基于模糊逻辑的数字滤波器设计

针对传统滤波器单一模型设计的缺点,提出一种将滤波噪声统计特性与模糊逻辑相结合的数字滤波器设计方法.该方法根据噪声统计特性,利用缩短三角型模糊函数,确定模糊逻辑滤波器的参数.通过滤波噪声标准偏差与最大滤波噪声门限的比较,提出了分别采用单值滤波和均值滤波的方法,并推导了均值滤波方法中的周期抽样数目.给出了算法具体设计步骤,包括:滤波方法选择、连续信号周期采样、模糊化数据处理和去模糊化.仿真表明,基于模糊逻辑的数字滤波器对不同标准偏差的噪声均有良好的滤波性能,并且设计经济合理,时间开销小,软硬件易实现,有工程应用价值.

Digital filter design via fuzzy logic

Aimed at the shortcoming of single model for traditional filter design,a new digital filter design method is introduced,which combines the statistical property of noise to be filtered with the fuzzy logic theory.The truncated triangular fuzzy function is used to define the parameters of the digital filter with the noise statistics.According to the comparison between the standard deviation of input noise and the maximal threshold of filtering noise,filters both via single value and via statistical averaging are proposed respectively.The sample numbers of a period is established for statistical averaging filters.The proposed algorithm contains four steps: filtering method selection,period signal sampling,signal fuzzification and signal defuzzification.The simulation results indicate that the fuzzy logic digital filter is powerful for different standard deviations and might be valuable in project applications.