PS2-14. Monitoring Cattle Respiration Rate Using a Wearable Accelerometer and Digital Signal Processing.

Abstract Technologies that enable continuous, individual-level monitoring of livestock physiology are increasingly important for digital management systems. Respiration rate (RR) is a practical indicator of thermoregulatory demand and respiratory health, but routine, continuous RR monitoring remains challenging in commercial settings. Non-contact video approaches can estimate RR from thoracoabdominal motion, yet performance may be sensitive to occlusion, posture changes, and unrelated body movements. This study evaluated a non-invasive alternative that uses a lightweight wearable accelerometer and signal-processing to extract respiratory motion from on-animal acceleration data. A tri-axial accelerometer (Axivity AX6; 16-bit; 100 Hz) was positioned near the abdomen in a pocket of a heated blanket worn by Holstein–Friesian cow to capture cyclic expansion associated with breathing. In this pilot feasibility evaluation (approximately 1 h of recording), raw acceleration signals were resampled as needed, band-pass filtered (0.3–1.0 Hz), and analyzed using fast Fourier transformation to identify the dominant respiratory frequency. Respiration rate estimates derived from acceleration were cross-referenced with synchronized video-based flank movement counts. Across the evaluated recording, the dominant respiratory frequency corresponded to 0.67 to 0.80 Hz (40 to 48 breaths/min), which aligns with commonly reported resting respiration ranges for adult dairy cattle. These results support the feasibility of extracting RR from blanket-mounted accelerometry and motivate evaluation across broader activity states and environmental conditions, with attention to placement stability and motion artifact handling. For image description, please refer to the figure legend and surrounding text. For image description, please refer to the figure legend and surrounding text.