Hydrophones sensors are used to listen and record the sounds under water (just like a microphone above water!). Individual sensors and arrays of sensors can produce large amounts of data that need to be processed into various formats for analysis. One of my largest projects has been revamping the system that acquires and processes that data.

When I came into the acoustic trials group, the in-house software they had been using for the acquisition system hadn’t been upgraded in many years, and data analyst had many requests for improvements. Previous efforts to revamp the software had stalled because of the complexity and critical nature of the system. Most of the original developers and maintainers had retired, leaving the system as a black box for many.

I started by gathering the source code of the system that I could find in order to develop a baseline. I worked with the current operators, data analysts, and developers (who had moved on!) in order to understand the current logic, shortcomings, and desired outcomes. I sought out the subject matter experts in acoustic analysis to develop and implement best digital signal processing (DSP) techniques for our use case. In some cases, this led me to uncover small bugs deep in the existing system.

I planned a multiphase approach the replace the existing software to enable a smooth rollout and ensure the accuracy of the data products. Each phase provided incremental and significant improvements to reliability, performance, and maintainability.

Through persistence, multiple iterations, and rigorous unit / full system testing, I was able to upgrade and simplify the system.