Unplucked: Robotics and Automation Take Center Stage
TLDR: Listen now:
Colin Usher, Senior Research Scientist and Robotics Systems and Technology Branch Head at Georgia Tech, has over two decades of experience in machine vision, robotics, and system integration. He joined PSA Executive Director Andy Vance on a recent Unplucked episode to share his insights into cutting-edge technologies designed to enhance efficiency, animal welfare, and sustainability in the poultry industry.
Usher currently collaborates with several multidisciplinary teams to develop and evaluate industrial automation prototype systems for a variety of clients in the food and agribusiness space.
“We like looking at novel automation techniques and opportunities where we can apply advanced technologies and engineering sciences into poultry processing and production, to see where it fits and what makes the most sense,” he says.
While some companies have attempted to deploy robotic solutions, applications inside poultry houses are still in their infancy. Even so, Usher has already invested more than a decade into developing robots for commercial poultry farms.
Initially, they created a mobile sensor platform that moved through flocks, characterizing traits such as uniformity, growth rate, and animal welfare. While farmers found it interesting, they couldn’t justify the cost, as they still needed to walk through the houses for flock management.
This led the team to begin exploring other applications beyond just sensing what utilitarian things robots could do. Today, they’ve developed robots for picking up eggs and deceased chickens, “reducing or eliminating the need for people to go into the flocks.”
Much like a Roomba vacuum, the robots carry out mundane, repetitive tasks every day. Egg picking, for example, is very much like that, Usher explains. “When a hen lays an egg on the floor, other hens see that egg and they say, ‘Oh, that’s a nest; I’m going to lay an egg on the floor, too.’ So, it’s very important to regularly remove the floor eggs throughout the day.”
Paying someone to just stand in the house all light hours of the day and pick eggs up off the floor isn’t practical, Usher adds, making egg picking a good application for automation. “Things that happen all day long are things that adapt well to these types of systems.”
Additionally, there are other applications where the robots can gather data more robustly than humans. Usher says there are instances where producers are doing more qualitative assessments of their flocks, including smelling the air and listening to the birds, but there is also opportunity for 3D profiles of the birds as well as uniformity measurements of the birds.
Return on investment
The infamous ROI — it isn’t the end-all, be-all, but for now, robots remain cost-prohibitive. The good news, Usher notes, is that the pace of technological advancement continues to accelerate, while costs are steadily decreasing.
“A chicken farmer isn’t going to buy a robot, no matter what it does, if it costs $75,000-$100,000 per house,” Usher says. “As the costs drop and as the capabilities become more advanced, we’re able to do more advanced things.”
Robotics in poultry houses will drive innovation, ultimately enabling producers to better manage flocks and optimize growth, enriching the production environment.
Another emerging consideration is modifying or designing poultry houses to better support automation. For instance, feeders could be automatically raised to allow robots to move through more efficiently. “This would, in effect, move chickens out of the way.”
Other possibilities include automating smart drinkers and feeders to notify producers of leaks and jams, or to control feeding.
Barriers and bottlenecks
Usher says people have shown great interest in the robots he and his teams have developed, but as a research institution, they are not selling them commercially. “Cost is still a bit of a challenge, and the payback cannot be quantified yet.”
Producers must assume the risk, and many aren’t ready to take that leap, which limits opportunities to prove the systems.
“Moving from research to commercialization is always a challenge,” Usher emphasizes.
There’s engineering work that needs to be done— such as redesigning the robot chassis for better durability and longevity, as well as ease of cleaning. “Ammonia is prevalent in these environments, and electronics hate that,” Usher adds.
Virtual reality
Historically, more work was done in the processing field with robot arms carrying out repetitive functions like deboning in unattractive, colder environments. However, accuracy has been a challenge.
To address this, researchers have incorporated virtual reality headsets to allow humans to guide the robots remotely. “We can allow them to make the decisions for the robot when it’s not able to make the decision for itself,” Usher explains.
The same process has been used for cone loading, giving humans the capability to tell the robot which chicken to pick up. “As fast as they can point and click, that robot is in an operation loading those cones.”
All of this can be done remotely—even from someone’s living room. Still, cost remains a major hurdle as it is hard to justify the expense of both purchasing the robot and continuing to pay someone for oversight.
Artificial intelligence may be part of the solution to some of these problems, but not without its own set of challenges.
Listen to the full, insightful discussion on PSA’s original podcast Unplucked here.