Fuzzy Neural Network In The Mine-clearing Plow System

Electro-hydraulic servo system has been used in mine sweeping plough system due to its advantages such as high precision, fast response, big output power. Traditional control methods such as feedback control, PID control can't obtain satisfactory control effect because electro-hydraulic servo system is a strong nonlinear and time-varied system. In recent years, Fuzzy Neural Network (FNN) attracts more and more scholars'attention. It uses the learning ability of neural network to achieve the extraction of fuzzy rules. Fuzzy neural network can be employed in control of nonlinear systems owing to its ability to express nonlinear mapping. This paper uses FNN for the modeling and control of electro-hydraulic servo system of mine sweeping plough. The following sections will introduce the main work finished:First, electro-hydraulic servo system of mine sweeping plough had been built in AMESim software by working principle and parameters of main components of system. The simulation results were analyzed in detail. Then, hydraulic circuit which controls movement of plough was studied. Input and output data within the scope of system were collected by a Simulink model when co-simulation was done.Adaptive Neural-Fuzzy Inference System (ANFIS), Fuzzy Neural Network based on Subtractive clustering (Sub-FNN) and Dynamical Fuzzy Neural Network (D-FNN) were used to settle the modeling problem of the system. The ability of generalization was discussed in detail. The results showed that the error of former two models associate with the input signal, while D-FNN model has little error, strong anti-interference ability.PID control, inverse control, fuzzy control and Generalized Dynamic Fuzzy Neural Network (GD-FNN) control were used to solve the control problem of electro-hydraulic servo system of mine sweeping plough. The simulation results showed that the system has overshoot and oscillation with PID control, the error of depth is too much with inverse control, while the other two controls have the high quality of performance, strong robustness.