基于模糊小波网络的伺服系统参数辨识研究

伺服系统大多是非线性系统,难以对其建立准确的控制模型,其位置、和速度检测信号易受干扰,而小波网络具有多分辨率特性和任意逼近能力。利用其非线性映射能力对系统的输入输出关系进行模拟,将小波网络和模糊规则结合对系统的位置和速度进行辨识,动态调整网络的权值W和模糊规则,将非线性映射的问题转化为求解系统最优解,从而产生一种新的系统辨识方法,并以永磁伺服系统为例,设计了辨识的结构模型和策略,实验表明该算法可以达到较高的系统控制要求。

Servo System Parameter Identification based on Fuzzy and Wavelet Network

The servo system is mostly nonlinear system, and is difficult to establish its accurate control model. Its positions and speed signals am susceptible to interferenoe. Wavelet networks have multi-resolution characteristics and any approximation abil- ity. It simulates the input-output relationship of the system by nonlinear mapping capability. Wavelet network combined with fuzzy rules is applied to identify the system lmsition and speed, and the network weights of W and fuzzy rules dynamically ad- justed. So nonlinear mapping is transformed into solving the optimization problem. It is a new system identification method. Taking magnet servo system for example, we design the identification of the structural model and strategy. The experiments show that the algorithm can meet strict system control requirements.