基于BP神经网络的路面不平度检测与仿真  

Road Roughness Detection and Simulation based on BP Neural Network

作　　者：崔丹丹[1] 张才千[1] 韩东[2] 

机构地区：[1]开封大学软件学院,河南开封475004 [2]海军大连舰艇学院信息与通信工程系,辽宁大连116018

出　　处：《计算机仿真》2014年第5期162-166,共5页Computer Simulation

基　　金：国家自然科学基金项目(11374001)

摘　　要：在对路面不平度优化检测问题的研究中,由于路面不平使车辆振动加剧,严重影响了路面的使用寿命和乘坐的舒适性。功率谱密度是评价路面不平度的常用指标,为了识别分析路面功率谱密度,提出了一种采用BP神经网络的路面不平度检测方法。以四自由度车辆振动模型为基础,把ADAMS/CAR中车辆平顺性仿真得到的汽车质心垂直加速度谱和俯仰角加速度谱为输入样本,以路面功率谱密度为输出样本,应用BP神经网络建立非线性映射。将仿真数据代入已训练好的网络中进行路面功率谱识别,仿真结果表明:上述方法识别出的功率谱密度与实际功率谱密度的平均误差仅为1.23%,具有较强的抗噪声能力和较理想的识别精度。The road roughness increases the vibration of the vehicle which seriously affects the life of the road and ride comfort. In order to identify and analyse road surface power spectral density, a method based on BP neural network to detect road roughness was proposed. The four degrees of freedom vehicle vibration model was used as the foundation, and the vertical acceleration spectrum of the vehicle center of mass and pitch acceleration spectrum were obtained by simulation in ADAMS/CAR as the input samples and the road surface power spectrum density as output samples, the nonlinear mapping was found by the application of BP neural network. Another simulation input data were used in the trained network as the road spectrum identification. The results show that this method has better ability of anti-noise and ideal identification accuracy, and the road surface spectrum of identification fits the imitated road surface spectrum.

关 键 词：路面不平度 功率谱密度 车辆振动模型 检测与仿真 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]