NeuroAqua: Developing an Optimized Artificial Intelligence and Internet of Things-Based Aquaponics System

This paper aims to solve challenges concerning cost, manual labor, and efficiency in aquaponics by designing and implementing NeuroAqua, an optimized and automated AI and IoT-based aquaponics system, in both a lab setting (small-scale) and a field setting (large-scale) in Ouroboros Farms in Half Moon Bay, CA. In the system, wireless sensors monitor key factors for system effectiveness and send data to a cloud database. NeuroAqua integrates 6 types of sensors, collecting data regarding the environment, water, and nutrients, such as air temperature, humidity, pH, light, Nitrogen, Phosphorus, Potassium, and Total Dissolved Solids. Camera module sensors were also developed to capture and monitor plant growth automatically. These 6 types of sensors were implemented in both settings and successfully collected 8 plant growth samples (2 types of plants in 4 different settings) each across 2 weeks from NeuroAqua. NeuroAqua transformed the data from the sensors into a structured database, which was fed to the machine learning (ML) models for data analysis. It also feeds data to our online monitoring system and sends out alarms when the tested values fall below or exceed the normal range. NeuroAqua is the first aquaponics system to implement a comprehensive automatic machine learning system into a real farm setting.