Research on fuzzy modeling and predictable control in electro-hydraulic servo systems

In the field of control systems, constructing an accurate mathematical model aimed at drive element is an important task as it is closely related to control quality. Since hydraulic elements present uncertain factors, such as hysteresis, friction, dead-zone, thermal behaviors and viscosity, magnetic saturation, it is difficult for these systems to produce an accurate mathematical model to meet the desired performance. In this contribution, we propose a new control method by combining fuzzy modeling and inverse predictable compensation with a conventional controller to solve this problem. Firstly, we obtain a fuzzy model with fast fuzzy clustering, taking into account the information provided by input-output samples. Next, a new control mechanism is developed, which can incorporate conventional controller with inverse predictable compensation. In order to improve the control effect, the learning method makes use of current data to adjust the online of the fuzzy model, which leads to reveal state transition of systems in real time. The related performance of the control mechanism in terms of feasibility and effectiveness is also discussed based on the results of simulation and experiments for the concerned system.