基于灰色神经网络的汽车动态称重数据处理  

Vehicle Dynamic Weighing Data Processing Based on Grey Neural Network

作　　者：谭爽[1] 李丽宏[1] 

机构地区：[1]太原理工大学信息工程学院,太原030024

出　　处：《传感技术学报》2016年第8期1205-1209,共5页Chinese Journal of Sensors and Actuators

摘　　要：针对目前汽车动态称重方法称量精度低和所需数据样本数量大的缺陷,提出了一种新的动态称重数据处理方法,通过GM(1,1)模型对车辆动态称重数据进行预处理,获得每个称重数据的误差补偿量,建立了以车辆速度、加速度、动态称重残差序列为输入变量的灰色神经网络模型,使系统称量误差小于1%,称量准确度等级达到1级指标。研究表明,该方法在动态汽车衡数据处理中的可行性强,实现了在贫信息、少数据情况下对动态称重数据的高精度处理,为汽车动态称重系统精度的进一步提高提供了理论与技术支撑。Aiming at the defects of low vehicle dynamic weighing accuracy and requiring a large number of samples in vehicle dynamic weighing data processing currently,this paper presented a new way of vehicle dynamic weighing data processing,using GM（1,1）model to pretreat the data and obtain the error compensation of each weighing data, and a gray neural network model was established,which used vehicle speed,acceleration and dynamic weighing re？siduals as input variables,to make the system’s weighing error less then 1%,and the weighing accuracy attained the level-one index. The results showed that this method had strong feasibility in the dynamic truck scale data pro？cessing,and achieved high precision about data processing under the condition of poor information,which provided theoretical and technical support for further improving the accuracy of vehicle dynamic weighing system.

关 键 词：动态称重 数据处理 灰色神经网络 GM(1 1)模型 BP算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]