Precision feeding of sows reduces feed costs by improving sow’s efficiency.
•
Behaviour and health status can strongly influence individual nutrient requirement.
•
Thermal conditions and housing type can affect nutrient requirement.
•
Machine learning to automatize and improve precision feeding.
Abstract
Taking into account individual variability while feeding a group of sows allows feed cost reductions and therefore improves animal efficiency. This precision feeding strategy is based on 1) nutritional models, which are able to predict daily individual nutrient requirements; 2) automatons, that can deliver individual rations; and 3) new technologies such as sensors which provide real-time information on the animal performance and life conditions that should be integrated into the estimation of requirements. Up to now, only production data (body weight, backfat thickness) have been integrated into the calculation of individual nutrient requirements.
However, the literature reported that health status and behavior, such as physical activity, social behavior, and location in the pen, can strongly influence nutrient requirements. A change in the feeding or drinking behavior can also indicate a health or welfare problem. Sensors, automatons and cameras are now able to detect some diseases or injuries, and record certain on-farm behaviors. Therefore, nutrient requirements should be adjusted based on these health and behavioral parameters. Environmental factors such as thermal conditions, housing type and noise level have also been reported to affect nutrient requirements. On-farm sensors can easily be installed to record these parameters to be integrated into the nutritional model and improve its precision. A decision support system can be used to integrate these new measurements into the nutritional model for gestating sows. It would also be helpful to trigger alerts and propose corrective actions when behavior changes or health issues are detected.
