作物需水量自适应神经网络模糊系统的设计研究

通过自适应神经网络对大量实验数据(太阳辐射、空气湿度)的学习，得到学习后的作物需水量模糊系统的隶属度函数和模糊规则，所建立的Sugeno型模糊推理系统通过与自适应技术的结合，使建立的模糊模型很好地匹配了输入数据。该模糊模型能很好地解决需水量多影响因素之间的不确定性和非线性，模型的预测精度较高，对精确灌溉和节水农业不仅有一定的理论意义，更具有巨大的实用价值。

Study on Self-Adaptive Neural Network Fuzzy System Design of Crop Water Requirement

In this article, through training of large quantity of experiment dada on solar radiation and air humidity by using self-adaptive neural network method (SANN), the subordinate degree function and fuzzy rules of crop water requirement fuzzy model are derived. The Sugeno-type fuzzy inference system established is integrated with the SANN technique to make the fuzzy model developed well adaptive to the experiment data entered. The fuzzy model can solve the uncertainty and non-linearity of multiple influencing factors of water requirement, and give high precision forecasting. So the model developed has not only theoretical significance, but also great practical value for precise irrigation and water-saving agriculture.