DocumentCode
2608016
Title
Restricted structure adaptive control including series 2 DOF tracking
Author
Grimble, M.J.
Author_Institution
Ind. Control Centre, Strathclyde Univ., Glasgow, UK
fYear
2000
fDate
2000
Firstpage
420
Lastpage
425
Abstract
A new approach to adaptive control is presented based on a multiple model control calculation, where the system model is identified using extended least squares. The control law calculation is for a restricted structure controller, such as a simple PID form. The restricted structure algorithm enables a simplified cost-function to be optimized numerically. This approach can be used for systems described by multiple models, where the restricted structure controller is optimised across a set of model/cost-function pairs. If these models are obtained by a plant identification step an adaptive algorithm results. The main advantage of the approach is the cautious nature of the solution. The new update of the controller involves averaging the cost across old and new models and this provides cautious robust adaptive control action. The solution also provides a degree of robustness that can be predefined, assuming the possible plant models are all within a given set
Keywords
adaptive control; identification; least squares approximations; robust control; tracking; PID; adaptive algorithm; extended least squares; multiple model control; plant identification; restricted structure adaptive control; restricted structure algorithm; restricted structure controller; robust adaptive control; series 2 DOF tracking; simplified cost-function; system model; Adaptive algorithm; Adaptive control; Buildings; Costs; Electronic mail; Industrial control; Least squares methods; Optimal control; System identification; Three-term control;
fLanguage
English
Publisher
ieee
Conference_Titel
Adaptive Systems for Signal Processing, Communications, and Control Symposium 2000. AS-SPCC. The IEEE 2000
Conference_Location
Lake Louise, Alta.
Print_ISBN
0-7803-5800-7
Type
conf
DOI
10.1109/ASSPCC.2000.882511
Filename
882511
Link To Document