METHOD OF SYNTHESIS OF A SLIDING FUZZY CONTROLLER IN THE OBSERVATION MODE

Abstract

The article considers the method of synthesis of a sliding fuzzy controller in the observation mode. This method is most widely used when the full state vector of the system is not available for measurement, but it is necessary to estimate it for the implementation of full state feedback or for diagnostic purposes. The most common and theoretically justified approach to solving this problem is the synthesis of a parallel sliding fuzzy controller. The use of a controller of this type is one of the most common means for controlling robots, UAVs, electric drives and all tasks where high accuracy and robustness to significant uncertainties and disturbances are required. That is, the sliding fuzzy controller is used for objects of industrial automated process control systems (ACS TP), which are characterized by significant nonlinearity, parametric changes and external disturbances. When synthesizing the controller, hybrid schemes based on Mamdani fuzzy logic are proposed, which eliminate the main drawback of the classical sliding mode – by replacing the relay component with a smooth fuzzy approximation. Sufficient conditions for stability and achievement of the sliding mode are given in the form of linear matrix-fuzzy constraints, which are adapted for typical models of ACS ТР (aperiodic links of higher orders, objects with distributed parameters). An adaptive sliding fuzzy controller with automatic adjustment of scaling factors λ is considered, which provides continuous control and improvement of the qualitative characteristics of the transient process. According to the results of theoretical studies in mathematical modeling, it was established that when the scaling factor λ increases to a value of 0.5, the transient process for the controlled parameter acquires an exponential form, which allows reducing the dynamic error.