A systematic literature review on deep learning and IoT for livestock management, monitoring, and anti-theft applications

In recent years, the integration of Deep Learning (DL) and the Internet of Things (IoT) has brought new possibil- ities to livestock management, offering smart ways to monitor animal health, behavior, and security. Yet, several challenges remain. These include the high cost of deploying advanced sensors in rural areas, inconsistencies in data collected from different environments, and the limited ability of models to adapt to varying farm conditions. There’s also a lack of standard datasets and difficulty in achieving real-time, reliable results at scale. In this paper, we present a detailed review of the current state of DL and IoT technologies in livestock systems. Using the PRISMA framework, we reviewed 50 studies from reputable sources such as Scopus, IEEE Xplore, and Web of Science. Our analysis covers key use cases—including animal identification, tracking, health monitoring, and theft prevention—and highlights the deep learning models most commonly used, such as CNNs, RNNs, LSTMs, and SVMs. This study contributes by offering:
(1) a clear picture of how DL and IoT are being applied in real-world livestock settings, (2) a categorization of models and methods by application area, and (3) insights into ongoing technical and deployment challenges. Looking ahead, future work should explore the creation of open and diverse datasets, the development of lightweight AI models suitable for farm-based edge devices, and privacy-aware solutions that ensure both data security and scalability for smart agriculture.