BovineTalk2.0

Machine learning and IoT based approaches for monitoring dairy cattle health and welfare using bioacoustic data

The BovineTalk2.0 project seeks to refine and validate a novel, non-invasive methodology for assessing the welfare and health of dairy cows using vocal indicators. Building on preliminary results (BovineTalk project, code PN-III-P1-1.1-TE-2021-0027, implementation 2022-2024) that demonstrated the ability to record and classify cattle vocalizations with high accuracy, we aim to transition this research to a higher Technology Readiness Level (from TLR 2 to TLR 4 at the end of the project) by collaborating with industry partners specialized in precision agriculture. This will involve developing a prototype platform for real-time, on-farm monitoring, involving a custom built wearable device for audio recording, enhanced machine learning models and a prototype user interface to assist farm veterinarians in assessing animal welfare. The overarching objective of our work is to improve the welfare and health of dairy cows, which, as seen in the One Health concept, is intrinsically linked to better health for all of us. For example, improved welfare can lead to better herd health, reducing the need for the use of antimicrobials and thus reducing inputs of chemical synthesis in the bioeconomy. Moreover, stress-free animals often have better feed conversion ratios and milk yields, enhancing farm productivity and efficiency. This aligns with climate-resilient practices by ensuring that agriculture systems can adapt to and recover from adverse conditions while minimizing environmental impact.