基于小波包分解的阈值消噪在车载CAN总线上的应用  

Application of threshold denoising based on wavelet packet decomposition to vehicular CAN-bus

作　　者：高印寒[1] 马喜来[2] 何丽桥[3] 杨开宇[1] 周晓华[2] 

机构地区：[1]吉林大学测试科学实验中心,吉林长春130025 [2]吉林大学仪器科学与电气工程学院,吉林长春130025 [3]吉林大学物理学院,吉林长春130025

出　　处：《光学精密工程》2007年第3期434-439,共6页Optics and Precision Engineering

基　　金：高等学校博士学科点专项科研基金资助项目(No.20050183019);吉林省科技发展计划(No.20030527)

摘　　要：为了实现在车载CAN总线控制器故障产生不可预知干扰信号的情况下正确读出CAN总线上的故障码,提出了一种基于小波包分解的阈值消噪方法。将信号进行基于最小Shannon熵原理的最佳树小波包分解后,针对CAN总线传输信号特点,对高频和低频系数采取不同阈值处理的混合阈值消噪处理。对于高频系数,采用基于Stein无偏似然估计原理的自适应阈值;在干扰信号的统计特性未知的情况下,对低频系数阈值的选取采取保持频率单一和能量损耗小的双重标准。最后,由阈值处理后的系数重构信号,得到正确的传输码。实验结果表明,该混合阈值消噪处理方法对消除这种由CAN总线上的控制器故障引起的干扰,较传统的消噪技术提高1/3以上的信噪比,增加了后期读取CAN总线故障码的准确率。In order to identify the correct failure code in a vehicular CAN-bus disturbed by the unforeseen damage of the controllers, application of threshol denosing based on wavelet packet decomposition was proposed here. The disturbed signal was decomposed into the multiscale wavelet domain by minimal Shannon entropy criteria. Then, the threshold based on Stein＇s unbiased risk estimation was used to eliminate the noise in the high frequency coefficients and the double criteria,and holding frequency singleness and little energy losses were used to set the threshold of the low frequency coefficients. Finally, the denoised signal was achieved by the invert WPD. Experimental results show that the S/N radio of the denoising technique in proposed method has been increased 1/3 over that of tranditional method,which can effectively get the correct facult-code while the some controller is damaged in CAN bus.

关 键 词：小波包分解 最佳树 混合阚值 风险函数 CAN总线 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]