Systematic Literature Review on Deep Learning-Based Approaches for Broiler Chicken Growth Forecasting
Year: 2026
Type: review
Study design: Systematic Review
Abstract: Ensuring optimal biomass growth and automated health monitoring in broiler chickens remains a critical challenge for precision livestock farming. This systematic literature review comprehensively examines the integration of deep learning architectures, specifically focusing on Differential Recurrent Neural Networks (DRNN) and the Improved Feature Fusion Single Shot MultiBox Detector (IFSSD), to facilitate real-time growth forecasting and anomaly detection. Adhering strictly to Kitchenham, PICOC, and PRISMA methodologies, 30 high-quality studies were synthesized from an initial pool of 150 database records. The empirical synthesis demonstrates outstanding model capabilities, highlighting that DRNN structures forecast broiler development with a marginal error rate of 1.8%, while IFSSD architectures achieve a 99.7% accuracy threshold in clinical health diagnostics. Despite these high precision levels, multidimensional bottlenecks persist, specifically computational overhead, data scarcity, and high susceptibility to open-house microclimate fluctuations. Beyond evaluating model efficiencies, this study outlines critical technical and non-technical barriers, offering strategic pathways for hybrid geospatial-temporal model scaling. Ultimately, this review bridges the gap between neural network engineering and poultry science, establishing a foundational roadmap for sustainable, data-driven livestock productivity.
Institution: Institut Teknologi Budi Utomo
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
SDG: Zero hunger
Open Access status: hybrid
OpenAlex ID: W7225155366