Acoustic Sensing for City Flow: Quasi-Supervised Recognition of Sirens and Traffic for Urban Mobility Intelligence

Notice

This is an unedited manuscript accepted for publication and provided as an Article in Press for early access at the author’s request. The article will undergo copyediting, typesetting, and galley proof review before final publication. Please be aware that errors may be identified during production that could affect the content. All legal disclaimers of the journal apply.

Year : 2026 | Volume : 16 | 01 | Page :
    By

    Bhargav Chebrolu,

  1. Research Scholar, Department of Project Management, The University of Texas at Dallas, Richardson, Texas, United States

Abstract

This paper frames environmental audio as a mobility telemetry source, extending a benchmark urban-sound corpus with transportation-critical classes—ambulance, firetruck, police, and traffic—and training spectrogram-based models under a quasi-supervised regime to support real-time city operations; leveraging 10-fold protocols, class-weighted objectives, and audiospecific augmentations (time stretch, pitch shift, SpecAugment, PatchAugment), the system benchmarks multiple CNN backbones combined with self-supervised learning paradigms enable the extraction of rich, discriminative acoustic representations, achieving strong multi-class classification accuracy alongside robust ROC performance and reliable Grad-CAM–based interpretability. These properties support transparent model validation and trustworthy decision-making. As a result, the approach enables practical use cases such as real-time signal preemption, automated incident detection, traffic state inference, and congestion analytics across complex urban environments. Moreover, the pipeline’s lightweight deployment options, low computational overhead, and fully reproducible evaluation framework position acoustic classifiers as scalable, cost- effective edge sensors that enhance urban transportation management systems without requiring additional roadside hardware or intrusive infrastructure upgrades.

Keywords: Acoustic sensing, urban mobility, audio classification, spectrogram analysis, self-supervised learning, smart cities

How to cite this article:
Bhargav Chebrolu. Acoustic Sensing for City Flow: Quasi-Supervised Recognition of Sirens and Traffic for Urban Mobility Intelligence. Trends in Electrical Engineering. 2026; 16(01):-.
How to cite this URL:
Bhargav Chebrolu. Acoustic Sensing for City Flow: Quasi-Supervised Recognition of Sirens and Traffic for Urban Mobility Intelligence. Trends in Electrical Engineering. 2026; 16(01):-. Available from: https://journals.stmjournals.com/tee/article=2026/view=236074


References

  1. Venkatesh S, Moffat D, Miranda ER. You only hear once: a YOLO-like algorithm for audio segmentation and sound event detection. Applied Sciences. 2022 Mar 24;12(7):3293.
  2. Abdoli S, Cardinal P, Koerich AL. End-to-end environmental sound classification using a 1D convolutional neural network. Expert Systems with Applications. 2019 Dec 1;136:252-63.
  3. Chen X, Wang M, Kan R, Qiu H. Improved Patch-Mix Transformer and Contrastive Learning Method for Sound Classification in Noisy Environments. Applied Sciences. 2024 Oct 24;14(21):9711.
  4. Tripathi AM, Mishra A. Self-supervised learning for environmental sound classification. Applied Acoustics. 2021 Nov 1;182:108183.
  5. Chen F, Zhu Z, Sun C, Xia L. Evaluating metric and contrastive learning in pretrained models for environmental sound classification. Applied Acoustics. 2025 Mar 15;232:110593.
  6. ANasiri A, Cui Y, Liu Z, Jin J, Zhao Y, Hu J. Audiomask: Robust sound event detection using mask r-cnn and frame-level classifier. In2019 IEEE 31st International Conference on Tools with Artificial Intelligence (ICTAI) 2019 Nov 4 (pp. 485-492). IEEE.
  7. Yeom J, Li G, Loianno G. Geometric fault-tolerant control of quadrotors in case of rotor failures: An attitude based comparative study. In2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2023 Oct 1 (pp. 4974-4980). IEEE.
  8. Wilkinghoff K. Self-supervised learning for anomalous sound detection. InICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2024 Apr 14 (pp. 276-280). IEEE.
  9. Moummad I, Farrugia N, Serizel R. Self-supervised learning for few-shot bird sound classification. In2024 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW) 2024 Apr 14 (pp. 600-604). IEEE..
  10. Zhao J, Liu X, Zhao J, Yuan Y, Kong Q, Plumbley MD, Wang W. Universal sound separation with self-supervised audio masked autoencoder. In2024 32nd European Signal Processing Conference (EUSIPCO) 2024 Aug 26 (pp. 1-5). IEEE..
  11. Vu L, Tran T, Lim WH, Phan R. Toward end-to-end interpretable convolutional neural networks for waveform signals. arXiv preprint arXiv:2405.01815. 2024 May 3.
  12. Ntalampiras S, Potamitis I. Acoustic detection of unknown bird species and individuals. CAAI Transactions on Intelligence Technology. 2021 Sep;6(3):291- 300.
  13. Kadandale VS, Montesinos JF, Haro G, Gómez E. Multi-channel u-net for music source separation. In2020 IEEE 22nd international workshop on multimedia signal processing (MMSP) 2020 Sep 21 (pp. 1-6). IEEE.
Ahead of Print Subscription Review Article
Volume 16
01
Received 13/01/2026
Accepted 15/01/2026
Published 17/01/2026
Publication Time 4 Days
79

Login


My IP

PlumX Metrics