Closed-Loop Sustainable Precision Animal Agriculture

Overview: Current animal farming practice relies on discrete individual sensors to monitor animal health and farm efficiency, however, no cyber-physical systems (CPS) or closed-loop methodologies exist to monitor and control the efficiency of individual animals within a herd or the entire herd in real-time. We believe this limits the production efficiency and opens the possibility of accidental health hazards. Precision animal agriculture provides closed-loop intensified management of the individual animal and herd in areas of nutrition, health, productivity, and efficiency. Using trans-disciplinary expertise, this 3-year medium project will create a generic CPS for precision animal agriculture. Individual dairy cows in conjunction with a dairy herd within a farm will be utilized as a testbed for CPS development. Design tools, networking, sensing and control artifacts, and control software will be systematically developed to create a CPS for sustainable precision animal agriculture. The proposed integrated system will improve the sustainability of the US dairy industry by addressing key inefficiencies in animal metabolism and health, which are the primary drivers of overall farm efficiency. This new system will also serve as an exemplar for the utility of CPS in other animal agriculture applications. This program will have significant impact on the engineering of CPS by providing a reference architecture for precision animal agriculture that applies to any herd of animals (e.g, cows, goats, chickens, fish). It also contributes to the technology of CPS because novel and trusted principles of systems engineering processes will be used to design and integrate components of the dairy CPS.

Suggested Key words: precision agriculture, active sensing, distributed analytics, animal metabolism models

Intellectual Merit: In order to achieve a generic CPS for precision animal agriculture, a reference architecture that comprises a multi-layer networked system of active sensors and analytic units, with protocols, is proposed. The system will selectively stream real-time, personalized data of each herd member as well as the overall herd, to analytic units appropriately distributed from the edge of the network to the cloud, for learned and pre-programmed analyses at different levels. A structured hierarchy of networks from the internet to mobile ad-hoc networks to wireless body sensor networks will methodically extend the reach of sensors to right inside of the animals themselves (rumen environment). A hierarchy of analytics protocols both in-situ and ex-situ will permit active sensing at all levels, as the computational topology must adjust to the needs for communication. The innermost level of the hierarchy will be millibot sensor nodes used for intra-body data gathering. Control-theoretic adaptive algorithms will ensure data-driven adaptation to every individual in the herd and the overall herd. An exemplar system based on dairy cows will guide our network, controller, and sensor development to increase the production efficiency and wellbeing of every animal on the farm, as well as the collective efficiency of the herd/flock/school.

Broader Impacts: There is a persistent need to mitigate the negative effects of livestock production on non-renewable and renewable resources. Our strong contribution to satisfy this need is the development of a generic CPS for precision animal agriculture that can be applied across the farming industry. This new CPS will allow for efficient feeding and health management to optimize milk production while simultaneously reducing the environmental footprint of the US dairy industry. Improving efficiency and/or the quality of life for farmers and animals will have an enormous societal benefit with global implications. Further, this grant will provide training opportunities for 4 PhD students and 2 MS student. Four students will be trained as engineers and two will be trained as an animal scientist. In addition, the team will use the concept of "the connected cow" in outreach to K-12 children at science camps in both West Lafayette and Blacksburg as part of existing programs to excite our largely agrarian communities about STEM. Lastly, we plan to expose the potential benefits of this CPS research to practitioners and researchers with a novel workshop format that includes a live hack-a-thon.