Kolodyazhniy V. Control of stochastic systems under uncertainty using fuzzy neural models

The problem of control of various dynamic stochastic plants under the conditions of structural and parametric uncertainty is considered. The control laws based on tuned models are improved so that the constraints on the control error and the actuating variable dynamics could be accounted for. An adaptive high-performance identification procedure is proposed. New fast adaptive learning and self-organization procedures for fuzzy neural models are derived. The control method based on inverse fuzzy model is modified. New control methods based on locally linearized fuzzy models with the use of online learning and self-organization procedures are proposed. Computer simulation of the developed control methods and learning procedures is carried out. Real-world problems of controlling complex nonlinear dynamic plants with the application of the developed models and control methods are solved.