Feedback control of soil moisture in precision-agriculture systems: Incorporating stochastic weather forecasts

Management of soil moisture in precision-agriculture systems is complicated by the high-dimensional spatio-temporal dynamics of soil water content, the limitations in (and sparsity of) sensing/actuation capabilities, and impact of uncertain precipitation and insolation (solar radiance) futures. In this article, we envision a feedback-control framework for managing soil moisture in precision-agriculture systems via irrigation, that can take advantage of stochastic precipitation and insolation forecasts. Our proposed framework comprises three components: 1) a network modeling tool for tracking soil-moisture dynamics and placing sensors; 2) algorithms for generating representative futures of relevant weather parameters (precipitation, insolation); and 3) a stochastic-receding-horizon approach for scheduling irrigation given the representative weather futures. Our focus in this exploratory article is to provide a conceptual introduction to this framework, within the context of irrigation-system controls and the broader network-contingency-management literature.