Real-Time Acoustic Anomaly Detection in Vehicles Using AI Processor and Machine Learning

Year 2025, Volume: 9 Issue: 1st Future of Vehicles Conf., 96 - 100, 17.12.2025

Attila Aradi , Attila Károly Varga

https://doi.org/10.30939/ijastech..1769036

Abstract

This study investigates real-time acoustic anomaly detection in vehicles, focusing on diesel particulate filter (DPF) faults in a pickup truck as a case study. Utilizing two WAV audio recordings—one from normal idling and another during a DPF fault—features such as Mel-Frequency Cepstral Coefficients (MFCC) and Discrete Wavelet Transform (DWT) were extracted. Machine learning models, including Support Vector Machines (SVM) and Convolutional Neural Networks (CNN), were deployed on the Hailo-8 AI processor for edge-based, real-time inference. The approach achieved high accuracy in anomaly classification, demonstrating the viability of embedded AI for predictive vehicle maintenance. This non-invasive method enhances fault detection speed and efficiency, reducing downtime and emissions in automotive applications. The edge computing approach demonstrates significant advantages over cloud-based alternatives: sub-millisecond inference latency enables immediate fault detection critical for preventing vehicle damage; continuous operation without network dependency ensures reliability in areas with poor connectivity; local processing eliminates security risks associated with transmitting sensitive vehicle data; and bandwidth requirements are reduced as only anomaly alerts require transmission. Hailo-8's power-efficient architecture enables continuous monitoring without significant battery drain, while its 26 TOPS performance capability ensures real-time processing of complex acoustic signatures. This non-invasive method enhances fault detection speed and efficiency, reducing downtime and emissions in automotive applications while providing a scalable solution deployable across entire vehicle fleets without cloud infrastructure dependencies.

Keywords

Acoustic anomaly detection , Machine learning , edge-computing , cloud computing , AI processor

References

• [1] Erdoğan FA, Küçükmanisa A, Kilimci ZH. Detection of fault from acoustic signals in automobile engines using deep learning techniques. Kocaeli J Sci Eng. 2023;6(2):122-130. https://doi.org/10.34088/kojose.1225591
• [2] Nasim F, Masood S, Jaffar A, Ahmad U, Rashid M. Intelligent sound-based early fault detection system for vehicles. Comput Syst Sci Eng. 2023;46(3):3175-3190. https://doi.org/10.32604/csse.2023.034550
• [3] Glowacz A. Fault diagnosis of internal combustion engines using acoustic signals and advanced signal processing. Measurement. 2021;167:108478. https://doi.org/10.1016/j.measurement.2020.108478
• [4] Wu JD, Liu CH. An expert system for fault diagnosis in internal combustion engine using wavelet packet transform and neural network. Expert Syst Appl. 2009;36(3):4278-4286. https://doi.org/10.1016/j.eswa.2008.03.008
• [5] Akbalık F, Yıldız A, Ertuğrul ÖF, Zan H. Engine fault detection by sound analysis and machine learning. Appl Sci. 2024;14(15):6532. https://doi.org/10.3390/app14156532
• [6] Zhao X, Yao J, Deng W, Jia M, Liu Z. Intelligent fault diagnosis of gearbox under variable working conditions with adaptive intra-class and interclass convolutional neural network. IEEE Trans Neural Netw Learn Syst. 2022;34(10):6339-6353. https://doi.org/10.1109/TNNLS.2021.3135877
• [7] Glowacz A, Glowacz Z. Diagnosis of stator faults of the single-phase induction motor using acoustic signals. Appl Acoust. 2017;117:20-27. https://doi.org/10.1016/j.apacoust.2016.10.012
• [8] Hossain MN, Rahman MM, Ramasamy D. Artificial intelligence-driven vehicle fault diagnosis to revolutionize automotive maintenance: a review. Comput Model Eng Sci. 2024;141(2):951-996. https://doi.org/10.32604/cmes.2024.056022
• [9] Yan J, Liao JB, Gao JY, Zhang WW, Huang CM, Yu HL. Fusion of audio and vibration signals for bearing fault diagnosis based on a quadratic convolution neural network. Sensors. 2023;23(22):9155. https://doi.org/10.3390/s23229155
• [10] Fedorishin D, Birgiolas J, Mohan DD, Forte L, Schneider P, Setlur S. Fine-grained engine fault sound event detection using multimodal signals. arXiv [Preprint]. 2024:arXiv:2403.11037. https://doi.org/10.48550/arXiv.2403.11037
• [11] Heo YJ, Kim JH, Park SH, Lee KH. Deep-learning-based approach to anomaly detection techniques for large acoustic data in machine operation. Sensors. 2021;21(16):5446. https://doi.org/10.3390/s21165446
• [12] Al-Momani M, Alauthman M, Alweshah M, Atoum J. Machine learning-based anomaly detection for securing in-vehicle networks with autoencoder and LSTM. Electronics. 2024;13(10):1962. https://doi.org/10.3390/electronics13101962
• [13] Lei Y, Yang B, Jiang X, Jia F, Li N, Nandi AK. Applications of machine learning to machine fault diagnosis: a review and roadmap. Mech Syst Signal Process. 2020;138:106587. https://doi.org/10.1016/j.ymssp.2019.106587
• [14] Tran VT, AlThobiani F, Ball A. An approach to fault diagnosis of reciprocating compressor valves using Teager–Kaiser energy operator and deep belief networks. Expert Syst Appl. 2014;41(9):4113-4122. https://doi.org/10.1016/j.eswa.2013.12.026
• [15] Purohit H, Tanabe R, Ichige K, Endo T, Nikaido Y, Suefusa K, et al. MIMII dataset: sound dataset for malfunctioning industrial machine investigation and inspection. In: Proceedings of the Detection and Classification of Acoustic Scenes and Events 2019 Workshop (DCASE2019); 2019. p. 209-213. https://doi.org/10.33682/m76f-d618
• [16] El-Sharkawy AM, Abouelazm MN. Deep learning with TinyML driven real-time anomaly detection for predictive maintenance in IoT devices. Ain Shams Eng J. 2025;16(3):102552. https://doi.org/10.1016/j.asej.2024.102552
• [17] Michau G, Fink O. Unsupervised fault detection in varying operating conditions via variational autoencoders. Mech Syst Signal Process. 2021;152:107459. https://doi.org/10.1016/j.ymssp.2020.107459
• [18] Wang Y, Tian J, Chen Z, Liu X. Acoustic emission signal analysis based on deep learning for pipeline leak detection. Mech Syst Signal Process. 2022;163:108151. https://doi.org/10.1016/j.ymssp.2021.108151
• [19] Samek W, Montavon G, Lapuschkin S, Anders CJ, Müller KR. Explaining deep neural networks and beyond: a review of methods and applications. Proc IEEE. 2021;109(3):247-278. https://doi.org/10.1109/JPROC.2021.3060483
• [20] Khan S, Yairi T. A review on the application of deep learning in system health management. Mech Syst Signal Process. 2018;107:241-265. https://doi.org/10.1016/j.ymssp.2017.11.024
• [21] Abdeljaber O, Avci O, Kiranyaz S, Gabbouj M, Inman DJ. Real-time vibration-based structural damage detection using one-dimensional convolutional neural networks. J Sound Vib. 2017;388:154-170. https://doi.org/10.1016/j.jsv.2016.10.043
• [22] Lu C, Wang ZY, Qin WL, Ma J. Fault diagnosis of rotary machinery components using a stacked denoising autoencoder-based health state identification. Signal Process. 2017;130:377-388. https://doi.org/10.1016/j.sigpro.2016.07.028
• [23] Zhao R, Yan R, Chen Z, Mao K, Wang P, Gao RX. Deep learning and its applications to machine health monitoring. Mech Syst Signal Process. 2019;115:213-237. https://doi.org/10.1016/j.ymssp.2018.05.050