PARAMETER ESTIMATION FOR DUAL-RATE SYSTEMS WITH FINITE MEASUREMENT DATA 1

This paper on parameter estimation is motivated by practical consideration that the output sampling rate is often limited and that the data length is finite. In par- ticular, for dual-rate systems in which the output sampling period is an integer multiple of the input updating period, we obtain frequency-domain models, study the properties of the least squares type algorithms in detail in the stochastic framework, and derive conver- gence rates and upper bounds of parameter estimation errors (PEE) for the least-squares (LS) algorithm, instrumental variable least squares (IVLS) algorithm, and forgetting-factor least squares (FFLS) algorithm, using directly the finite dual-rate input-output data. The analysis indicates that the mean square PEE upper bounds of LS and IVLS algorithms are proportional to 1=k and converge to zero as the data length k increases, and the PEE upper bound of the FFLS algorithm approaches a finite constant. Finally, we illustrate and verify the theoretical findings with example systems, including an experimental water-level