ABSTRACT
Abstract. A portable electronic nose was constructed to detect the odours released by cattle in oestrus. It consisted of conducting polymer sensors, a system for sampling, humidity management (known as the olfactory lens) and data analysis. It was used in a number of experiments to detect oestrus in cows with different polymer based electronic noses. The method developed was to take swabs of cervical mucus from the cows and place them in a sample chamber. The humidity emanating from the swab was measured and the humidity surrounding the sensors was adjusted to minimise the change in humidity experienced by the sensors. The change in electrical signal measured by the sensors was recorded. A novel method of data analysis - the wavelet function was adapted and applied to the transient signals from the odour sensors. The wavelet function coefficients characterised the signal from each sensor. By taking the ratio between the coefficients representing the most active sensors in the array a two dimensional vector plot of each days sample was drawn. The method was applied to four cows of whom three showed a clear change in odour found in their mucus samples between the day of oestrus and the day of ovulation. This project has shown that with improvements in sample management (the olfactory lens) and signal processing (wavelet analysis) existing electronic noses can be used to detect oestrus in dairy cows but with an invasive sampling technique.
