Sustainable Hydroponic Strawberry Growth by Minimizing Waste Through Nutrient Solution Approximation Using an Artificial Neural Network

Sustainable hydroponic strawberry cultivation requires precise, real-time management of nutrient solutions to reduce waste and improve resource efficiency for home growers. This work presents an artificial neural network (ANN)-based system that estimates and approximates optimal nutrient formulations for strawberry hydroponics, with a primary focus on nutrient and water savings. The ANN-based system is implemented on hardware for real-time operation and was compared against a traditional PID controller to evaluate performance in nutrient solution management. The ANN receives temperature, humidity, water pH, nutrient concentrations, and lighting as inputs and outputs recommended nutrient solution parameters. Training used gradient descent and reached convergence after an average of 2500 epochs, achieving a mean error of 0.001 and an operating frequency of 84 MHz. Mechanical design was optimized for compact 3D-printable assembly to facilitate adoption by small-scale and domestic producers. The validation with 15 min sampling demonstrated the ability of the system to maintain target nutrient conditions, producing measurable increases in seedling and fruit weight and size while substantially reducing water and nutrient waste. These results indicate that ANN-driven, real-time nutrient estimation can make precision hydroponics more accessible and sustainable for home growers. In the comparative analysis, the ANN-based system demonstrated superior performance compared to the PID controller in maintaining precise nutrient concentrations and reducing resource waste, achieving more accurate tracking of target nutrient levels.