Monitoring livestock behavior in real-time promises to open new possibilities to achieve early disease and lameness detection, bet-tet reproduction decisions and farm management, aggression and cannibalism monitoring, welfare assessment, and even moreaccurate environmental control, among others.1 Collecting real-time behavioral metrics in commercial farms is no simple task. A good livestock behavior tracker would be both precise and affordable. In the last years, researchers and companies have trieddifferent types of behaviot data acquisition systems, but the resulting solutions are still not widely adopted. Sectors in which animals have a greater value such as dairy have partially adopted accelerometer and indoor positioning systems while other sectors like swine or poultry have not been able to integrate those in their business model due to the high cost of the systems and the difficulty to implement them at a large scale. Radio Frequency Identification (RFID) has also been tried1' with different species to monitor visits to some predetermined areas, giving interesting results, but its physical implementation is complex, and the generated data is somehow poor. In fact, RFID requires antennas at different locations and tells when an animal is detected at those positions only. Therefore, we have no information about the behavior of the animal in between detections at the antennas.
展开▼