Distributed UIO for Multiple Sensor Systems With Unknown Inputs and Measurement Noise

Abstract

In the article, a distributed unknown input observer (DUIO) constructing technique for a system with unknown input (UI) and measurement noise (MN) is come up with. Specially, multiple sensor groups are equipped to measure the output of the target system, and the output information measured by each sensor group is not enough for estimating the system states individually. It is assumed that each sensor group suffers from MN. To begin with, an output transformation and a state shift transformation are made such that the transformed subsystem contains neither UI nor MN. Next, for each transformed subsystem, a distributed state observer is designed to obtain the asymptotic convergent state estimation, and also the MN reconstruction (MNR). Furthermore, an output interval estimation method is proposed, and a correlation of the UI and the state is come up with algebraically. Then, by using the relationship, an asymptotic convergent UI reconstruction (UIR) method is developed. Finally, a simulation example is given to verify the effectiveness of the proposed method.