Abstract
This paper discusses the problem of estimating state variables and nonstationary parameters for adaptive control system in the presence of random disturbance and mesurement noise.In this paper, a parameter is assumed to be “Frequency Band Limited Process”, so that the parameter equation can be described approximately with m prior parameters and white gaussian noise by introducing the sampling theory. This approximate parameter equation is applied to the nonlinear filtering theory for estimating a parameter from indirect data with the presence of noise. The number of m is changed properly according to the nonstationary variation of the parameter.
Experimental computer simulation of 1st order system confirms the practicability of the method.
