At the speed of automation
Nov. 20, 2012
by Lynn Petrak
If the advent of automation altered the meat-processing floor, the latest robotic systems could practically revolutionize it.
Robotics has been much-talked about in the meat and poultry industry in discussions about ways to improve line speeds and enhance worker safety. Meat cutting, after all, has long been considered one of the most dangerous and challenging jobs. And the drumbeat for greater throughput, faster speeds, improved accuracy and lower costs is stronger than ever, given current industry trends and demands.
There have been sophisticated, human-less features in several of the latest automated meat-processing systems, from carcass splitters to high-tech cutting and slicing systems. Robotic systems also have been used in case and tray packing and warehousing/distribution. Yet, true robotics has generally remained a concept on the processing side, often due to challenges with product consistency and uniformity and the sheer cost of transforming an operation.
As the industry heads into 2013, that concept is edging closer to reality on the processing line, with research and development projects that have resulted in systems featuring high-tech robot hardware, sensing/vision systems and advanced software.
Sensing an opportunity
One ongoing project that has sought to make robotics a reality in modern meat and poultry processing facilities recently garnered national attention, thanks to coverage in the Wall Street Journal. Robotics experts at Georgia Tech Univ., through the Georgia Tech Research Institute (GTRI), have for years been refining a human-less prototype called the Intelligent Cutting and Deboning System that can rapidly debone an entire chicken, leading to greater line speeds and improved worker safety, hygiene and food safety.
Essentially, an advanced vision system makes 3-D measurements of a bird within various points, which are then used with algorithms to identify the right place to cut for the maximum yield. The location of bones, tendons and ligaments guide the process.
Gary McMurray, division chief of the food processing technology division at GTRI, is a principal research engineer for the robotic cutting and deboning system. “This is a project we have been working on for a while, and for us to see the ideas that we envisioned a long time ago come together, it’s been good. For us, we are a bunch of engineers, and we love to solve problems,” he says.
The timing for a robotic cutting and deboning system is especially good for this industry, McMurray notes. “With the downturn in the economy, the struggle with labor, the attention to yield, costs of worker comp and competition from overseas have all come together to make processors aware that they need to automate more. When times are good, you can put off things until later. Now it’s time to get better,” he points out.
As he set out to solve challenges in meat cutting, McMurray enlisted the help of industry experts, including meat scientists. An industrial advisory board has also helped guide the process. “We approached this problem with the basic concept that you cut and run into bone and turn and do something different. How do you detect bone, then?” he recalls.
McMurray and his team determined that a key step was the cut in the first hundred milliseconds, which determined whether something was bone or muscle. “The critical things happen in those first hundred milliseconds, where you can get up to six data points. We need to collect data at those milliseconds, and need 80 or 90 data points before we get can determine if something is meat, tendon or bone,” he explains.
A congruence of technology enabled that capability. “For us, it was also about the ability to have faster algorithms because we have to be able to do these things at high rates of speed,” McMurray adds.
Advanced sensing technology, including the 3-D vision system, fueled the development of the Intelligent Cutting and Deboning System prototype at GTRI. In addition, today’s cameras are superior, and there is an improved ability to transfer more data in a short period of time. “Previously, you had to do a tradeoff with the resolution of the camera with line speed, and now we are able to say we can do more with both,” he says.
A challenge has been with consistency, given the nature of animal proteins, which is why hand deboning and hand trimming have remained tried-and-true methods over the years. “Every bird is different, and the location of the bird [on the cone in front of the visions system] also is different, and makes for more variation,” McMurray says. To overcome that hurdle, the GTRI team utilized better sensors, with statistical modeling, to best predict the bone structure of inconsistent products.
The process of fine-tuning the new robotic system is ongoing, according to McMurray. Most recently, the group at GTRI worked through some hardware issues, ultimately using a different robot.
Although the Intelligent Cutting and Deboning System is currently designed for poultry processing, McMurray foresees broader usage. Some in the pork industry have already reached out to the researchers at Georgia Tech.
The robotics expert also underscores the fact that the system would be different for each industry and user.
Looking ahead to 2013, McMurray says GTRI will have a booth the International Production & Processing Expo (formerly known as the IPE) in January and soon will set up a running demonstration of the robotic Intelligent Cutting and Deboning System.
He believes the technology has the capacity to change poultry and meat processing by speeding up operations while enhancing yield and improving labor issues.
Other robotic systems
Greater automaton through robotics is going on in other areas of the industry as well, in an effort to boost line speeds and efficiencies. A Japanese equipment company, Mayekawa Manufacturing Co., recently introduced a new robotic deboning machine, the Toridas Mark 2.
Using an X-Ray scanner, the Mark 2 can debone up to 1,000 chicken whole legs per hour, with yields that compare to human deboning. As with the GTRI prototype, this system was designed to take into account variances in each piece of poultry or meat, and uses state-of the-art sensing technology. Mayekawa says the speed of that process is four times greater than deboning by hand.
Meanwhile, Quality Meat Scotland in Europe has utilized robotics to quickly test the quality of meat. That system, with a robotic arm that immediately collects information from ultrasound probes on the surface of the meat, is designed to work at line speed, providing information on quality without sacrificing time.
Lynn Petrak is a contributing editor based in the Chicago area.