AI in Vertical Farming: Light and Nutrient Optimization

Abstract

Vertical farming is also quite problematic in regards to energy intensity and accurate nutrient control, which makes it less scalable and sustainable. This paper describes an AI-based optimization system that changes light soils (PAR, far-red, USA) and hydroponic nutrient dosing per-plant needs as they change in real-time. Our network showed a 22 percent reduction in energy consumption and an 18 percent growth in crop yield over that of the static protocols based on a 3-month experiment of lettuce growing. The most significant innovations involve a cheap sensor platform to phenotype plants and a digitally simulated environment to train safe policies. The flexibility of the system to a variety of crops and the possibility of expanding it to large-scale crop production make it a revolutionary instrument of sustainable urban agriculture.