基于EMD-RBF网络的车辆动态称重信号处理方法

针对动态称重数据处理的复杂性,以常见的两轴车辆为研究对象,对汽车轴完全驶上称重台的信号进行经验模分解(EMD)求得剩余分量的平均值,并以前轴、后轴的剩余分量的平均值、平均速度、上台信号上升斜率、下台信号下降斜率为径向基函数(RBF)网络的输入,利用RBF网络对系统进行非线性建模。以静态测量结果为相对真值进行仿真,其出现的最大相对误差为1.4%,而简单平均值的测试误差达到2.1%,结果表明:该方法比直接输入台上信号的平均值具有更高的测试精度。

Signal analysis method of vehicle weigh-in-motion based on EMD-RBF network

In allusion to the complication of processing dynamic weighing data,the expirical mode decomposition(EMD) method is used to obtain the average residue.The experiments concentrate on two-axle vehicle.The mean residue of front,the one of rear axle,the velocity,the ascending slope and descending slope of signal on the plate entirely are inputed into the RBF network.The system composed by vehicle and weighing plate is non-linear.The static weight signals are used for simulating relative real value.The relative error is decreased from(2.1 %)to 1.4 %.The simulation results show that higher precision can be achieved instead of simple average.