By Colin Usher
Georgia Tech Research Institute
Special to Poultry Times
ATLANTA — We live in unprecedented times. A novel virus (COVID-19) is spreading across the globe like wildfire, shutting down entire countries and wreaking havoc on all industries including food production. On top of that, a highly pathogenic avian influenza (HPAI) H7 was confirmed in a commercial poultry flock along the Mississippi Flyway in early March, the first such outbreak in the United States since 2015. This has dealt a double blow to the poultry industry and highlights the need for safe and effective bio-safety practices, both for humans and chickens.
This new environment we find ourselves in has created a surge of interest in robotic systems. China and Italy recently bought dozens of robots utilizing 254 nanometer UVC light for disinfecting hospital beds and rooms. China even deployed robots to assist with taking patient’s temperatures and delivering food, allowing the doctors and nurses to have reduced contact with virus-stricken patients. Even big box retailers are starting to utilize robots for carrying out janitorial tasks. These are but a few examples highlighting how robots could prevent human contact from spreading the virus.
Researchers in the Georgia Tech Research Institute’s Food Processing Technology Division have been working for several years on the development of a robotic system that can tend to poultry flocks. Like these robots introduced in the medical field, a primary motivating factor in this research is to reduce the chance of contamination and introduction of disease into the flocks.
The idea is a simple one: the more that a robot can do, the less a person needs to intervene, reducing their need to enter the houses and therefore reducing the opportunity for contamination.
Currently, there are a handful of robot systems available for working in commercial houses. Each of these systems primarily carry out a single task, such as monitoring the flock, or encouraging flock movement, and even disinfecting. With these single purpose systems, humans still have to enter the house on a daily basis to carry out manual tasks such as removing mortality or picking up floor eggs.
With this in mind, the research team set out to design a robot that can carry out as many tasks as possible, reducing and even eliminating the need for a human to enter the house at all.
To this end, the robot has to be able to move autonomously through the flock, actively sensing the birds and reacting to them so as not to harm them. This work was completed in 2015 and was demonstrated with successful field tests in 2016 where the robot navigated among a flock in a commercial house.
To date, the robot has operated fully autonomously in commercial houses and smaller scale test houses replicating commercial facilities for multiple hundreds of hours.
However, autonomous navigation is only the first task the robot must be able to do if one wants to remove humans from the equation. The robot must be able to identify chickens and equipment and provide information to farm management.
Artificial intelligence routines were added to the robot allowing it to detect chickens, feeders, drinkers, and even eggs. These developments were demonstrated in 2017 and 2018.
With the ability to detect chickens and eggs, a small robotic arm was added to the platform, and routines for approaching and automatically removing the floor eggs were developed. This required the development of novel routines that allow the robot to search spaces, ensuring the entire area is covered.
The robot is now able to search an entire commercial broiler-breeder house floor, find eggs, drive up to them, pick them up, and remove them from the house. Field testing in a small-scale test house yielded a success rate of greater than 90 percent in picking up floor eggs. Testing in actual commercial houses was planned for this year but has been postponed due to the virus outbreaks.
With the ability to autonomously navigate the houses, identify chickens and equipment, and find and pick up floor eggs, one significant task remains that requires humans to enter the houses on a regular basis. This task is removal of mortality. Current work on the robot is focusing on the ability to detect and remove mortality. To this end, a custom robot arm was developed that is both low-cost and strong enough to pick up full-grown chickens. A prototype arm was developed and demonstrated in 2019 for picking up chickens weighing up to 20 pounds. This effort is currently ongoing.
While the world focuses on hospitals and hot spots for humans, opportunity exists for improving the state of the industry with animals as well. In an editorial published in the journal Science Robotics, experts suggested that more needs to be done now.
They also remind us that similar plans for robotic assistance were created after the 2015 Ebola outbreak, but the funding and motivation dropped off when the outbreak was resolved. Perhaps interest in these systems can be sustained a little longer this time.
Colin Usher is a senior research scientist in the Georgia Tech Research Institute’s Food Processing Technology Division. This article is reprinted from PoultryTech, a publication of ATRP of the Georgia Tech Research Institute, a program conducted in cooperation with the Georgia Poultry Federation with funding from the Georgia legislature.
