| dc.description.abstract | "An automated system can be used to monitor and classify different behaviors in animals. By monitoring these behaviors, health problems of animals can be detected. It is useful, if these systems are used to monitor cattle such as dairy cows. It is important for farming and dairy farmer to monitor status of cows from an animal welfare and dairy farmer’s economic point of view. In this study, we have developed two monitoring systems, with different needs that needed two different transmission protocols. The basic components are discussed in the thesis in detail. In Bluetooth based system, Bluetooth transmission protocol is used while
and in ISM (Industrial, scientific, and medical radio band) based system, ISM protocol is used for radio transmitter and receiver. In ISM based system, multiple transmitters-single receiver protocol has also been developed. The classification of the activities is done on the basis of threshold based classification, binary decision tree and Support vector machine (SVM) and compared. The threshold based classifier gives the better accuracy of 99.99%, while the accuracy with binary decision tree classifier is 89% and with the SVM is 91.48%. The estimation of battery for Bluetooth based system is 11 days approximately, while for ISM based system is 255 days approximately. Finally, results are discussed with the system tested within a barn to monitor the activities of the cow like sleeping, standing, walking and lying while attaching the device on one of the legs." | |
