Skip to main content Skip to secondary navigation

Large-Scale and Continuous Subsurface Monitoring Using Distributed Acoustic Sensing in Urban Environments SEP-191 (2023)



Table of contents

  • Chapter 1: Introduction and Background
  • Chapter 2: Vehicle-induced Signal Characterization with DAS
  • Chapter 3: A Scalable and Accurate Car-based Fiber Mapping Approach with DAS
  • Chapter 4: Spatial Deep Deconvolution U-Net for Traffic Analyses with Distributed Acoustic Sensing
  • Chapter 5: Using Vehicle-induced DAS Signals for High Spatiotemporal Resolution Near-surface Characterization
  • Chapter 6: Conclusions and Future Work


This thesis investigates using Distributed Acoustic Sensing (DAS) as a scalable and cost-effective approach for continuous, large-scale monitoring in urban settings. DAS capitalizes on existing telecommunication infrastructure to establish high-resolution seismic recording arrays. The research specifically targets two critical domains: traffic management and near-surface monitoring. 

The central innovation of this work lies in leveraging moving vehicles as cost-efficient, non-specialized indicators for both traffic and seismic monitoring. This involves tackling various obstacles such as pinpointing the geographical positions of DAS channels, improving the clarity of traffic signals, and overcoming the processing constraints and diminished resolution typical of conventional ambient noise interferometry methods.

I first conduct comprehensive field experiments and simulations to characterize vehicle-induced seismic signals, thus providing foundational knowledge for more sophisticated monitoring techniques. Subsequent chapters introduce scalable and precise mapping methods that utilize concurrent onboard GPS and DAS recordings. These methods have proven essential for tasks like fiber interruption localization, enhancing vehicle tracking accuracy, and near-surface characterization. To tackle the issue of degraded signal resolution in urban DAS applications for traffic monitoring, an advanced machine learning algorithm is developed, yielding improved vehicle detection and tracking. Finally, to overcome the limitations of conventional ambient noise interferometry, I present a targeted interferometry approach based on the Kalman filter algorithm. This method enables high-resolution, cost-effective, and time-sensitive characterizations of the near-surface environment, effectively capturing phenomena such as rainfall-induced soil saturation changes.

In summary, this research signifies a substantial advancement in the application of DAS technology for urban monitoring, offering solutions that are both state-of-the-art and economically sustainable.

Reproducibility and source codes

This thesis has been tested for reproducibility. The source code and reproducibility steps are available at the GitHub repository: 


Defense Presentation

Siyuan Yuan
Publication Date
October, 2023