An IoT Architectural Framework for Dynamic Cattle Supplementation

Abstract

The growth of the world population challenges the demand for beef production. In this context, the animal nutrition process must be essential to achieving the desired herd production performance. In the world's largest beef-producing and exporting countries, although forage is the primary source of animal nutrition, the degradation of pastures forces livestock farmers to provide their herds with supplements as a complementary source of animal nutrition. In addition to being a process that requires a large workload and time, the bovine supplementation activity is impacted by losses caused by incorrect dosages resulting from imprecise nutritional programs. Livestock farmers have used agricultural machinery to assist in the supplementation process, such as Programmed Automatic Feeders (PAFs), which provide pre-determined doses of supplement to the herd in a programmed manner, reducing the intensive work spent in the treatment process. Despite all the technological equipment that PAFs may have, such as radio frequency antennas for identification (RFID) of animals, rain sensors to prevent the supplement from being supplied during rainy periods, etc., PAFs currently do not have integration with Decision Support Systems (DSS) and with technologies that take into account animal characteristics, pasture performance, and environmental parameters to assist in the optimized decision-making process for supplementation. There is a technological gap that could help the decision-making process regarding supplementation to be more optimized and dynamic. Taking these aspects into account, the objective of this work is to propose an IoT architectural framework capable of integrating animal welfare monitoring technologies (bioacoustic collars, accelerometers, etc.), pasture condition monitoring technologies (spectral sensors), environmental sensors (temperature and humidity sensors, for example), and nutritional programs to develop the dynamic supplementation process in cattle herds. This thesis is made up of a collection of five academic articles, where Chapters 1 and 2 propose a set of references framework to motivate the development of the work; Chapter 3 presents a literature review article, which details the main existing techniques and technologies that contribute to improving the supplementation process and demonstrates the first design of an IoT architecture for the dynamic supplementation process; Chapter 4 presents two articles: the first of which details the design and development of machine learning models for predicting forage conditions and the second demonstrates the data collection process to create a dataset to validate the proposition of the IoT architectures designed in the later sections; Chapter 5 consolidates and validates IoT architecture design for dynamic cattle supplementation through two articles; and Chapter 6 presents the conclusions and contributions of this thesis.