AIoT-Enabled Hybrid ML–GA Framework for Elderly-Friendly Greenhouse Optimization

The rapid aging of the agricultural workforce underscores the need for technologies that ensure both productivity and usability. This study introduces an AIoT-enabled elderly-friendly greenhouse that integrates ergonomic and agronomic parameters into a unified optimization framework. Experiments with ten elderly participants (60–75 years) combined anthropometric assessments, environmental monitoring, and machine learning–based irrigation modeling. Results showed that an optimal planting table height of 75 cm maximized comfort (4.44 ± 0.34) and minimized fatigue (1.89 ± 0.66). Work–rest scheduling identified early morning (06:00–09:00) and late afternoon (15:00–18:00) as periods with reduced heat strain. Ventilation at 60% fan speed-maintained comfort ranges while stabilizing microclimate conditions. For irrigation, Random Forest Regression achieved the best accuracy ( R 2 ≈ 0.75), with soil moisture as the dominant predictor. A Genetic Algorithm (GA) further improved outcomes, increasing comfort scores by 30% and reducing water use by 20%. By embedding ergonomic ( X opt , T comfort , V comfort ) and agronomic ( W , I , θ opt ) variables as objectives, the system creates greenhouses that are both “user-aware” and “plant-aware.” This dual approach enhances productivity, sustainability, and usability, offering practical insights for AIoT-enabled smart greenhouses in aging societies.