WAIT4

Improving animal welfare is a key element of the sustainability of animal production systems. The agroecological transition of these systems will have significant impacts on animal welfare, with expected positive effects as well as specific vulnerabilities. To facilitate an agroecological transition of livestock systems that ensures animal welfare, it is essential to develop new tools for assessing animal welfare and decision support. Improvement strategies depend on the precise understanding and evaluation of the physiological and behavioral dimensions of animal welfare, as well as the perception that each animal has of its environment.

The WAIT4 project will leverage new opportunities offered by digital technologies to measure the various components of animal welfare in real-time and implement new artificial intelligence approaches to integrate the large volumes of heterogeneous data collected through these devices. To achieve this, the project is organized into 5 work axis:

• Test and develop equipment and sensors to assess animal behaviors, physiological constants, and emotions ;
• Develop artificial intelligence algorithms to integrate this heterogeneous data in terms of nature and temporality, and extract relevant indicators (proxies) of animal welfare ;
• Provide real-time monitoring of changes in these proxies for different species (pigs, small and large ruminants) raised indoors or outdoors (grazing), in conventional or alternative systems (organic farming, agropastoralism), and in contrasting environmental contexts (tropical or metropolitan climate uncertainties) ;
• Identify early warning signals (early deviations) of changes in animal welfare and health ;
• Foster dialogue and cross-disciplinary collaboration between scientists with different expertise (from ethology to data science) and stakeholders to facilitate the adoption and dissemination of results.

The project's ambition is to go beyond traditional approaches to the study of animal welfare by adopting a holistic approach that better considers the various components of animal welfare and their interactions in the context of the agroecological transition of livestock systems under the constraint of climate change. It will generate new knowledge and develop proxies to better measure animal welfare for each animal within its group, ranging from a few days to several months, and even to the effects of seasons in an agroecological perspective. This will lead to the design of tools for measuring and improving animal welfare. The results will, among other things, contribute to proposing individual or group rearing practices, refining animal welfare assessment grids, and assisting in the definition of selection schemes to accelerate the agroecological transition.

The WAIT4 project brings together a consortium of French research and educational institutes (INRAE, CEA, INRIA/Université Rennes 1, INSA), a living lab (LIT Ouesterel), and a small enterprise (AIHERD). This diversity of actors represents expertise in electrochemistry, physiology, ethology, precision agriculture, data science, and data mining.

Theses

Sarah Mauny (INRAE, UMR 0791 – MoSAR): DigitWelfare – A hybrid modelling approach to characterise the activity profiles of dairy goats associated with welfare.

Adèle Denis (INRAE, UMR 1348 PEGASE): Automated detection and analysis of behaviour between piglets using artificial intelligence.

Joseph Allyndrée (INRAE, UMR 1300 – Bioepar): Development of methods for characterising the behaviour and social interactions of cattle: assessment of health and welfare in the context of agro-ecological transition.

Lucie Lepetit (Inria, LACODAM): Data mining approach for the automatic discovery of farm animal behaviour with regard to animal welfare.

Sacha Germain (Inria): Detection and explanation of individual and collective behaviour within a group to assess their well-being.

Vincent Elouan (INSA, DM2L): Self-supervised learning on heterogeneous graphs: contribution to the analysis of animal behaviour and social dynamics.

Morgane GENIN (UMR 791, MoSAR): Effect of heat stress in ewes: Search for indicators in milk using mid-infrared spectra to assess animal health and welfare.

Postdoctoral researchers

Lucile Riaboff (INRAE, GenPhySe): Use of accelerometers and artificial intelligence methods to assess the welfare of grazing sheep.

Mathile Valenchon (INRAE, Mosar): Behaviour of dairy goats in an enriched environment: behavioural signature associated with improved welfare using data from sensors and machine learning analysis.

Alessio Ragno (INSA, DM2L): Characterisation of animal activities and their social interactions using interpretable graph neural networks.

Caroline Xavier (INRAE, PEGASE): Effect of heat stress on the synchronisation of circadian rhythms of metabolism and behaviour in pigs.