模糊C-均值聚类算法在动态汽车衡中的应用

Application of Fuzzy C-means Clustering Algorithm in Dynamic Truck Scales

为了提高动态汽车衡的测量精度,针对测量数据的信号处理问题,采用小波变换对动态汽车衡测量数据进行数字滤波。小波变换不但能滤除测量数据中的噪声信号,而且能很好地保留信号的突变部分。同时,通过提取5尺度小波系数作为模糊C-均值聚类算法的聚类样本,有效识别出最接近车辆实际质量的有用称重数据,提高了称重精度。试验结果表明,采用小波变换对动态汽车衡测量数据进行阀值滤波,并利用模糊C-均值聚类算法识别有用数据,对提高车辆称重的精度具有良好效果。

In order to improve the measurement precision of dynamic truck scales, aiming at the issue of signal processing for measuring data, digital filtering for the data is conducted by using wavelet transform. The method eliminates the noise signal in measuring data, and well retains the mutant signal. In addition, through picking up the 5-sclae wavelet coefficient as the clustering sample for C-means clustering algorithm, the useful weighing data that mostly close to the actual mass of the vehicle can be identified, thus the weighing precision is enhanced. The experimental result shows that excellent effect of enhancing weighing precision is oblained by using wavelet transform to conduct threshold filtering for the measuring data, and using fuzzy C-means clustering algorithm to identify useful data.