Closed-loop model set validation under a stochastic framework

Authors:
Tong Zhou;Ling Wang;Zhengshun Sun
Affiliations:
Department of Automation, Tsinghua University, Beijing 100084, China;Department of Automation, Tsinghua University, Beijing 100084, China;Department of Automation, Tsinghua University, Beijing 100084, China
Venue:
Automatica (Journal of IFAC)
Year:
2002

Citing 5
Cited 1

Quantification of uncertainty in transfer function estimation: a mixed probabilistic—worst-case approach

Automatica (Journal of IFAC)
On some key issues in the Windsurfer approach to adaptive robust control

Automatica (Journal of IFAC)
Identification and control—closed-loop issues

Automatica (Journal of IFAC) - Special issue on trends in system identification
Estimation of model quality

Automatica (Journal of IFAC) - Special issue on trends in system identification
Closed-loop identification revisited

Automatica (Journal of IFAC)

H∞ set membership identification: A survey

Automatica (Journal of IFAC)

Quantified Score

Hi-index	22.15

Visualization

Abstract

This paper deals with probabilistic model set validation. It is assumed that the dynamics of a multi-input multi-output (MIMO) plant is described by a model set with unstructured uncertainties, and identification experiments are performed in closed loop. A necessary and sufficient condition has been derived for the consistency of the model set with both the stabilizing controller and closed-loop frequency domain experimental data (FDED). In this condition, only the Euclidean norm of a complex vector is involved, and this complex vector depends linearly on both the disturbances and the measurement errors. Based on this condition, an analytic formula has been derived for the sample unfalsified probability (SUP) of the model set. Some of the asymptotic statistical properties of the SUP have also been briefly discussed. A numerical example is included to illustrate the efficiency of the suggested method in model set quality evaluation.