A new study conducted by researchers from Penn State, University of Kentucky, and University of Vermont has found that monitoring dairy calves with precision technologies based on the “internet of things” (IoT) can lead to earlier diagnosis of bovine respiratory disease, which is highly fatal for calves. The researchers believe that this novel approach can help dairy producers improve the economies of their farms.
Lead researcher Melissa Cantor, assistant professor of precision dairy science in Penn State’s College of Agricultural Sciences, emphasizes that this strategy is far different from traditional dairy farming methods. Advancements in technology have made it more affordable for farmers to detect animal health problems early, allowing for intervention and saving both the calves’ lives and the investment they represent.
IoT refers to devices embedded with sensors, processing and communication abilities, software, and other technologies that enable them to connect and exchange data with other devices over the Internet. In this study, Cantor explained that wearable sensors and automatic feeders equipped with IoT technologies were used to closely monitor and analyze the condition of the calves.
These IoT devices generate a significant amount of data by closely monitoring the behavior of the cows. By adopting machine learning, which is a branch of artificial intelligence that identifies hidden patterns in data, the researchers made the data easier to interpret and utilized it to detect health problems in the calves.
Cantor said, “We equipped the calves with leg bands that record activity behavior data in dairy cattle, such as the number of steps and lying time. We also used automatic feeders that dispense milk and grain and gather feeding behavior data, such as the number of visits and liters of consumed milk. The information from these sources enabled us to identify when a calf’s condition was deteriorating.”
Bovine respiratory disease is an infection of the respiratory tract and is the primary reason for antimicrobial use in dairy calves. It is also responsible for 22% of calf mortalities. The costs and effects of this disease can be devastating to a farm’s economy since raising dairy calves is a significant investment.
Cantor stated, “Diagnosing bovine respiratory disease requires specialized and intensive labor, which can be difficult to find. Therefore, precision technologies based on IoT devices, such as automatic feeders, scales, and accelerometers, can help detect behavioral changes before outward clinical signs of the disease appear.”
The researchers collected data from 159 dairy calves using precision livestock technologies and by conducting daily physical health exams on the calves at the University of Kentucky. They compared the results obtained from automatic data collection with those obtained from manual data collection.
In their findings published in the peer-reviewed scientific journal, IEEE Access, the researchers reported that their proposed approach was able to identify calves with bovine respiratory disease earlier. The system achieved an accuracy of 88% in labeling sick and healthy calves. Additionally, 70% of sick calves were predicted four days prior to diagnosis, and 80% of calves that developed a chronic form of the disease were detected within the first five days of sickness.
Cantor said, “We were surprised to find that the behavioral changes in these animals were different compared to animals that responded to treatment. This had not been studied before. We concluded that if these animals exhibited different behavior, IoT technologies empowered with machine learning techniques could potentially identify them earlier, before it is visible to the naked eye. This offers producers more options.”
The research was conducted by Enrico Casella from the Department of Animal and Dairy Science at the University of Wisconsin-Madison; Melissa Cantor from the Department of Animal Science at Penn State University; Megan Woodrum Setser from the Department of Animal and Food Sciences at the University of Kentucky; Simone Silvestri from the Department of Computer Science at the University of Kentucky; and Joao Costa from the Department of Animal and Veterinary Sciences at the University of Vermont.
Support for this research was provided by the U.S. Department of Agriculture and the National Science Foundation.