Fuzzy predictive controller for chaotic flows based on continuous signals

In this paper, we propose a fuzzy model predictive control method, which can be used in the control of highly nonlinear and complex systems, like chaotic ones. This method only uses the obtained time series of the system and does not require any prior knowledge about the system's equations. In our proposed method, a fuzzy model is created using a combination of Gaussian basis functions. The model is developed using initial part of the time series, sampled from an observed signal from the nonlinear chaotic system (learning phase). Then, the developed fuzzy model is used to modify the controller. The controller, which is tuned in each sample of the time series, is subsequently applied to an interval of the continuous signal and holds the system in the desired state. We investigate the efficiency of this new control method using a chaotic system with no equilibrium point, which belongs to category of chaotic systems with hidden attractor.