| 基于广义回归神经网络的车辆制动距离预测 |
| |
| 引用本文： | 强添纲,辛雨蔚,田广东,靳良真,魏文博,侯建.基于广义回归神经网络的车辆制动距离预测[J].森林工程,2014(1):73-75. |
| |
| 作者姓名： | 强添纲 辛雨蔚 田广东 靳良真 魏文博 侯建 |
| |
| 作者单位： | [1]东北林业大学交通学院,哈尔滨150040 [2]山东省潍柴重机股份有限公司,山东潍坊261000 |
| |
| 基金项目： | 中央高校基本科研业务费专项基金资助（DL11BB32） |
| |
| 摘 要： | 制动性能对车辆主动安全至关重要,其主要评价参数之一就是制动距离.以初速度、峰值附着系数、滑动附着系数和反应时间作为影响制动距离的主要因素变量,应用SIMULINK的制动模型,获得不同状况下的制动距离的结果,并将其作为神经网络的训练数据.基于广义回归神经网络（general regression neural network,GRNN）对于小样本能够实现精确预测的特点,提出应用其进行车辆制动距离的预测分析.结果表明：预测最大误差不超过11.50,预测相对误差不超过7.0％,表明预测精度较高.因此,应用GRNN可有效实现车辆制动距离的精确预测.
|
| 关 键 词： | 制动距离 广义回归神经网络 预测 |
Prediction of Vehicle Braking Distance Based on GRNN |
| |
| Qiang Tiangang,Xin Yuwei,Tian Guangdong,Jin Liangzhen,Wei Wenbo,Hou Jian.Prediction of Vehicle Braking Distance Based on GRNN[J].Forest Engineering,2014(1):73-75. |
| |
| Authors: | Qiang Tiangang Xin Yuwei Tian Guangdong Jin Liangzhen Wei Wenbo Hou Jian |
| |
| Institution: | 1. Traffic College, Northeast Forestry University, Harbin 150040; 2. Weichai Heavy Machinery Co. , Ltd. , Weifang 261000, Shandong Province) |
| |
| Abstract: | Brake performance is very important to vehicle active safety, and one of its evaluation parameters is braking distance. Taking the initial speed, adhesion coefficient, sliding adhesion coefficient and response time as main factor variables influencing the braking distance, combined with the brake model in SIMULINK, the results of braking distance under different conditions are ob- tained, which is taken as training data of neural network. Because the general regression neural network （GRNN） has the character- istic of exact prediction of small sample, it is used to forecast and analyze vehicle braking distance. The results show that the biggest prediction error is less than ll. 50 and the relative prediction error is less than 7.0% , which means that the prediction accuracy is high and good. Therefore, the GRNN can realize the exact prediction of vehicle braking distance. |
| |
| Keywords: | braking distance general regression neural network prediction |
| 本文献已被 维普 等数据库收录！ |
|
