基于遗传-支持向量回归的车内稳态噪声声品质预测  

Sound Quality Prediction of Steady Internal Noise of VehiclesBased on GA-SVR

作　　者：朱仝 郑松林[1] 袁卫平[2] ZHU Tong;ZHENG Songlin;YUAN Weiping(School of Mechanical Engineering,Shanghai University of Science and Technology,Shanghai 200093,China;SAIC Motor Commercial Vehicle Technique Center,Shanghai 200438,China)

机构地区：[1]上海理工大学机械工程学院,上海200093 [2]上海汽车集团股份有限公司商用车技术中心,上海200438

出　　处：《噪声与振动控制》2020年第3期170-174,193,共6页Noise and Vibration Control

摘　　要：为了稳定、精确地评价车内稳态噪声声品质,以车内稳态噪声为研究对象,进行主观评价试验,计算客观心理声学参数并完成了相关性分析。建立基于支持向量回归(Support Vector Regression,SVR)的车内稳态噪声声品质预测模型,并使用遗传算法(Genetic Algorithm,GA)对支持向量回归的超参数进行优化。其后建立基于反向传播神经网络(Back Propagation Artificial Neural Network,BPANN)的声品质预测模型。对比分析发现遗传-支持向量回归(GASVR)模型预测精度高于BP神经网络。结果表明,遗传-支持向量回归适用于车内稳态噪声声品质预测,能够较大提高车内稳态噪声声品质预测精度和工程效率。In order to accurately evaluate the sound quality of internal noise of vehicles,subjective test and psychoacoustic metrics calculation were carried out for the correlation analysis between subjective and objective evaluations.On this basis,a sound Quality evaluation model based on Support Vector Regression(SVR),which parameters were optimized by Genetic Algorithm(GA),was built.Then,another sound quality prediction model based on Back Propagation Artificial Neural Network(BPANN)was established.Through the comparative analysis,the GA-SVR model was found to have better performance than the BPANN model.The result shows that the GA-SVR is suitable for establishing the sound quality evaluation model for steady internal noise prediction of vehicles with high accuracy and efficiency.

关 键 词：声学 车内稳态噪声 声品质 支持向量回归 遗传算法 主客观评价 

分 类 号：TB535[理学—物理] U467.4[理学—声学]