基于数据驱动的模糊系统辨识研究

Research on Fuzzy System Identification Using Data Drive

针对实际中某些过程无法确定其精确表达式的问题,研究了基于输入-输出数据的模糊建模方法。采用Mamdani模型和模糊基函数,以系统的输入-输出数据为基础,通过One-Pass、误差反向传播、查表法、最小二乘法这4种基于数据驱动的建模方法分别建立了Mackey-Glass混沌系统的模糊模型,用Mackey-Glass混沌数据验证了它们的有效性和实用性。对这4种方法的性能和适用场合做了分析说明,为实际过程的建模提供参考依据。在实际中,可根据需要选择合适的建模方法。

Due to the difficulty in getting the exact expression of some actual processes, fuzzy modeling method by input-output data is discussed in this paper. One-pass, back propagation, seeking table and least square are used to build Mackey-Glass chaos fuzzy model by Marndani fuzzy model and fuzzy basic function, which based on input-output datas. The effectiveness and practicality of these four methods are illustrated by Mackey-Glass chaos data. The performance and application conditions of these four methods are stated, providing reference for actual modeling. In practice, a suitable modeling method can be selected according to actual process.