Li Tan, Hongfei Hou and Qin Zhang. Proceedings of IEEE 17th international conference in Information Reuse and Integration (IRI'16). Bolong and Li Tan.
Precision agriculture is a data-driven farming practice that uses intra- and inter-field information to optimize farming operations. The “brain” of precision agriculture is a decision support system (DSS) that acquires data from various sources, analyzes them, and recommends actions to farmers. Recently cloud computing has been used to improve the scalability and reliability of a DSS. Cloud-based DSSs present some major challenges for software design:(1) how can a cloud-based DSS process a diversified profile of intra- and/or inter-field data from various sources? (2) how can a cloud-based DSS accommodate and support the diversity of farming operations? (3) how can a cloud-based DSS automate the entire decision process and control field devices directly? we proposed an extensible cloud-based software platform that integrated 3 novel components to address these questions: (1) a meta data-model-based data acquisition and integration module that accepts data in different formats and semantics; (2) an adaptive soft- ware architecture supporting on-the-fly reconfiguration of decision modules; and (3) software-defined control, a new software design paradigm we proposed for taming control diversity. It enables a DSS to control various field de- vices through unified software-defined interfaces. We implemented the platform in Agrilaxy, a cloud-based DSS, and deployed it on Amazon Web Services (AWS). An early version of Agrilaxy has been used in a USDA-sponsored project on canopy management for specialty crops.