Between the scarcity of labor and live birds hiding the deceased ones, poultry farmers face major challenges finding and disposing of mortalities on the farm.
In response, a commercial manufacturer has created a mobile robotic system equipped with multiple cameras that traverses the ceiling to identify mortalities in the barn. At the 2025 International Processing and Production Expo, Tanner Thornton, a graduate research assistant at the University of Tennessee Institute of Agriculture, discussed a test of the system, which he said is an example of precision livestock farming (PLF) technology.
Thornton concluded that while the robotic system can detect mortalities, it needs to improve its peripheral vision and ability to disperse birds in order to achieve maximum results. This research was the first time this PLF had been studied in a commercial setting.
Farms face labor shortages
Thornton noted that poultry is the fastest-growing livestock industry. In 2019, chicken became the most consumed meat product in the world, surpassing pork, he reported.
Meanwhile, all segments of the agricultural industry are experiencing labor shortages. Farms experience high turnover rates due to challenging conditions for workers, including exposure to ammonia, dust, and zoonotic diseases.
One of the most difficult tasks for poultry workers is walking the farm daily to search for and remove the carcasses of deceased birds. On a more positive note, Thornton said these walks encourage movement of live birds and help improve their leg strength.
“Timely removal of mortalities is key to prevent disease spread, maintain optimal air quality, and reduce the attraction of pests and rodents,” Tanner explained. However, for a farmer with multiple 600-foot (183-meter) broiler houses, for example, this task is time-consuming and leaves less time for other chores such as maintenance and picking up supplies.
Novel PLF robot system offers solution
One proposed solution, a ceiling-mounted robotic system equipped with multiple cameras, is an example of PLF, which Thornton said offers “real-time continuous monitoring of animals via cameras, sensors … in an effort to improve animal quality of life and efficiency.” Compared to the swine and dairy industries, however, the poultry industry has been slow to adopt PLF technology.
Thornton and his team studied the potential use of Scout, a mobile robotic system on poultry farms. Scout is equipped with a set of cameras mounted on a rail suspended from the ceiling. It travels a set route in the barn then docks and recharges. Cameras are located on the bottom and both sides of the robot.
As Scout travels, it collects a variety of data, including ammonia, temperature and humidity, as well as data related to animal welfare, including mortalities, free usable space and noise level in the barn.
The advantages of the system include its ability to traverse the entire poultry house and to collect this essential data. The data are put into a map online and sent as an email summary for the producer to review daily.
Thornton’s research team hypothesized that mortality detection with Scout would be limited because of the high density of animals and the presence of equipment blocking its view.
In the study, investigators wanted to compare expected mortalities detected by Scout with the producer’s mortality records. They also wanted to see how well Scout detected mortalities in different areas of the barn and how the stocking density affected Scout’s detection abilities. Based on their findings, the team could then provide recommendations to producers on using and optimizing Scout-type systems.
System challenges
While Scout performed well in some areas, the researchers found challenges in other areas. Scout’s mortality-recognition data may involve duplications in the mortality data as it made passes throughout the barn, which overestimated the total number of dead birds. The duplicate mortality identified by Scout was then manually removed by the researchers; however, the system’s performance in reporting the actual mortality numbers remained subpar.
To determine why, the researchers conducted a series of tests to identify the factors affecting mortality detection by Scout, including staged mortalities along the rail path for Scout’s three cameras to identify, both inside and outside of a specific trial pen.
The research team found that most of the time Scout found those mortalities directly underneath it. The side cameras, however, performed poorly, mainly because equipment and feed and water lines occluded the views from the sides.
Scout also found “significantly more” mortalities inside the trial pen than outside it, suggesting that “the stocking density in the house played a major role in Scout’s ability to find birds” as occlusion by live birds compromises the performance of mortality detection by Scout’s cameras, especially the side cameras.
Meanwhile, though Scout’s camera captured birds up to 20 feet away from the camera, it only processed data for up to 10 feet. So the bird carcasses placed farther away from the Scout system were never found. This could be addressed by upgrading camera resolution, but that would increase the cost of the technology, Thornton noted.
Thornton concluded that the Scout system is “effective at finding mortalities when in clear view of a mortality, but it still needs to improve field of vision and develop a way that it can disperse the birds so that it can get a clear field of vision of the birds.”
