HenTwin: Digital Twin for Laying Hen Biological Monitoring

Monitoring the early-life health and development of laying hens is crucial for poultry farming, but current methods are often fragmented and lack comprehensive system-level state representations. To address this, researchers have developed HenTwin, a multimodal digital twin framework. Implemented as a five-layer IoT architecture, HenTwin formalizes the biological state dynamics of an entire flock from hatching up to 25 weeks of age. The framework defines a four-dimensional biological state vector, incorporating data such as body surface temperature, acoustic energy entropy, band energy ratio, and optical-flow-based motion, with environmental factors like temperature-humidity index as exogenous inputs. A discrete-time state transition model, estimated from 25 weeks of longitudinal data from 150 hens, reveals modality-specific persistence and asymptotic stability. Perturbation analysis demonstrates how environmental changes impact biological states, and cross-room validation suggests partial transferability of structural parameters, supporting a two-tier IoT deployment. HenTwin represents a significant step towards formal, state-aware digital twin inference in precision livestock farming.