DigitWelfare. A hybrid modelling approach to characterize dairy goat's activity profiles associated with welfare

PhD fellow: Sarah Mauny: https://orcid.org/0009-0006-1986-9330

Starting date : 01/11/2022. Funding: 50% MathNum, INRAE ; 50% WAIT4 PEPR Agroécologie et Numérique

Animal behavior is known to be a key indicator of animal welfare, as it can provide early signs of variations in welfare. Artificial Intelligence and sensors such as accelerometers are promising tools for predicting animal behavior individually and objectively. This PhD project aims to develop a generic pipeline using a supervised classification algorithm to automatically predict animal behavior from accelerometer data. The pipeline will be used to detect specific animal behaviors and then characterize patterns in the detected behaviors, referred to as activity profiles. After a perturbation, the goal is to detect changes in these activity profiles, which can indicate disruptions in production performance and serve as early signs of health and welfare issues. In the short term, this approach could enhance real-time monitoring and immediate intervention capabilities, improving animal management practices. Over time, it could contribute to a deeper understanding of the link between animal behavior and welfare, supporting the development of more sustainable and welfare-friendly livestock systems.