Self-Supervised Velocity Field Learning for High-Resolution Traffic Monitoring with Distributed Acoustic Sensing

Distributed Acoustic Sensing (DAS) is a technology that can be employed to record vibrations along fiber optic (telecommunication) cables, including those generated by human activities. Since the optical fiber cables are often deployed along existing traffic infrastructures, DAS has the potential to record vehicular traffic flows, which permits high-resolution traffic analysis and long-term monitoring. In this work, we propose a Machine Learning (ML) model for estimating the speed of vehicles using DAS data. A major component of the proposed model is based on Continuous Piecewise Affine (CPA) transformations, which allows us to extract the speed as a function of space and time. We demonstrate the efficiency of our approach, which is significantly faster than non-ML solutions in estimating the vehicle speed.