Premium
Controller falsification based on multiple models
Author(s) -
Agnoloni Tommaso,
Mosca Edoardo
Publication year - 2003
Publication title -
international journal of adaptive control and signal processing
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.73
H-Index - 66
eISSN - 1099-1115
pISSN - 0890-6327
DOI - 10.1002/acs.745
Subject(s) - statistic , lemma (botany) , divergence (linguistics) , control theory (sociology) , controller (irrigation) , test statistic , stability (learning theory) , residual , computer science , mathematics , mathematical optimization , control (management) , statistical hypothesis testing , statistics , algorithm , artificial intelligence , ecology , linguistics , philosophy , poaceae , machine learning , agronomy , biology
The paper addresses the controller falsification problem in feedback‐control systems where the plant dynamics are uncertain and possibly time‐varying. The adopted approach considers a statistic (or residual) in the form of a ratio of closed‐loop variables. The test hinges upon a small‐gain stability lemma which ensures that, under divergence, the statistic certainly exceeds a precomputable bound. Use of multiple models appropriately distributed over the uncertainty region of the plant dynamics allows one to generate a multi‐statistic such that the denser the model distribution the larger the value taken on by the mentioned bounds. This allows one to arbitrarily widen the gap between the level that the statistic is guaranteed to stay below at stability and the one that will exceed at divergence, thus removing limitations on the applicability of the test based on a single model. Copyright © 2003 John Wiley & Sons, Ltd.