基于神经网络自适应滤波的车辆动态称重系统研究  

Research of Vehicle Weigh-in-Motion System Based on Neural Network Self-adaptive Filtering

作　　者：张瑞[1] 吕文红[2] 张瑞玺 

机构地区：[1]山东科技大学信息科学与工程学院,山东青岛266510 [2]山东科技大学信电学院,山东青岛266510 [3]山东恒通化工股份公司化肥厂,山东郯城276100

出　　处：《公路交通科技》2010年第7期138-141,146,共5页Journal of Highway and Transportation Research and Development

基　　金：山东科技大学"春蕾计划"资助项目(2008BZC014)

摘　　要：针对车辆动态称重信号的特点,提出了一种车辆动态载重信号的处理方法。该方法采用基于神经网络的自适应滤波变步长LMS算法滤掉称重信号在各个频段内的噪声。在不同的环境下,对不同的车型,该技术适应性好,精度高,速度快。基于这种信号处理方法,开发了基于神经网络自适应滤波的车辆超载动态监测系统,选择高性能TMS32C2812芯片,设计了高效的软硬件系统,能够准确、及时测量高速公路上所经车辆的重量。According to the characteristics of vehicle weigh-in-motion signals, WIM signals was proposed. This method adopts self-adaptive filtering variable a new processing method of vehicle step size LMS algorithm based on neural network to filter out the noise in each band of WIM signal. It has good adaptability, high accuracy and high speed for different vehicle types in different circumstances. Based on this method of signal processing, a vehicle WIM system based on neural network self-adaptive filtering was developed, for which a high- performance TMS32C2812 chip was selected to design a high-performance software/hardware system to accurately and timely measure the weight of vehicle on expressway.

关 键 词：交通工程 交通监测 动态称重 自适应滤波 神经网络 DSP 

分 类 号：U491.116[交通运输工程—交通运输规划与管理]