Intelligent control using neural networks

Abstract
Intelligent control, using neural networks, of complex dynamical systems in the presence of structural failures is addressed. After a failure occurs, the dynamical system is assumed to be in one of a finite number of configurations, with a stabilizing controller corresponding to each. A two-level controller structure is used, in which the higher level detects a failure using pattern recognition techniques and activates a stabilizing controller. At the second level, an adaptive controller is used to optimize the stabilized system. Multilayer neural networks are used at the lower level for the identification and control of the plant and as a pattern recognizer in the higher level for the detection of a failure in the system. Simulation results are presented, showing that efficient multilevel controllers using neural networks can be designed for nonlinear dynamical systems.

This publication has 4 references indexed in Scilit: