Monitoring behavior of poultry based on RFID radio frequency network

Poultry behavior monitoring is an important basis for the poultry disease warning. Manual monitoring is mostly used nowadays. In this work, the automatic monitoring system for assisting manual monitoring was examined. Sophisticated data mining techniques were used to leverage the data collected by RFID devices. Specifically, (1) weighing sensors and wireless networks of Multiple RFID-tag-collector groups were used to monitor the poultry behavior; (2) RFID tags were putted on individual poultry so that the moving time of the poultry between two RFID-tag-collectors could be recorded. Thus, the characteristic functions of poultry behaviors such as speed, ability to snatch food and resting time could be extracted based on the distance between two RFID-tag-collectors and the relevant time parameters; (3) the sick, normal, active and other poultry groups were categorized by using the K-means method which utilizing the behavior characteristics and poultry weight data in data mining. The results demonstrated that accurate classifications could be obtained according to the poultry characteristics, and the clustering results matched with the results obtained by manual method to identify the poultry groups. Consequently, the technique in this paper has great potential for large-scale poultry disease warning and poultry classification.