Rapid prediction of vehicle interior wind noise via a hybrid convolutional neural network-transformer model and geometric styling features

  • Hongwei Yi

    Vehicle Measurement Control and Safety Key Laboratory of Sichuan Province, Xihua University, Chengdu 610039, China; Engineering Research Center of Intelligent Control and Simulation Test Technology for New Energy Vehicles of Sichuan Province, Xihua University, Chengdu 610039, China
  • Penghu Li

    School of Mechanical Engineering, Southwest Jiaotong University, Chengdu 610031, China
  • Jifeng Wang

    School of Mechanical Engineering, Southwest Jiaotong University, Chengdu 610031, China
  • Yuwei Deng

    Global R&D Center, China FAW Corporation, Limited, Changchun 130013, China
  • Xiaorong Huang orcid

    Vehicle Measurement Control and Safety Key Laboratory of Sichuan Province, Xihua University, Chengdu 610039, China; Engineering Research Center of Intelligent Control and Simulation Test Technology for New Energy Vehicles of Sichuan Province, XihuaUniversity, Chengdu 610039, China
  • Haibo Huang

    School of Mechanical Engineering, Southwest Jiaotong University, Chengdu 610031, China
Article ID: 4536
DOI: https://doi.org/10.59400/sv4536
Keywords: interior wind noise; styling features; convolutional neural network; transformer

Abstract

The rapid evaluation of interior aerodynamic noise during the Concept A Surface design stage is important for vehicle acoustic development, but conventional methods are limited by high cost and low efficiency. This study proposes a convolutional neural network transformer-based prediction method for vehicle interior wind noise by integrating vehicle styling features and acoustic technical parameters. An optimal Latin hypercube sampling method was used to generate design combinations, and wind tunnel tests were conducted at 120 km/h. Key vehicle styling parameters, including A-pillar geometry, side mirror dimensions, front windscreen angle, side mirror-to-body spacing, and side window inclination, together with glazing material properties, glass thickness, acoustic transfer function, and interior reverberation time, were selected as input features to predict the driver’s left-ear wind noise spectrum. Based on five-fold cross-validation, the proposed model was compared with convolutional neural network (CNN), long short-term memory (LSTM), and transformer models. The CNN-Transformer model achieved the best performance, with mean absolute percentage error (MAPE) and root mean square error (RMSE) values of 2.23% and 0.94 dB, respectively. Compared with the transformer, CNN, and LSTM models, the proposed method reduced MAPE by 24.91%, 36.29%, and 49.59%, and reduced RMSE by 22.95%, 35.17%, and 48.07%, respectively. The model also maintained reliable performance on an independent test set, with MAPE and RMSE values of 4.82% and 1.44 dB. The mean impact value method was further applied to identify the influence of design parameters on interior wind noise, guiding for early vehicle acoustic optimization.

Published
2026-08-15
How to Cite
Yi, H., Li, P., Wang, J., Deng, Y., Huang, X., & Huang, H. (2026). Rapid prediction of vehicle interior wind noise via a hybrid convolutional neural network-transformer model and geometric styling features. Sound & Vibration, 60(5). https://doi.org/10.59400/sv4536
Issue
Vol. 60 No. 5 (2026)
Section
Article

Copyright (c) 2026 Hongwei Yi, Penghu Li, Jifeng Wang, Yuwei Deng, Xiaorong Huang, Haibo Huang

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

References

[1]Pang J, Mao T, Jia W, et al. Prediction and Analysis of Vehicle Interior Road Noise Based on Mechanism and Data Series Modeling. Sound & Vibration. 2024; 58(1): 59–80. doi: 10.32604/sv.2024.046247

[2]Gad AS, Darakhshan Jabeen S, El-Zomor HM, et al. Optimizing Ride Comfort and Stability in In-Wheel Motor Drive Electric Vehicles: A Correlated Semi-Active Suspension System with Kalman Filter and Spring-Type Analysis. SAE International Journal of Vehicle Dynamics, Stability, and NVH. 2026; 10(4). doi: 10.4271/10-10-04-0030

[3]Wang T, Liu Z, Li X, et al. An end-to-end data-driven framework for intelligent evaluation of automotive noise and vibration performance. Engineering Applications of Artificial Intelligence. 2026; 181: 115422. doi: 10.1016/j.engappai.2026.115422

[4]Zhang Y, He T, Wang Y, et al. Research on the Influence of Door and Window Sealing on Interior Wind Noise Based on Statistical Energy Analysis. SAE International Journal of Vehicle Dynamics, Stability, and NVH. 2025; 9(3): 423–437. doi: 10.4271/10-09-03-0026

[5]Liu K, Liao Y, Wang H, et al. Road Noise Improvement Method for Battery Electric Vehicles Based on Bushing Stiffness Optimization. SAE International Journal of Vehicle Dynamics, Stability, and NVH. 2025; 9(4): 555–569. doi: 10.4271/10-09-04-0032

[6]Song D, Hong S, Ha C, et al. Acoustic analysis and data-driven control of vehicle NVH: A framework for manufacturing process optimization. Applied Acoustics. 2025; 233: 110618. doi: 10.1016/j.apacoust.2025.110618

[7]Yang M, Dai P, Yin Y, et al. Predicting and optimizing pure electric vehicle road noise via a locality-sensitive hashing transformer and interval analysis. ISA Transactions. 2025; 157: 556–572. doi: 10.1016/j.isatra.2024.11.059

[8]Mahmood A, Cockerill T, de Boer G, et al. Heat transfer modeling and optimal thermal management of electric vehicle battery systems. Energies. 2024; 17(18): 4575. doi: 10.3390/en17184575

[9]Wang Y, Du M, Su C, et al. Numerical investigation on the contribution of underbody flow-induced noise on vehicle interior noise. Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering. 2021; 235(10–11): 2667–2678. doi: 10.1177/09544070211004460

[10]Duroyon M, Susini P, Misdariis N, et al. A lexicon to describe specific sounds of the electric car cabin: A verbal approach to comfort improvement. Acta Acustica. 2026; 10: 21. doi: 10.1051/aacus/2026013

[11]Masri J, Amer M, Salman S, et al. A survey of modern vehicle noise, vibration, and harshness: A state-of-the-art. Ain Shams Engineering Journal. 2024; 15(10): 102957. doi: 10.1016/j.asej.2024.102957

[12]Jin T, Liu X, Chen C, et al. The impact of environmental noise on drivers' cognitive abilities: A case study on in-vehicle voice interaction interfaces. Applied Ergonomics. 2024, 117: 104247. doi: 10.1016/j.apergo.2024.104247

[13]Zhu H, Zhao J, Wang Y, et al. Improving of pure electric vehicle sound and vibration comfort using a multi-task learning with task-dependent weighting method. Measurement. 2024; 233: 114752. doi: 10.1016/j.measurement.2024.114752

[14]Yang Y, Feng Y, Easa SM, et al. Sound effects on physiological state and behavior of drivers in a highway tunnel. Frontiers in Psychology. 2021; 12: 693005. doi: 10.3389/fpsyg.2021.693005

[15]Fan D, Dai P, Yang M, et al. Research on maglev vibration isolation technology for vehicle road noise control. SAE International Journal of Vehicle Dynamics, Stability, and NVH. 2022; 6(3): 233–245. doi: 10.4271/10-06-03-0016

[16]Xiang X, Yi H, Hou J, et al. Uncertainty Optimization of Vibration Characteristics in Engineering Machinery Powertrain Mounting Systems Using Monte Carlo and Genetic Algorithms. SAE International Journal of Vehicle Dynamics, Stability, and NVH. 2026; 10(1): 109–128. doi: 10.4271/10-10-01-0006

[17]Wan Y, Xia J, Yang L, et al. Clear imaging beamforming method for the noise source identification in car cabin. Sound & Vibration. 2025; 59(1): 1839. doi: 10.59400/sv1839

[18]Jamaluddin NS, Oettle N, Staron D. Experimental Investigation into Modulated Aeroacoustic Cabin Noise Arising from Unsteady Flow Conditions. SAE International Journal of Advances and Current Practices in Mobility. 2025; 7(6): 2925–2939. doi: 10.4271/2025-01-0029

[19]Yu J, Davies P. Development and verification of a method to simulate non-stationary vehicle interior wind noise. The Journal of the Acoustical Society of America. 2024; 155(3_Supplement): A198. doi: 10.1121/10.0027292

[20]Tajima A, Ikeda J, Tsubokura M. Modulated wind noise in a road vehicle due to atmospheric fluctuations and reduction measures. Journal of Wind Engineering and Industrial Aerodynamics. 2025; 266: 106210. doi: 10.1016/j.jweia.2025.106210

[21]Guleria A, Novacek J, Ihi R, et al. Validation of Wind Noise for Class-8 Truck Using Lattice Boltzmann Method. SAE International Journal of Commercial Vehicles. 2025; 18(3): 391–408. doi: 10.4271/02-18-03-0021

[22]Josefsson E, Sebben S, Urquhart M. Characterisation of the flow around passenger vehicle wheels with varying tyre profiles. International Journal of Heat and Fluid Flow. 2023; 103: 109191. doi: 10.1016/j.ijheatfluidflow.2023.109191

[23]Nusser K, Becker S. Numerical investigation of the fluid structure acoustics interaction on a simplified car model. Acta Acustica. 2021; 5: 22. doi: 10.1051/aacus/2021014

[24]Jia Q, Zhang J, Liang W, et al. A mathematical‐boundary‐recognition domain‐decomposition Lattice Boltzmann method combined with large eddy simulation applied to airfoil aeroacoustics simulation. International Journal for Numerical Methods in Fluids. 2024; 96(7): 1250–1275. doi: 10.1002/fld.5287

[25]Guo Z, Lei Z, Zeng K, et al. Non-integer-dimensional architected materials enabling synergistic acoustic, mechanical, and fluid coupling. Materials Horizons. 2026; 13(2): 748–762. doi: 10.1039/d5mh01768h

[26]Guo Z, Du L, Yang Z, et al. Switchable bidirectional sound absorption via exceptional point modulation in acoustic metastructures with interleaved resonator coupling. Advanced Science. 2025; 12(44): e08951. doi: 10.1002/advs.202508951

[27]Cho H, Jeong J, Jang J, et al. VAE-based surrogate model of rotor aerodynamic load distribution for noise prediction. Aerospace Science and Technology. 2026; 178: 112412. doi: 10.1016/j.ast.2026.112412

[28]Huang Y, Boerboom M, Wolff K, et al. Find Optimal Suspension Kinematics Targets for Vehicle Dynamics Using Reinforcement Learning. SAE International Journal of Vehicle Dynamics, Stability, and NVH. 2025; 10(1): 25–41. doi: 10.4271/10-10-01-0002

[29]Zhang Q, Wang X, Yang D, et al. Data-driven prediction of aerodynamic noise of transonic buffeting over an airfoil. Engineering Analysis with Boundary Elements. 2024; 163: 549–561. doi: 10.1016/j.enganabound.2024.04.006

[30]Zhao Y, Liu H, Guo W, et al. Research on sound quality prediction of vehicle interior noise using the human-ear physiological model. The Journal of the Acoustical Society of America. 2024; 156(2): 989–1003. doi: 10.1121/10.0028130

[31]Qi Z, Liu P, Duan H, et al. RDAS: A Shared Control Strategy Considering Real-Time Driving Ability in Driver Takeover Tasks. SAE International Journal of Vehicle Dynamics, Stability, and NVH. 2026; 10(2): 303–326. doi: 10.4271/10-10-02-0014

[32]Yang H, Yuan W, Zhu W, et al. Wind turbine airfoil noise prediction using dedicated airfoil database and deep learning technology. Applied Energy. 2024; 364: 123165. doi: 10.1016/j.apenergy.2024.123165

[33]Roznowicz D, Stabile G, Demo N, et al. Large-scale graph-machine-learning surrogate models for 3D-flowfield prediction in external aerodynamics. Advanced Modeling and Simulation in Engineering Sciences. 2024; 11(1): 6. doi: 10.1186/s40323-024-00259-1

[34]Hou J, Yi H, Xiang X, et al. Identification of vehicle suspension shock absorber rattle noise based on wavelet packet feature fusion and GWO-LSTM. Sound & Vibration. 2025; 59(2): 1941. doi: 10.59400/sv1941

[35]Zhao J, Yin Y, Chen J, et al. Evaluation and prediction of vibration comfort in engineering machinery cabs using random forest with genetic algorithm. SAE International Journal of Vehicle Dynamics, Stability, and NVH. 2024; 8(4): 491–512. doi: 10.4271/10-08-04-0027

[36]Dai R, Zhao J, Zhao W, et al. Exploratory study on sound quality evaluation and prediction for engineering machinery cabins. Measurement. 2025; 253: 117684. doi: 10.1016/j.measurement.2025.117684

[37]Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need. Advances in Neural Information Processing Systems. In: Proceedings of the 31st Conference on Neural Information Processing Systems (NIPS 2017); 4–9 December 2017; Long Beach, CA, USA.

[38]Islam S, Elmekki H, Elsebai A, et al. A comprehensive survey on applications of transformers for deep learning tasks. Expert Systems with Applications. 2024; 241: 122666. doi: 10.1016/j.eswa.2023.122666

[39]Soydaner D. Attention mechanism in neural networks: Where it comes and where it goes. Neural Computing and Applications. 2022; 34(16): 13371–13385. doi: 10.1007/s00521-022-07366-3

[40]Hosseinzadeh R, Sadeghzadeh M. Attention mechanisms in transformers: A general survey. Journal of AI and Data Mining. 2025; 13(3): 359–368. doi: 10.22044/jadm.2025.15584.2679

[41]Thirunavukarasu R, Kotei E. A comprehensive review on transformer network for natural and medical image analysis. Computer Science Review. 2024; 53: 100648. doi: 10.1016/j.cosrev.2024.100648

[42]Ampazis N, Sakketou F. Diversifying multi-head attention in the transformer model. Machine Learning and Knowledge Extraction. 2024; 6(4): 2618–2638. doi: 10.3390/make6040126

[43]Sanger TD. Transformer Models for Signal Processing: Scaled Dot-Product Attention Implements Constrained Filtering. Neural Computation. 2025; 37(10): 1839–1852. doi: 10.1162/neco.a.29

[44]Su L, Zuo X, Li R, et al. A systematic review for transformer-based long-term series forecasting. Artificial Intelligence Review. 2025; 58(3): 80. doi: 10.1007/s10462-024-11044-2

[45]Cong S, Wang H, Zhou Y, et al. Comprehensive review of Transformer‐based models in neuroscience, neurology, and psychiatry. Brain‐X. 2024; 2(2): e57. doi: 10.1002/brx2.57

[46]Isomura T, Shimizu R, Goto M. Sparse attention is all you need for pre-training on tabular data. Neural Computing and Applications. 2025; 37(3): 1509–1522. doi: 10.1007/s00521-024-10698-x

[47]Lee A, Hwang E, Kim D. A Practical CNN–Transformer Hybrid Network for Real-World Image Denoising. Mathematics. 2026; 14(1): 203. doi: 10.3390/math14010203

[48]Gao L. CNN-Transformer Hybrid Models for Object Detection: A Comprehensive Review. Science and Technology of Engineering, Chemistry and Environmental Protection. 2025; 1(4).

[49]Wang Z, Liu J, Sun H, et al. A CNN–LSTM–SVM Framework for Mapping Driving Styles to Adaptive Suspension Modes. SAE International Journal of Vehicle Dynamics, Stability, and NVH. 2026; 10(2): 237–258. doi: 10.4271/10-10-02-0011

[50]Kim Y. Convolutional neural networks for sentence classification. In: Moschitti A, Pang B, Daelemans W (editors). Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP); Association for Computational Linguistics; 2014. pp. 1746–1751. doi: 10.3115/v1/D14-1181

[51]Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems. 2012; 60(6): 84–90. doi: 10.1145/3065386

[52]Peng C, Cheng S, Sun M, et al. Prediction of sound transmission loss of vehicle floor system based on 1d-convolutional neural networks. Sound & Vibration. 2024; 58(1): 046940. doi: 10.32604/sv.2024.046940

[53]Dursun ÖO. Deep Learning-Based Prediction of Commercial Aircraft Noise: A CNN–Transformer Hybrid Model Versus Support Vector Regression and Multi-Layer Perceptron. Aerospace. 2025; 12(11): 1031. doi: 10.3390/aerospace12111031

[54]Wu Y, Wang Y, Cai H, et al. An incremental intelligent fault diagnosis method for marine diesel engines based on CNN-Transformer and cosine similarity. Sound & Vibration. 2025; 59(6): 3617. doi: 10.59400/sv3617

[55]Jiao D, Zhou H, Huang T, et al. Numerical Study on Aerodynamic Noise Reduction in Passenger Car with Fender Shape Optimization. Symmetry. 2024; 16(6): 651. doi: 10.3390/sym16060651

[56]Xu L, Zhu X, Wang Q, et al. The New China Automotive Engineering Research Institute Co., Ltd.: Full-Scale Aero-Acoustic Wind Tunnel. In: Proceedings of the WCX SAE World Congress Experience; 8–10 April 2025; Detroit, MI, USA. doi: 10.4271/2025-01-8779

[57]Duell EG, Walter J, Yen J, et al. Progress in aeroacoustic and climatic wind tunnels for automotive wind noise and acoustic testing. SAE International Journal of Passenger Cars-Mechanical Systems. 2013; 6(1): 448–461. doi: 10.4271/2013-01-1352

[58]ISO 532-1:2017. Acoustics—Methods for Calculating Loudness—Part 1: Zwicker Method. 2017.

[59]Zaareer M, Mourad AH. Effect of vehicle side mirror base position on aerodynamic forces and acoustics. Alexandria Engineering Journal. 2022; 61(2): 1437–1448. doi: 10.1016/j.aej.2021.06.049

[60]Xiong Z, Zhang X, Liu P, et al. Optimization of Window Glass Parameters on Vehicle Weight and Wind Noise. SAE International Journal of Passenger Vehicle Systems. 2025; 18(3): 281–298. doi: 10.4271/15-18-03-0018

[61]Toer GA, Kim G. An empirical study on the impact of feature scaling and encoding strategies on machine learning regression pipelines. International Journal of Informatics and Information Systems. 2026; 9(1): 242–256. doi: 10.47738/ijiis.v9i1.293

[62]Shen X, Shen J, Zhang Y, et al. A multi-source feature engineering-enhanced framework for mid-to-long-term EV charging load forecasting: Integrating self-adaptive optimization and BiLSTM-iTransformer predictor. Energy. 2025; 139020. doi: 10.1016/j.energy.2025.139020