This paper describes a computer simulation analysis of modular networks and fuzzy-logic controllers for the motion control of 2-D nonlinear flexible structures. The Stochastic-Gradient Learning Algorithm using modular networks is presented and applied as two inputs one output controller. The self learning process of the Modular Networks consists of a set of LQR. Each one of the LQR was optimized to control a different zone, a state space. The trained controller is used to actually control the system. The two methods are compared quantitatively though the work done, performance index (settling time, energy consumption and over shoot). The comparison shows the advantage of the MN control method having better performance index results and requires less effort in the tuning stage.
|Number of pages||5|
|State||Published - 1 Dec 1997|
|Event||Proceedings of the 1997 Annual Meeting of the North American Fuzzy Information Processing Society, NAFIPS'97 - Syracuse, NY, USA|
Duration: 21 Sep 1997 → 24 Sep 1997
|Conference||Proceedings of the 1997 Annual Meeting of the North American Fuzzy Information Processing Society, NAFIPS'97|
|City||Syracuse, NY, USA|
|Period||21/09/97 → 24/09/97|