Internet of Dairy: Monitoring of udder health in dairy cattle using innovative sensor data and intelligent data processing

This research focuses on how to use data to monitor udder health in dairy cattle and give early warnings for mastitis, a severe disease in dairy farming. Mastitis is the most widespread and economically challenging disease in dairy cattle [26], as it can directly affect milk production and quality, costing up to 277₤ per infection [13] and affecting up to 70% of a given herd [6]. The average economic costs for clinical mastitis are estimated to be over 300 billion dollars per year in China. In this research, I am going to develop algorithms for mastitis detection and monitoring based on milk quality information that is measured by innovative on-farm sensors. Using traditional statistical method based on change point detection, the change point of cattle from healthy state to mastitis is estimated and a recovery parameter can be derived. Several change point detection methods, both supervised and unsupervised methods, will be implemented and evaluated. By comparing the performance of these methods, the most effective way will be chosen and analysis about its performance will be made. In turn, the research would provide an early warning and monitoring system that allow us to concentrate treatment efforts on infected cows and follow up on the recovery and cure.