Advanced Control System for Real-Time Regulation of Dissolved Oxygen in Aquaculture Systems

Abstract

Dissolved oxygen (DO) is a critical characteristic in aquaculture systems, regulating the health, development, and production of aquatic species. Maintaining adequate DO levels is essential for avoiding hypoxia, which can cause stress, sickness, and even death in fish and other animals. Traditional DO regulation relies on manual interventions and fixed aeration strategies, which cannot quickly adapt to environmental changes, causing inefficiencies and potential aquaculture productivity risks. Research presents a control system for real-time regulation of dissolved oxygen in aquaculture systems. The proposed system uses Intelligent Satin Bowerbird tuned Dynamic Logistic regression (ISB-dynamicLR) to effectively forecast DO levels while addressing excessive noise and poor data quality. The sensor data are collected continuously, providing a basis for real-time monitoring of DO levels. The data was preprocessed and decomposed into multiple frequency components using Discrete Wavelet Transforms (DWT). The Control system adjusts aeration rates and water circulation in response to predicted DO levels, providing a dynamic and adaptive solution for DO regulation. The proposed system combines dynamicLR for regression-based estimation and ISB to optimize dynamicLR parameters and kernel functions, providing robust and efficient prediction. The results demonstrate that the proposed model achieved excellent accuracy, with various error parameters such as RMSE (0.0091), MSE (0.0005) and operating time (1.92s). The system also demonstrated superior computational efficiency and outperformed traditional models. The high throughput, accuracy, and real-time capability of this system make it an ideal choice for automated DO regulation in water quality monitoring systems for aquaculture.