Modeling of surface vessels using distributed acoustic sensing data and physics-based optimization

Technological advances in global communications depend significantly on robust and efficient long-distance infrastructures. One notable example is the submarine cable network. Installed on the ocean floor, these cables use fiber optic technology to transmit large volumes of data at high speed and low latency between continents. Beyond their primary communication function, recent innovations allow these cables to serve as Distributed Acoustic Sensing (DAS) systems, effectively converting tens of kilometers of passive fiber into massive, coherent arrays of vibration sensors. The primary objective of this project is to enhance maritime surveillance capabilities by integrating DAS technology with advanced kinematic modeling. This paper establishes a rigorous physical and mathematical framework for interpreting the acoustic signatures of surface vessels detected by bottom-mounted fibers. We derive the complete opto-acoustic transfer function, formulate the hyperbolic moveout equations based on a moving point-source solution to the wave equation, and implement a stochastic inversion scheme using Differential Evolution. By optimizing a correlation-based loss function, we demonstrate the ability to recover vessel trajectory, speed, and depth from complex interferometric patterns with speed estimation errors consistently below 1%. This approach allows for the extraction of quantitative physical parameters from raw optical data, bridging the gap between photonics and hydroacoustics.