Title :
Parameter learning for performance adaptation
Author :
Peek, Mark D. ; Antsaklis, Panos J.
Author_Institution :
Tellabs Res. Center, Mishawaka, IN, USA
Abstract :
A parameter learning method is introduced and used to broaden the region of operability of the adaptive control system of a flexible space antenna. The learning system guides the selection of control parameters in a process leading to optimal system performance. A grid search procedure is used to estimate an initial set of parameter values. The optimization search procedure uses a variation of the Hooke and Jeeves multidimensional search algorithm. The method is applicable to any system where performance depends on a number of adjustable parameters. A mathematical model is not necessary, as the learning system can be used whenever the performance can be measured via simulation or experiment. The results of two experiments, the transient regulation and the command following experiment, are presented.<>
Keywords :
adaptive control; aerospace control; distributed parameter systems; large-scale systems; learning systems; optimisation; satellite antennas; search problems; Hooke and Jeeves multidimensional search algorithm; adaptive control; command following; flexible space antenna; grid search procedure; optimization search; parameter learning; performance adaptation; transient regulation; Adaptive arrays; Adaptive control; Artificial intelligence; Control systems; Learning systems; Machine learning; Optimal control; Programmable control; Robust control; System performance;
Journal_Title :
Control Systems Magazine, IEEE
