ABSTRACT
The standard problem of reduced-order filtering has been extensively reported for the case where some of the observations are assumed to be noise free [1-3]. In this chapter, unlike the previous studies, we consider the case where none of noise-free observation is assumed. This chapter deals with the design problem of a reduced-order filter which estimates a specific linear function of the state for the linear stochastic system [4-6]. It can be shown that this problem can be reformulated as a standard Kalman filtering problem for the reduced-order system obtained through an appropriate system transformation. Thus a method for designing a reduced-order filter offers the possibility of significant reduction in computational requirement and less complexity in physical implementa tion. However, the price to pay for these benefits is some loss of performance compared with the full-order Kalman filter. Finally the relation to the full-order Kalman filter is also carefully discussed in this chapter.
