Safe tracking of farm animals with a smart leash

In today's farms, the milk yield of animals is an important parameter for the farm. In animal diseases, milk yield may decrease or even end with the death of the animal. It is extremely important to detect the disease in the animal at its initial stage. For this reason, it is desired to use technological opportunities in farms. It is aimed to observe with sensors instead of analysis and tracking done by the human procedure. In this thesis, an electronic design has been made for the collar, pedometer and stomach apparatus. In line with the design, products were developed and prototypes were prepared. Through the sensors on the collar, the number of ruminations made by the animal and the amount of water drinking was found. The distance taken from the pedometer, the number of steps taken, the lying time and the standing time were reached. Stomach temperature and pH values were obtained in the gastric apparatus. To monitor all data, animals can be analyzed and tracked in the web interface developed with .NET Core. Maximum efficiency is tried to be achieved by applying special identification to each animal with sensor values from each animal. In addition, estrus classification was made using the data obtained from the pedometer and using Xgboost, Adaboost, Support vector machine and Naive Bayes classifiers. The maximum performance was obtained from the Xgboost classifier with 83.6%. Thus, it is aimed to make milk yield and insemination at maximum efficiency for fattening farms. It is thought that the data to be obtained will contribute to future studies.