See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/392660011 An AI-Driven IoT Framework for Aquaculture Management Conference Paper · June 2025 CITATIONS 0 READS 38 5 authors, including: Andrew Omambia University of Eastern Africa, Baraton 4 PUBLICATIONS   19 CITATIONS    SEE PROFILE Anthony Wambua Wambua Daystar University 16 PUBLICATIONS   37 CITATIONS    SEE PROFILE Benard Maake Kisii University 36 PUBLICATIONS   169 CITATIONS    SEE PROFILE All content following this page was uploaded by Anthony Wambua Wambua on 13 June 2025. The user has requested enhancement of the downloaded file. https://www.researchgate.net/publication/392660011_An_AI-Driven_IoT_Framework_for_Aquaculture_Management?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_2&_esc=publicationCoverPdf https://www.researchgate.net/publication/392660011_An_AI-Driven_IoT_Framework_for_Aquaculture_Management?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_3&_esc=publicationCoverPdf https://www.researchgate.net/?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_1&_esc=publicationCoverPdf https://www.researchgate.net/profile/Andrew-Omambia?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_4&_esc=publicationCoverPdf https://www.researchgate.net/profile/Andrew-Omambia?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_5&_esc=publicationCoverPdf https://www.researchgate.net/institution/University_of_Eastern_Africa_Baraton?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_6&_esc=publicationCoverPdf https://www.researchgate.net/profile/Andrew-Omambia?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_7&_esc=publicationCoverPdf https://www.researchgate.net/profile/Anthony-Wambua?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_4&_esc=publicationCoverPdf https://www.researchgate.net/profile/Anthony-Wambua?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_5&_esc=publicationCoverPdf https://www.researchgate.net/institution/Daystar_University?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_6&_esc=publicationCoverPdf https://www.researchgate.net/profile/Anthony-Wambua?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_7&_esc=publicationCoverPdf https://www.researchgate.net/profile/Benard-Maake?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_4&_esc=publicationCoverPdf https://www.researchgate.net/profile/Benard-Maake?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_5&_esc=publicationCoverPdf https://www.researchgate.net/institution/Kisii_University?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_6&_esc=publicationCoverPdf https://www.researchgate.net/profile/Benard-Maake?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_7&_esc=publicationCoverPdf https://www.researchgate.net/profile/Anthony-Wambua?enrichId=rgreq-dad18232a35bd8a0698abc84d71416cf-XXX&enrichSource=Y292ZXJQYWdlOzM5MjY2MDAxMTtBUzoxMTQzMTI4MTQ5OTI3NjA2MEAxNzQ5ODM0MTAxMTk5&el=1_x_10&_esc=publicationCoverPdf IST-Africa 2025 Conference Proceedings Miriam Cunningham and Paul Cunningham (Eds) IST-Africa Institute and IIMC, 2025 ISBN: 978-1-905824-74-8 Copyright © 2025 The authors www.IST-Africa.org/Conference2025 Page 1 of 9 An AI-Driven IoT Framework for Aquaculture Management Andrew OMAMBIA1, Anthony WAMBUA2, Fredrick AWOUR3, Benard MAAKE3, Paul ORINA4 1University of East Africa- Baraton, P.O Box 2500, Eldoret, 30100, Kenya Tel: +254723140175, Email: omambiaa@ueab.ac.ke 2Daystar University, P.O Box 44400, Nairobi, 00100, Kenya Tel: +254721589325, Email: awambua@daystar.ac.ke 3Kisii University, P.O Box 408, Kisii, 40200, Kenya Tel: +254715867778, Email: fawuor@kisiiuniversity.ac.ke, bmaake@kisiiuniversity.ac.ke 4Kenya Marine and Fisheries Research Institute, PO Box 3259 - 40200 Kisii, Kenya  Tel: +254710774477, paulorina@gmail.com Abstract: Aquaculture is a rapidly growing food production sector globally that provides fish protein to millions worldwide. However, the sector faces many challenges, including water quality management. Most small-scale aquaculture farmers, unlike commercial farms, rely on experience in managing water quality challenges. The limitation of real-time monitoring tools has contributed to significant economic losses at the aquaculture fish production level. This study explores existing water quality monitoring technologies in small-scale aquaculture to develop an IoT-based framework for improved efficiency. It also provides recommendations to enhance adoption and sustainability in aquaculture management. The framework adopts IoT and machine learning with a further integration of LLMs to automate aquaculture management. Water quality parameter- specific sensors are key components of the framework and their interlinkage to systems operator devices such as mobile phones and computers. The framework can be scaled and commercialized upon successful testing and refinement, benefiting a wider range of stakeholders, including large-scale fish farmers and caged aquaculture operations. This expansion would further enhance sustainability and resilience in the aquaculture industry, ensuring broader adoption of data-driven fish farming solutions. Further, other supply chain stakeholders, such as suppliers, can monitor the feed supply in real time. The framework aims to enhance water quality in aquaculture production systems, thus sustainably upscaling farmed fish production, productivity, and quality. Implementing this plan can provide several advantages. Small-scale fish farmers can enhance their decision-making processes by automating the monitoring and management of water quality. By utilizing advanced technology, they can observe key parameters such as aquatic turbidity, dissolved oxygen, ammonia, alkalinity, and acidity, enabling them to assess the suitability of water for fish farming Keywords: IoT, AI, Aquaculture, Water quality, Fish farming. 1. Introduction and Background Globally, fish farming has a significant positive impact on food security in underdeveloped, developing and developed nations. Fish is a reliable source of protein and income in many urban and rural households. However, during the production phase, fish farmers have many difficulties, including poor quality seed and feed, the most important of which is keeping an eye on and controlling the production resources [1]. Water capacity and quality are critical but not always keenly monitored resources in aquaculture. Low productivity and fish kills have resulted in investment losses due to ineffective water quality monitoring and a lack of mailto:omambiaa@ueab.ac.ke mailto:awambua@daystar.ac.ke mailto:fawuor@kisiiuniversity.ac.ke mailto:bmaake@kisiiuniversity.ac.ke mailto:paulorina@gmail.com Copyright © 2025 The authors www.IST-Africa.org/Conference2025 Page 2 of 9 prompt intervention when quality aspects, such as temperature, ammonia buildup, and dissolved oxygen drop or rise beyond the recommended limits of the culture species [2]. Many aquaculture owners run family-based farms and have not adopted information technology advancements to improve operational efficiency [3]. When it comes to making decisions regarding control actions, most farmers rely on their experience. However, because of the complexity involved, these decisions are frequently prone to error [4]; additionally, depending on humans to take necessary action would require round-the-clock supervision – this raises the cost of fish farming. Although aquaculture is one of the food industries with the fastest growth rate worldwide, environmental issues need to be resolved [5]. Recent advances in Internet of Things (IoT) technologies have made it possible to develop real-time monitoring systems that continuously track and manage water quality in aquaculture environments [6, 7]. IoT devices equipped with specialized sensors can measure critical water parameters, offering real-time insights and enabling prompt action. Artificial Intelligence (AI), particularly through machine learning, enhances these systems by analyzing historical data to predict and mitigate potential issues before they escalate. IoT and AI together create an advanced aquaculture management ecosystem that optimizes water quality, boosts operational efficiency, minimizes resource waste, and enables remote aquaculture systems aquaculture production system management [8]. An accessible source of clean water, a channel for wastewater disposal, and sturdy water containment systems are the three most critical infrastructural components needed to establish fish farming. Various small-scale fish farmers face difficulties feeding their fish routinely. Other issues they encounter include difficulty in getting to know the condition of the water and replacing the water when there is an issue. Effectively relies heavily on human knowledge and experience, which presents a significant problem [9]. In most cases, automation of fish management has been partial, either adopting IoT with or without machine learning and rarely exploring the power of LLMs and thus not fully automating the aquaculture environment. The proposed framework in this study is framed in keeping with the adoption of the IoT to automate fish farm management and facilitate remote control and monitoring of aquaculture production systems. Fish have a defined tolerance range for several environmental factors, just like many other living things. Thus, a few requirements must be met to farm fish species. For this reason, to preserve the living fish environment, employees in fish farming aquaculture production systems should work nonstop to ensure water quality parameters are monitored for optimal production. [10]. To address this, the framework adopts IoT and machine learning, further integrating LLMs to automate aquaculture management. 2. Objectives This section outlines the core objectives of developing an AI-driven IoT framework aimed at transforming aquaculture management practices. 1. To review technologies used for water quality monitoring in small-scale aquaculture farmers. 2. To investigate existing technologies and approaches used by small-scale aquaculture farmers. 3. To develop an integrative IoT-based water quality framework used by small-scale aquaculture farmers. 4. To make recommendations for the utilization and benefits of IoT-based water quality aquaculture monitoring systems The novelty of the present study lies in its integrated approach that combines water monitoring, the use of LLMs for personalized farmer support, and remote management with actions such as automatic dispensing of feed and water. Copyright © 2025 The authors www.IST-Africa.org/Conference2025 Page 3 of 9 The rest of the paper is organized as follows: The Literature Review section presents an overview of work done in this domain. This is followed by the Technology Description section, which highlights the major components of the systems and their workings. The paper then summarizes the benefits of adopting the proposed framework within the Business Benefits section. Lastly, the Conclusion section presents an overview of the work, recommendations, and further work towards implementing the framework. 3. Literature Review Sensors are key in monitoring water temperature, pH, turbidity, dissolved oxygen, ammonium, and other environmental factors in constructing an IoT-based fishery management system. A key unit of control gets data shared by the sensors and then works on it to notify the aquaculture production system managers when the preset thresholds have been surpassed. Additionally, a mobile application in the proposed system enables fish farmers to oversee and manage their fishing activities from a distance. According to literature in the area, these IoT-based systems can aid in improving the productivity of fish farms while reducing environmental effects. This has a net effect of boosting the productivity of these farmers [2]. Water quality in aquaculture systems is important for the wellbeing and development of aquatic life, and continuous monitoring is necessary to identify and address possible environmental problems. Temperature, dissolved oxygen, pH, and ammonia are water quality metrics monitored by various IoT-based devices and sensors. Furthermore, India, China, and Malaysia have deployed IoT-based aquaculture with good outcomes. The report also emphasizes the potential advantages of IoT technology application in aquaculture, including enhanced environmental sustainability, lower operational costs, and higher production and efficiency [11]. The research on the Internet of Things (IoT)-Based Smart Aquaculture Practices emphasizes how critical it is to leverage IoT technologies to address issues like disease control, feed management, and water quality management that the aquaculture sector faces. An overview of the Internet of Things (IoT) technologies utilized in innovative aquaculture is given in the paper. These technologies include automated feeders, control systems for controlling oxygen and temperature in the water, and sensors for tracking fish behaviour and water quality. The deploying of artificial intelligence and ML algorithms to evaluate sensor data and make defensible conclusions is also covered in this study [12]. The proposed study describes the creation of an intelligent cage aquaculture management system utilizing artificial intelligence (AI) techniques. The study applies AI techniques to implement a Smart aquaculture Management System (SAMS). The suggested system uses sensors and cameras to gather information on fish behaviour, ambient factors, and water quality. The study explains how they analyzed data gathered from the sensors and cameras using AI techniques. The study did not explore the integration of LLMs for personalized support of fish farmers. In [3], the authors integrated AI and IoT technologies for cage culture management. The system used underwater biological analysis images and an AI-enabled feeding model. The framework illustrates a feasible system that can constantly acquire information about the health status of fishponds and underwater images, thus providing training data for AI. It can also analyze fish mortality, fish body length, and weight using AI and image processing. With the collected data, the system can determine how to feed, maintain water quality, and protect fish against illnesses. The system's components included sensors, cameras, microcontrollers, and cloud-based data processing and storage. The authors also talk about the outcomes of their tests and how the system was implemented in an actual aquaculture scenario. The study lacks remote control and personalized support for fish farmers and can only manage water quality. Copyright © 2025 The authors www.IST-Africa.org/Conference2025 Page 4 of 9 Authors in [8] developed a system equipped with multiple sensors like pH sensor, temperature sensor, dissolved oxygen sensor, and turbidity sensor connected with controller Arduino, which makes it strong in terms of accuracy even though it lacks important sensors like ammonia for detecting Biochemical Oxygen Demand (BOD): BOD measures the amount of oxygen microorganisms require to decompose organic matter in water, it also lacks personalized support to fish farmers. The creation of an IoT-based system for real-time water quality monitoring in aquaculture is described in the study IoT for Aquaculture 4.0: Smart and Easy-to-Deploy Real-Time Water Monitoring with IoT [13]. The system comprises inexpensive wireless-based sensors capable of detecting dissolved oxygen, temperature, pH, and other parameters. The data is then sent to a cloud-based platform for analysis and storage. They emphasize other benefits of the system, such as its affordability, simplicity of setup, and capacity for real-time monitoring. They also review how the system may monitor and regulate the environmental factors in any aquaculture environment to maintain the best possible growth and health for the fish. The IoT sensors, cloud-based platforms, wireless communication protocols, and other hardware and software components utilized in the system are all described in detail in this article. The authors also present the experiment findings, which show the system's precision and dependability in real-time water quality parameter measurements. In June 2018, Dupont, et al. [13], using a mobile app, an IoT-based fish farm control system is developed, as detailed in the research Realization of IoT-based Fish Farm Control Using Mobile App. A microprocessor is attached to several sensors that measure the fish farm's temperature, oxygen level, and water quality. The microcontroller uploads sensor data to a server over Wi-Fi and receives control signals from the mobile app. The capacity of the system to deliver precise real-time data and manage the aqua farmer's expected parameters was demonstrated during testing on an actual fish farm. The authors concluded that by enabling remote monitoring and control, their method might increase the production and efficiency of fish farming [14]. The study by Kim, et al. [14] attests that an Internet of Things-based fish farm control system using a mobile app was feasible. The system enabled real-time monitoring and controlling the fishpond’s temperature, oxygen level, and water quality. Users could remotely, flexibly, and conveniently monitor and manage the fish farm parameters through a mobile app. The authors accentuate techniques that might increase profitability, lower labour expenses, and increase fish farming's productivity and efficiency. According to the literature, IoT technology can change how fish farming is run, creating a more lucrative and sustainable sector. The system made accurate real-time monitoring and control of the fish farm's temperature, oxygen level, and water quality possible. The user could monitor as well as manage the fish farm parameters remotely with more flexibility and convenience, simply through a mobile app farming is run, creating a more lucrative and sustainable sector. Authors in [15] proposed a system aimed at managing fish health, feeding, and water quality in real-time on aquaculture farms. The system has the network layer, the application layer, the cloud layer, and the sensing layer make up the four layers of the system architecture. Numerous sensors are included in the sensing layer to track variables, including pH, water level, dissolved oxygen, and water temperature. A data transmission gateway is part of the network layer, while a mobile application and a web- based user interface comprise the application layer. The cloud layer offers processing and storage to handle the massive amounts of data the sensors produce. The system was put to the test in a pilot study on a trout farm, establishing the viability and efficacy of the suggested remedy in raising the sustainability and efficiency of aquaculture operations. The goal of the research on developing and implementing an Internet of Things-based water quality monitoring system for aquaculture in the Mekong Delta was to develop and implement such a system in the Vietnamese aquaculture region. The system collected and transmitted water Copyright © 2025 The authors www.IST-Africa.org/Conference2025 Page 5 of 9 temperature, pH, dissolved oxygen, and water level data to a cloud-based server using sensors and an Arduino-based microcontroller. Farmers could modify their farming techniques by using real-time information from the data analysis on water quality conditions. The system was considered accurate and reliable in monitoring water quality and could help improve aquaculture productivity and sustainability in the region. A water quality monitoring system based on Narrowband IoT (NB-IoT) technology was proposed in [16] for aquaculture pond water quality monitoring system. The system uses a cloud platform, an NB-IoT gateway, and several sensor nodes. The sensor nodes gathered data on the water's temperature, turbidity, pH, and dissolved oxygen levels. The NB-IoT gateway transmitted the gathered data to the cloud platform, where it was stored and analyzed. The system's ability to precisely monitor water quality metrics and give farmers real-time feedback was demonstrated during testing in a prawn pond. The study found that the NB-IoT-based water quality monitoring system is an effective and affordable solution for aquaculture ponds, which can improve production efficiency and reduce the risks associated with fish diseases. 4. Technology Description 4.1 System Overview The study proposes a system that uses sensors attached to the aquaculture production system backbone for managing the pond in small-scale fish farmers. The study proposes an end-to- end system that automates water parameter sensing using sensors and takes the necessary intervention actions. It further automates decisions on water quality by sending alerts to the appropriate end-users based on set parameters. Unlike most systems that use IoT with limited sensors and machine learning for automating decisions, the proposed system integrates the use of LLMs and a wider range of sensors into the aquaculture environment. The complete system overview is depicted in Figure 1, and the roles of all components are explained thereafter. 4.2 System Components and Working – the IoT-based Aquaculture Management System A) Sensor data acquisition and aggregation module: This module consists of sensors attached to the aquaculture production system for sensing parameters and aggregating data. The roles of the sensors proposed in the system are described below: 1. The temperature sensor monitors the water temperature. A sudden temperature change could signify the presence of chemical or industrial effluents. The temperature change could also result from weather patterns, which need to be monitored for optimal fish production. 2. The turbidity sensor monitors the cloudiness and haziness of the water in the aquaculture production system, which can signify the presence of suspended particles. Increased haziness can be because of industrial effluents or excessive dirty water which is dangerous in the pond as it can clog fish gills, lower the amount of dissolved oxygen, and subsequently lead the death of fish. 3. Water level sensors detect the depth or height of water in a pond or other aquatic environments. A pressure sensor can also be used for the same reason since the pressure of the water is relative to the amount of water in the pond. Excessive evaporation owing to weather changes can affect water levels, leading to reduced oxygen levels. Monitoring water level information is vital for maintaining the appropriate water levels required for the health and growth of aquatic organisms. 4. The dissolved oxygen sensor monitors the amount of oxygen in water. Oxygen is one of the most critical parameters in any aquaculture production system. Sudden fluctuations could signify chemical reactions. Excessive evaporation of water or water leaks can Copyright © 2025 The authors www.IST-Africa.org/Conference2025 Page 6 of 9 decrease water levels, leading to low oxygen levels. Water temperature and salinity can also affect oxygen concentrations, leading to fish death due to oxygen deficiency. This can be fatal, and monitoring this parameter is critical for aquaculture production system owners. Figure 1: Proposed System Framework Overview 5. The pH sensor monitors the amount of acidity or alkalinity in water on a scale of 1-14, with a pH value of 7 being neutral. pH values below 7 indicate that the water is acidic, while values above 7 indicate that the water is alkaline. This affects the aquaculture production system environment and the health and production of fish. 6. Ammonia sensors detect the concentration of ammonia in the water. Ammonia is produced by decomposing organic matter, uneaten feed, and fish excretion through their gills in aquaculture systems. Elevated levels of ammonia can be toxic to aquatic organisms, leading to stress, illness, or even death. Copyright © 2025 The authors www.IST-Africa.org/Conference2025 Page 7 of 9 7. Conductivity sensors detect the ability of water to conduct electrical current, which increases with higher concentrations of dissolved salts and minerals. This measurement provides valuable information about the salinity levels and TDS content of the water in aquaculture systems. These sensors will be distributed at different points of the tanks or aquaculture production systems to measure and monitor these water parameters to determine how safe water is for the health and development of fish. Redundant sensors can be used, and their values are aggregated in this module. Use of redundant sensor is used to increase the reliability of sensor readings, hence providing backup in case of sensor failure or inaccuracies. B) Controller and driver unit: The controller and driver unit are the heart of the system’s ability to take necessary and timely automated interventions to maintain an optimal aquaculture production system environment. The module has a smart water inlet and outlet, which can be automatically by the IoT gateway to release water into the pond and out of the pond, respectively, in response to one or more of the parameters as read by the data sensor acquisition module. The controller unit can also be triggered remotely. Additionally, this module has a smart feed dispenser, meaning farmers can add water into the pond, release water from the pond, and even feed fish remotely. C) IoT Gateway: This is an essential and critical part of the system that can be implemented by use of a Raspberry Pi for cost effectiveness. The gateway links the sensor module to the controller unit, allowing the automatic triggering of water inlet, outlet, and feeder based on parameters reported by the sensor module. The IoT gateway also links the cloud system to the aquaculture production system allowing for remote actions such as releasing of feeds while making it possible to send data from the sensors to the cloud. D) Cloud system: The system receives data from the many farmers who use this system. It acquires and stores all the data in a database. The system facilitates remote access of the data to farmers. E) Application system: This is the main system for farmers and other end users with supporting functions such as data visualization over long or short-term periods, monitoring the data reported by the sensor attached to the aquaculture production systems, making predictions through machine learning as the system accumulates more data, setting of sensor parameter thresholds, triggering the driver unit for example, to add water to the aquaculture production system, sending alerts to farmers when set thresholds are of parameters are reached and access to a farmers’ discussion forum that aids sharing of skills and knowledge. Lastly, the system allows farmers to benefit from expert advice from NGOs or government aquaculture experts. F) LLMs module: This component is attached to the cloud system to allow for personalized recommendations from the data gathered by the system. While farmers can learn from other farmers and experts, LLMs can make more personalized and on-demand recommendations, allowing farmers to make timely decisions. H) User access: Admin users can access the cloud system for any necessary administrative actions, while farmers can access the application system through their laptops or mobile devices. Allowing access through mobile devices makes the system robust owing to the high penetration levels of mobile phones and the fact that farmers have access to their phones throughout. 5. Business Benefits This study suggests a framework for real-time monitoring of water quality indicators for fish health and development in small-scale farming using the Internet of Things and machine learning technology. Implementing this plan can provide several advantages. Small-scale fish farmers can enhance their decision-making processes by automating the monitoring and management of water quality. By utilizing advanced technology, they can observe key Copyright © 2025 The authors www.IST-Africa.org/Conference2025 Page 8 of 9 parameters such as aquatic turbidity, dissolved oxygen, ammonia, alkalinity, and acidity, enabling them to assess the suitability of water for fish farming. The integration of IoT sensors and machine learning algorithms allows farmers to effectively monitor and control essential water quality factors, including temperature, pH levels, dissolved oxygen, ammonia, and nitrate concentrations. This technological approach optimizes water quality management, leading to improved fish yields, faster growth, and higher survival rates. Early detection of issues related to pH imbalances or ammonia levels helps prevent disease outbreaks and reduces fish stock mortality. Research and prototyping projects in IoT and machine learning make cutting-edge technology accessible to small-scale fish farmers, leveling the playing field with larger enterprises. Implementing a smart water quality management system using IoT and machine learning reduces aquaculture costs by automating monitoring, minimizing manual testing, and optimizing resource usage. This lowers operational and labour expenses while improving efficiency. Real-time monitoring and predictive analytics help prevent fish mortality by enabling proactive interventions and reducing financial losses. 6. Conclusion This paper explores the possibilities of an all-around aquaculture management system based on IoT and AI. The suggested framework has immense potential for small-scale fish growers in underdeveloped and developing countries since it leverages IoT and machine learning to monitor and manage aquaculture production systems. The paper explores how the integration of IoT and AI in aquaculture management holds transformative potential for small-scale fish farmers, especially in regions developing and underdeveloped regions that are characterized by food insecurity. By enabling real-time monitoring and data-driven insights on critical water quality parameters and initiating immediate interventions, this technology empowers farmers to enhance fish health, optimize production, and mitigate risks related to disease outbreaks. The proactive approach facilitated by IoT sensors and machine learning not only boosts productivity and profitability for these farmers but also contributes to the overall resilience and sustainability of local aquaculture practices. It is expected some hands-on training will be conducted to enable farmers to use and maintain the system. To further this work, the authors will develop a prototype that implements the proposed system. The prototype will be deployed in several aquaculture production systems, and data will be collected for a period. Data will be analyzed and too if the system triggers the corrective actions at the expected time. After the prototype testing and necessary modifications made to the system, the authors will commercialize the system by developing actual systems that can be deployed to many farmers. Upon successful testing and refinement, the framework can be scaled and commercialized, benefiting a wider range of stakeholders, including large-scale fish farmers and caged aquaculture operations. This expansion would further enhance sustainability and resilience in the aquaculture industry, ensuring broader adoption of data-driven fish farming solutions. Further, other supply chain stakeholders, like suppliers, can monitor the supply of feed in real time. The framework can be deployed in phases, starting with prototype development and field testing, which will validate sensor accuracy and algorithm performance. Pilot programs with selected small-scale fish farms will refine the system and address implementation gaps. With successful trials and funding, commercialization can begin at a larger scale, supported by partnerships with aquaculture cooperatives, government agencies, and technology providers. Copyright © 2025 The authors www.IST-Africa.org/Conference2025 Page 9 of 9 Acknowledgements We confirm that there was the use of AI to refine sentence structure and improve readability while maintaining the paper's original meaning in the paper entitled “An AI-Driven IoT Framework for Aquaculture Management” References [1] F. E. Idachaba, J. O. Olowoleni, A. E. Ibhaze, and O. O. Oni, "IoT enabled real-time fishpond management system," in Proceedings of the world congress on engineering and computer science, 2017, vol. 1, pp. 25-27. [2] J. Janet, S. Balakrishnan, and S. S. Rani, "IOT based fishery management system," International Journal of Oceans and Oceanography, vol. 13, no. 1, pp. 147-152, 2019. [3] C.-C. Chang et al., "Applying artificial intelligence (AI) techniques to implement a practical smart cage aquaculture management system," Journal of Medical and Biological Engineering, vol. 41, pp. 652-658, 2021. [4] G. Gao, K. Xiao, and M. Chen, "An intelligent IoT-based control and traceability system to forecast and maintain water quality in freshwater fish farms," Computers and Electronics in Agriculture, vol. 166, p. 105013, 2019. [5] Food and N. Agriculture Organization of the United, The State of Food and Agriculture 2020: Overcoming Water Challenges in Agriculture. Rome, Italy: FAO, 2020. [6] N. A. M. Jais, A. F. Abdullah, M. S. M. Kassim, M. M. Abd Karim, M. Abdulsalam, and N. A. Muhadi, "Improved accuracy in IoT-Based water quality monitoring for aquaculture tanks using low-cost sensors: Asian seabass fish farming," Heliyon, vol. 10, no. 8, 2024. [7] A. Omambia, B. Maake, and A. Wambua, "Water quality monitoring using IoT & machine learning," in 2022 IST-Africa Conference (IST-Africa), 2022: IEEE, pp. 1-8. [8] M.-C. Chiu, W.-M. Yan, S. A. Bhat, and N.-F. Huang, "Development of smart aquaculture farm management system using IoT and AI-based surrogate models," journal of Agriculture and Food Research, vol. 9, p. 100357, 2022. [9] A. Junaidi and C. Kartiko, "Design of pond water quality monitoring system based on internet of things and pond fish market in real-time to support the industrial revolution 4.0," in IOP Conference Series: Materials Science and Engineering, 2020, vol. 771, no. 1: IOP Publishing, p. 012034. [10] T. Hartel et al., "The effect of fish and aquatic habitat complexity on amphibians," Hydrobiologia, vol. 583, pp. 173-182, 2007. [11] D. R. Prapti, A. R. Mohamed Shariff, H. Che Man, N. M. Ramli, T. Perumal, and M. Shariff, "Internet of Things (IoT)‐based aquaculture: An overview of IoT application on water quality monitoring," Reviews in Aquaculture, vol. 14, no. 2, pp. 979-992, 2022. [12] A. Yadav, M. T. Noori, A. Biswas, and B. Min, "A concise review on the recent developments in the internet of things (IoT)-based smart aquaculture practices," Reviews in fisheries science & aquaculture, vol. 31, no. 1, pp. 103-118, 2023. [13] C. Dupont, P. Cousin, and S. Dupont, "IoT for aquaculture 4.0 smart and easy-to-deploy real-time water monitoring with IoT," in 2018 global internet of things summit (GIoTS), 2018: IEEE, pp. 1-5. [14] Y. Kim, N. Lee, B. Kim, and K. Shin, "Realization of IoT based fish farm control using mobile app," in 2018 International Symposium on Computer, Consumer and Control (IS3C), 2018: IEEE, pp. 189-192. [15] L. V. Q. Danh, D. V. M. Dung, T. H. Danh, and N. C. Ngon, "Design and deployment of an IoT-based water quality monitoring system for aquaculture in Mekong Delta," International Journal of Mechanical Engineering and Robotics Research, vol. 9, no. 8, pp. 1170-1175, 2020. [16] J. Huan, H. Li, F. Wu, and W. Cao, "Design of water quality monitoring system for aquaculture ponds based on NB-IoT," Aquacultural Engineering, vol. 90, p. 102088, 2020. View publication stats https://www.researchgate.net/publication/392660011 An AI-Driven IoT Framework for Aquaculture Management 1. Introduction and Background 2. Objectives 3. Literature Review 4. Technology Description 4.1 System Overview 4.2 System Components and Working – the IoT-based Aquaculture Management System 5. Business Benefits 6. Conclusion Acknowledgements References