Smart monitoring and data-driven adaptive control of livestock environment in dairy cattle barns: development of a predictive modeling and forecasting framework to enhance animal welfare and energy efficiency

The growing demand for sustainable and welfare-oriented livestock production has intensified the need for more intelligent and adaptive environmental control systems in dairy farming. This research proposes a data-driven framework for forecasting key environmental indices in dairy cattle barns by integrating Internet of Things (IoT) technologies with neural-network-based predictive models, with the aim of enhancing both thermal comfort and energy efficiency. Specifically, two of the most widely used thermal indices the Temperature-Humidity Index (THI), and the Equivalent Temperature Index (ETI), were selected. These indices were modeled using high-resolution environmental data collected by a Smart Monitoring System deployed in two commercial dairy farms. The IoT system enabled the continuous acquisition of environmental parameters at several representative points within the barns, allowing the generation of multizonal representations of THI through spatial interpolation. This approach provided new insights into the spatial dynamics of thermal variability, including the estimation of conditions in unmeasured areas. Moreover, the integration of an anemometer into the monitoring system represents one of the principal innovations of this research, as it allowed for the inclusion of wind speed, a critical factor influencing animal thermal comfort, in the computation of ETI. A complementary case study enabled the simulation of energy consumption in the barn’s ventilation system by comparing traditional control methods based on THI measurements with an alternative approach using ETI as the control variable. The results demonstrated that the proposed ETI-based strategy, combined with multi-point monitoring, enables a more localized and adaptive ventilation control, resulting in significant energy savings while maintaining optimal animal welfare. These findings highlight the potential of predictive modeling and IoT integration to foster sustainable, data-driven management of environmental conditions in modern dairy farming, addressing one of the major challenges in the current environmental and geopolitical context.