基于Kalman滤波的Fuzzy系统参数估计

Fuzzy system parameter estimation based on Kalman filtering

模糊系统在现代工业、模式识别与人工智能、自动控制中起着重要的作用。论文研究了复模糊系统的参数估计 ,提出了用经典的 Kalman滤波来估计模糊系统参数的一种新方法。在实例分析中 ,将复模糊系统用于通信信道均衡 ,用专家知识对信道输出建立一个模糊规则库 ,可以提高算法的速度和准确率。用该文给出的经典的 Kalm an滤波与传统的迭代最小二乘 RL S法比较 ,结果是该文的方法不仅误码率比 RL S方法低 ,而且收敛速度更快 ,更加利于硬件实现。

Fuzzy systems play an important role in modern industry. This paper analyzed applications of complex fuzzy systems for parameter estimation and propose a new method for estimating fuzzy system parameters utilizing Kalman filtering. The method was applied to computer simulated channel equalization with linguistic fuzzy IF THEN rules used to further improve the algorithm's performance. The method converges faster than RLS method with a lower bit error ratio (BER).