javad Mohammadpour, Ali Hooshmand, Heidar Malki, Karolos M. Grigoriadis, Robert S. Provence · 2011
In the present paper, we develop a method to design a nonlinear observer using the solution to a linear matrix inequality (LMI) optimization problem. The proposed observer design method is then employed to the problem of estimating the position of a rover on the lunar surface. The nonlinear observer design method is then extended to the distributed case using the dynamic consensus filter approach proposed by Spanos et al. 1 to facilitate the real-time implementation of the filtering algorithm for the multi-rover case. Since the proposed observer design method does not make any assumption on the stochastic or deterministic nature of the noise, it is particularly suitable to provide improved estimation results compared to the Kalman filter in the presence of such noises. The measurement data from sensors with various measurement noises provided by NASA are used to validate the proposed centralized and distributed estimation methods.