Sowing Scheduling of Two Crop Types via Predictive Control in Adaptive Vertical Farms for Space

The adaptive vertical farm is proposed to produce a variety of fresh food for astronauts during missions or future space installations. It is an innovative vertical greenhouse that relies on the idea of gradually adjusting the distance between shelves by moving them automatically according to the growth of the plants. An optimal scheduling algorithm of sowings is presented for the cultivation of two types of crops using mixed-integer linear programming to fully exploit the available vertical space and maximize the production yield. Building algorithms that are robust against possible deviations from theoretical growth curves is crucial in applications for space, where plant growing is still at an extremely experimental stage. The proposed approach, based on the formalism of predictive control, is able to account for possible disturbances acting on the greenhouse that may cause deviations of the crop growth from nominal curves, owing to repeated optimizations over sliding windows. Numerical results are reported in a case study involving the simultaneous cultivation of wheat and lettuce to showcase the effectiveness of the approach.