Precision agriculture for grazing and animal health management : a case study in Colombia

Abstract

In this research, we address the problem of fattening management and animal health in rotational grazing. For the study of this problem, we have positioned ourselves in the framework of the paradigm of precision farming, a technological approach that uses advanced information and communication tools and techniques to optimize agricultural and livestock production processes. In this context, precision livestock farming focuses on the use of technologies to improve grazing management and animal health on cattle farms. Some objectives of precision livestock farming are to increase farm efficiency and productivity, improve product quality and reduce production costs. In addition, it also contributes to environmental sustainability by enabling more efficient management of natural resources and reducing negative impacts on the environment. Although precision livestock farming offers many opportunities to improve efficiency and sustainability in livestock production, it also presents challenges that must be addressed for successful implementation. To address this problem, our objective was to develop methodologies, models, and approaches to support decision-making related to productivity management and animal health. To achieve this objective, several sub-objectives were raised, the first one was to develop a precision livestock farming architecture based on emerging technologies (Industry 4.0, artificial intelligence, etc.), the second on developing generic knowledge models of precision livestock farming for animal health and herding management and finally, in the third to develop meta-intelligent models for precision livestock farming in the context of autonomous grazing and animal health management. In general, several research articles were developed to meet the objectives proposed in this thesis. Initially, a review article on the latest trends in precision livestock farming using machine learning techniques was carried out. On the other hand, for the first specific objective, an article was conducted where three autonomous cycles of data analysis tasks based on autonomous computing were proposed for a beef production process for precision livestock farming. To meet the second specific objective, three articles were proposed. The first is a beef cattle weight identification model using machine learning techniques for anomaly detection, the second presented a system for monitoring the cattle fattening process in rotational grazing using fuzzy classification, in the third, a multi-objective optimization model was developed to maximize weight gain of cattle in rotational grazing. Regarding the third objective, three articles were developed, the first one proposed an autonomous cycle of data analysis tasks for the self-supervision of animal fattening in the context of precision livestock farming, and the second article presents a management system for the cattle fattening process in rotational grazing by means of diagnostic and recommendation systems. Finally, the last article proposed the use of the meta-learning paradigm in a cattle weight identification system for anomaly detection. In each article, we evaluated the strategies/models using various datasets. The results showed the capacity of the developed methodologies and models for decision-making in the management of the livestock production process. Specifically, our proposals allow the management of fattening and animal health in rotational grazing, considering, among other things, monitoring, diagnosis, and optimization of the productive process, with good results in performance metrics.