基于高斯拟合的统计滤波算法及其应用

详细信息    查看官网全文

• 英文论文题名：Statistics Filter Based on Gaussian Fitting and its Application
• 论文作者：韩延明 ; 陈成 ; 张琨 ; 冯兴龙
• 英文论文作者：Han Yanming ; Chen Cheng ; Zhang Kun ; Feng Xinglong ; Brilliance Auto R&D Center(BARC)
• 年：2016
• 作者机构：华晨汽车工程研究院电子工程室;
• 论文关键词：传感器测量 ; 传感器参数标定 ; 统计滤波 ; 高斯函数 ; 最小二乘拟合
• 英文论文关键词：sensor measurement ; sensor calibration ; statistical filtering ; gaussian function ; the least-squares fitting
• 会议召开时间：2016-10-26
• 会议录名称：2016中国汽车工程学会年会论文集
• 英文会议录名称：Proceedings of 2016 SAE-China Congress
• 语种：中文
• 分类号：U463.6;TN713
• 学会代码：QCGC
• 会议名称：2016中国汽车工程学会年会暨展览会
• 会议地点：中国上海
• 主办单位：中国汽车工程学会
• 学会名称：中国汽车工程学会
• 页数：4
• 文件大小：424k
• 原文格式：O
• 会议级别：全国

摘要

本文着重介绍了基于高斯拟合的统计滤波算法(以下简称GFF),以及该算法在车载超声波传感器检测上的应用。该算法通过拟合高斯曲线的方式滤除噪声,与传统的平均值滤波方法相比,主要优点在于:①滤波结果由测量值的分布情况决定,从而抑制了数值变动对测量结果的影响,准确性和一致性得到提升,即在测量次数较少的情况下依然可以输出较稳定的滤波结果。②超声波传感器测量值的变化通常由大量随机因素组成,这与其他许多种类的传感器测量值类似,因此该算法的适用范围很广。③可以输出传感器的稳定性指标,用常规方法测量得到的传感器输出值即使是合格的,若稳定性指标超出范围也可以判定为不合格品。因此本方法增强了对产品特性的识别能力,具有很强的实用性。
A statistical filter algorithm based on Gaussian fitting is raised in this paper(Hereinafter referred to as the GFF).Its application in terms of vehicle- mounted ultrasonic sensors testing is illustrated.The algorithm can filter out the noise by fitting the Gaussian curve.The main advantages between GFF and traditional mean filtering methods are:1st,Filter results is determined by the distribution of the measurement data,this can inhibit the effects of data fluctuation and improve the measurement accuracy and consistency,even in the case of the number of the measurements are reduced,filter results remain stable.2nd,The cause of the sensor data fluctuation is consisted by many random factors,this is similar to many other kinds of sensor measurements,therefore,the scope of application of the algorithm is very wide.3rd,It can also output a parameter which can characterize the sensor stability,even though the measurement result of the sensor is qualified,it should be removed,as long as the stability parameter is out of range.Therefore,this method improved the ability to identify the products,has great practical value.

引文

[1]Steven W Smith.Digital Signal Processing:A Practical Guide for Engineers and Scientists[M].U S A,2003:26-32.
    [2]唐冲.基于Matlab的高斯曲线拟合求解[J].计算机与数字工程,2013,41(8):1262-1297.