Biofloc-Assisted Intelligent Aquaculture System Using Wireless Sensor Networks and Predictive Analytics

Abstract

The current rapid development in aquaculture has necessitated the need to develop smart and sustainable fish farming systems that are able to achieve high productivity, efficient use of resources, and enhanced management of aquatic health. Biofloc technology has been developed as a sustainable method of aquaculture with environmental benefits as it improves the quality of water and recycling of nutrients by the growth of desirable microbial communities. Nevertheless, it is important to note that the overall environmental conditions within biofloc-based aquaculture systems are a significant challenge because of the persistent variation in key parameters of water quality. In order to overcome these constraints, the proposal in this research is to develop a Biofloc -Assisted Intelligent Aquaculture System using Wireless Sensor Networks (WSN) and predictive analytics to real-time monitor the environment and make intelligent decisions. The suggested model uses distributed sensor nodes that constantly measure significant aquaculture parameters such as the pH, temperature, dissolved oxygen (DO), the concentration of ammonia, and turbidity. The sensor data are collected and sent via a power-saving WSN framework and analyzed by a machine learning algorithm, the random forest, to predictively analyze and determine intelligent water quality. The reason behind using the Random Forest model is that it is robust, has high levels of prediction accuracy and can analyze nonlinear environmental data produced within aquaculture systems. Through an experimental assessment, the suggested framework proves suitable to predict the state of water quality and promote proactive management of aquaculture with a higher degree of reliability and operational effectiveness. Accuracy, Precision, Recall, F1-score and Root Mean Square Error (RMSE) are used to analyze the performance of the predictive model. Findings show that the suggested intelligent aquaculture system has high predictability, minimized monitoring, increased survival rate of fish, and higher biofloc stability. Biofloc technology, WSN-driven monitoring, and machine learning-driven predictive analytics promises to be a reasonably scalable and sustainable solution to the next-generation smart aquaculture system and intelligent fisheries management system.