| 摘 要: 针对电动汽车内部噪声特性变化,构建适用于电动汽车的声品质评价预测模型。对预处理的车内噪声样本进行主客观评价分析,筛选出有效的主观评价结果,并利用随机森林特征分析,提取车内噪声客观评价特征,构建模型样本库。为提高预测精度和泛化能力,提出基于改进哈里斯鹰算法(IHHO)的支持向量机(SVM)模型。对比SVM、HHO-SVM和IHHO-SVM3个模型匀速和加速工况下的均方误差(MSE)和决定系数(R2)。其中,IHHO-SVM的R2分别为0.983和0.984,预测结果的相对误差更低;MSE分别为0.056和0.012。以上结果验证了IHHO-SVM 模型在电动汽车声品质评价中的优越性。 |
| 关键词: 电动汽车 声品质 哈里斯鹰算法 SVM模型 评价系统 |
|
中图分类号: TP277
文献标识码: A
|
| 基金项目: 海洋防务技术创新中心创新基金项目 |
|
| Research on an IHHO-SVM-Based Evaluation Model for In-Vehicle Acoustic Quality of Electric Vehicle |
|
WANG Jie, QIU Yiyang, LIU Tianlun, LI Mingrong,DING Yuxuan, XIA Zhouyang
|
(School of Automobile and Traffic Engineering, Wuhan University of Science and Technology, Wuhan 430065, China)
wangjie1980@wust.edu.cn; 645753223@qq.com; 1941851860@qq.com; 2114646362@qq.com; 1064636042@qq.com; 1366180422@qq.com
|
| Abstract: To address the changing characteristics of internal noise in electric vehicles (EVs), this study develops a predictive model for evaluating in-vehicle acoustic quality. Pre-processed noise samples underwent subjective-objective evaluation analysis, from which valid subjective assessment results were screened. Random forest feature analysis was then employed to extract objective evaluation features of the noise, establishing a model sample library. To enhance prediction accuracy and generalization capability, a Support Vector Machine (SVM) model optimized by the Improved Harris Hawks Optimization (IHHO) algorithm is proposed. Comparing the Mean Squared Error (MSE) and coefficient of determination (R2) of three models—SVM, HHO-SVM, and IHHO-SVM-under constant—speed and acceleration
conditions, the IHHO-SVM achieved R2 values of 0.983 and 0.984, respectively, with lower relative prediction errors.Its MSE values (0.056 and 0.012) further validate the superiority of the IHHO-SVM model for EV acoustic quality evaluation. |
| Keywords: electric vehicles acoustic quality Harris Hawks Optimization (HHO) algorithm SVM model evaluation system |
