We develop experimental systems for observing and operating recirculating aquaculture. Our research asks how sensing, deterministic automation, time-series analysis, and operator workflows can improve visibility without obscuring the biological and safety-critical limits of the farm.
Trident
Trident is the experimental operating layer Aqualabs uses to design, compare, and review controlled aquaculture trials. It keeps a structured record of water loops, feeding assumptions, sensor readings, interventions, failures, maintenance, and operator observations. That record lets us examine how a system changed, compare outcomes across trials, and distinguish useful signals from operational noise. Automation remains bounded by deterministic controls and human responsibility.
Research areas
We study how sensor placement, calibration, sampling frequency, missing data, and drift affect the reliability of water-quality and biological records. The first question is not what a model can predict, but whether the system is being observed well enough to support inference.
Measurement, interpretation, recommendation, and actuation carry different risk. We investigate architectures that keep emergency limits and equipment protection deterministic while using forecasting and anomaly detection for review, triage, and operator support.
Interventions only become useful evidence when their context and outcomes remain inspectable. We structure water conditions, feeding assumptions, maintenance, mortality, alarms, and operator notes so experiments can be compared rather than remembered informally.