The Study On Greenhouse Automatic Control Using Fuzzy Neural Networks

Equipment agriculture is a burgeoning industry, which can make full use of resource and have high-level of automation. It not only is technique and intelligence concentrated but also will bring you satisfy appreciation profit. Modern automatic control greenhouse is one of the realization forms of facility agriculture, which can make a man-made climate of the greenhouse or eliminate the disadvantages for crop. In such a greenhouse, the produce period of crop can be shortened. Meanwhile the aim of high-quality, high-output, high-benefit can be reached.Greenhouse is a multi-variable coupled time-varying and non-linear system. Its environment control is not only one of the key techniques in modern automatic control but also difficult to control the greenhouse climate. The modeling of greenhouse climate is the key to realizing greenhouse control. This paper has analyzed the main thermodynamics processes in the climate of the greenhouse systematically at first, on the basis of consulting foreign research results, direct against the regulation and control characteristic of the existing greenhouse of our country, have proposed containing the dynamic model of temperature and humidity of greenhouse which controls quantity; And then analyzed in detail the climate in the greenhouse system of the automatic control controls the question, including influence each main environmental factor and the target's control characteristic that the crop grows. On the basis of analyzing that is accused of the target's complexity in the greenhouse and under the prerequisite that the function resolve, design overall the environmental control system of the greenhouse, propose controlling the scheme in the room in summer and winter. Generally speaking, this chapter is about controlling the research of the method only as a kind of new try, wait further deepening.This text finally presents a kind of fuzzy neural network method which combines fuzzy knowledge representation with neural network self-learning ability and adopts a fast study algorithm. For the greenhouse heating systematic control in winter, I have been going on the simulation compare with the FNNC, routine PID control and fuzzy separately by the SIMULINK platform of MATLAB. In the end we can learn from the artificial result that FNNC has lighter exceeding suiting well to control, no shaking, fine stationarity, short time to reach the stable state, little error of the stable state. So its dynamic characteristic and static characteristic is the most superior. Therefore it has verified that the feasibility of using FNNC.