The increasing demand for sustainable aquaculture necessitates efficient water quality management to enhance fish health, reduce mortality rates, and improve overall productivity. However, conventional water quality monitoring relies on manual testing, which is labour-intensive, time-consuming, and ineffective in detecting rapid environmental fluctuations. To address these limitations, this study presents EcoGuard, an IoT-enabled smart aquaculture monitoring system that integrates edge computing and federated learning-based predictive analytics for real-time water quality assessment and management. EcoGuard continuously monitors the important water parameters, including pH, dissolved oxygen (DO), temperature, turbidity, and ammonia levels, through a wireless sensor network. The predictive analytics module, employing Random Forest and Long Short-Term Memory (LSTM) models, forecasts water quality trends, enabling early intervention and risk mitigation. A key feature of EcoGuard is its federated learning framework, which facilitates collaborative model training across multiple aquaculture farms while ensuring data privacy and security. The system utilizes the MQTT protocol for low-latency data transmission, while an integrated mobile application provides real-time alerts and decision support for optimized resource management. Experimental validation demonstrates that EcoGuard effectively reduces fish mortality, enhances operational efficiency, and supports sustainable aquaculture practices. By utilizing IoT, AI, and federated learning, the proposed system offers a scalable, cost-effective, and intelligent solution for modernizing aquaculture, contributing to food security, environmental conservation, and resilient fisheries management.

EcoGuard IoT ML Automation Aquaculture Monitoring Fish Farming Fish Recommendation Water Management Water Cleaning