Evaluation of smart fish feeding system using artificial intelligence and internet of things under desert regions. | ||||
| Mediterranean Aquaculture Journal | ||||
| Volume 12, Issue 1, June 2025, Page 40-51 PDF (528.95 K) | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/maj.2025.433749 | ||||
 View on SCiNiTO | ||||
| Authors | ||||
| Hesham M. Dorgham1; Hussein Mohamed K.2; Mohamed F. Badran3; Ashraf Y. El-Dakar1 | ||||
| 1Aquaculture and Biotechnology Department, Faculty of Aquaculture and Marine Fisheries, Arish University, Egypt | ||||
| 2Computer Science Department, Faculty of Computing and Information, Suez Canal University, Egypt. | ||||
| 3Aquatic Hatchery Production Department, Fish Farming and Technology Institute, Suez Canal University, Egypt. | ||||
| Abstract | ||||
| This study developed an Artificial Intelligence-driven Internet of Things-based feeding system for fish production, integrating real-time water quality monitoring with machine learning optimization to enhance feeding utilization. The experiment was conducted in desert conditions, with 80 Nile tilapia, Oreochromis niloticus fish/m³ stocked inside one cubic meter round fiberglass tanks. The system included Arduino-based sensors for temperature, pH, DO, TDS, Salinity and turbidity measurements, coupled with an XGBoost algorithm that adjusted feeding rates based on thermal growth coefficients (TGC = 0.12) and environmental factors. For a 125-day culture period, a comparison between the manual feeding technique MFT (until satiation) with the smart feeding technique SFT. SFT significantly improved performance (P≤0.05), with lower feed conversion ratio (1.24 ± 0.03 vs 1.76 ± 0.01), higher final weight (200.33 ± 3.24g vs 156.7 ± 0.75g), and increased protein efficiency ratio (2.7 ± 0.01 vs 1.9 ± 0.02) compared to MFT. Water quality parameters showed significant (P≤0.05) improvements, with Ammonia, NH3 (0.022 ± 0.01 vs 0.056 ± 0.01 mg/L) and nitrite (0.039 ± 0.01 vs 0.132 ± 0.01 mg/L) concentrations were significantly lower (P≤0.05) in treatment tanks. The system's edge computing architecture enabled low-latency adjustments without cloud dependency, while introducing a web-based system monitoring solution. The collected data over the culture period was stored in the cloud, and an integrated secure digital card module was used for analysis and validation of the system. These results validate the potential of AI-IoT integration in addressing key challenges of feed waste, which can cost up to 70% of total costs and water pollution in intensive aquaculture. The study demonstrates a scalable model for precision aquaculture that balances productivity with environmental sustainability. Future research should focus on introducing various water quality sensors, the culture of other fish species, and the introduction of behavioral analysis using underwater cameras. It may be concluded that SFT was more efficient for improving growth rate, FCR and nutrient utilization. It will be more useful in desert aquaculture Egyptian new agricultural farms. | ||||
| Keywords | ||||
| Key Words: AI-driven IoT; Smart feeding system; Nile tilapia; Water quality monitoring; Feed Efficiency | ||||
|
Statistics Article View: 95 PDF Download: 131 |
||||
