CAJ | 학술논문

以乘用车由50 km／h 加速到100 km／h 时的噪声信号为评价对象，用成对比较法对车内加速噪声品质偏好性进行主观评价实验，获得每个样本的偏好性评价值。计算各噪声样本的主要心理声学客观参数并进行相关分析。鉴于评价者对非稳态噪声主观评价过程的复杂性，建立支持向量机（Support Vector Machine，SVM）的主客观评价模型，并利用粒子群优化算法（Particle Swarm Optimization，PSO）对模型参数进行优化。为对比优化后预测效果，建立 BP 神经网络回归模型。结果表明，优化后的粒子群－向量机回归模型用于噪声声品质评价能获得更好的预测效果，可较大程度提高声品质预测精度。
이승용차유50 km／h 가속도100 km／h 시적조성신호위평개대상，용성대비교법대차내가속조성품질편호성진행주관평개실험，획득매개양본적편호성평개치。계산각조성양본적주요심리성학객관삼수병진행상관분석。감우평개자대비은태조성주관평개과정적복잡성，건립지지향량궤（Support Vector Machine，SVM）적주객관평개모형，병이용입자군우화산법（Particle Swarm Optimization，PSO）대모형삼수진행우화。위대비우화후예측효과，건립 BP 신경망락회귀모형。결과표명，우화후적입자군－향량궤회귀모형용우조성성품질평개능획득경호적예측효과，가교대정도제고성품질예측정도。
The noise signals from passenger car during acceleration from 50km /h to 100km /h were selected as the evaluation object and subjective testings were carried out via the paired comparison method to get the annoyance scores. The BK software was used to calculate the main psychoacoustics parameters and then to find the correlation between them. In view of the complexity of subjective evaluation process for nonstationary noise,support vector machine (SVM)model was used to simulate the process of noise subjective evaluation,and then the input parameters of support vector machine were optimized by using the particle swarm optimization (PSO).In order to compare the prediction effect after optimization,a BP neural network was established at the same time.The results show that the quality evaluation of the acceleration noise can get better prediction effect with PSO -SVMmodel.It could largely improve the predictive accuracy of sound quality evaluation and has reference value to the evaluation and analysis of vehicle noise.