This year’s International Food Automation Networking Conference (IFAN) at the Georgia Tech Hotel & Conference Center in Atlanta covered a wide range of meat and poultry processing technologies. Scholars and industry experts from all over the world convened to discuss and showcase the newest innovations and ideas for advancing automation in the food industry.

IFAN covered topics including automated robotic deboning for poultry and pork, artificial intelligence (AI) and virtual reality (VR) technologies of a robotic farm assistant for broilers to name a few. Additionally, industry suppliers touted their latest innovations including state-of-the-art robots as well as foreign material detection technologies and AI-based automation.

Presenters at IFAN stressed the importance of both food safety and worker safety, the elimination of human labor, or easing human labor, sustainability and efficiency. While many of the ideas and technology presented at IFAN were still in the research and development phases, they were all informative and showed promise.

Risk, reward and ROI

Stuart Shaw from Australian Meat Processor Corp. (AMPC), the specialist research and development provider for Australian red meat processors of all sizes, gave the keynote address for the first session titled “Advanced Manufacturing R&D Initiatives.” The presentation provided a backdrop to how new technologies can help processors even when current technologies provide adequate margins, and the risks associated with new tech seem to outweigh going with the latest equipment innovations.

AMPC exists to ensure Australian meat processors of all sizes become more efficient, sustainable and advanced in their current technologies and to provide extension services to processors.

“Our job is to find new technologies and to introduce those to the processing community and to help them to invest in these technologies and take on the risk, because there’s no doubt, everything I show you here today has got a risk involved,” Shaw said.

Shaw described the stages of implementation of new technology in “horizons.” For example, Horizon 3 represents the early development of a project. Horizon 2 indicates a technology has moved on to being developed with a solid chance to work in an operating environment in the field. Horizon 2 includes on-site testing and early-stage prototype work.

“Horizon 1 is where we want to be,” Shaw said. “That’s where something that you’ve developed is now starting to take hold, and people are interested, and you’re actually getting systems in plant and tested.”

Much of the technology covered at IFAN falls into this type of tiered system in which it is not necessarily ready to purchase off the shelf for meat and poultry applications currently but is close enough that the industry will likely see it in the near and middle future.

Gaps and opportunities

The first session of presenters at IFAN included industry professionals that shared their perspectives on the gaps and opportunities that exist in the food processing arena.

Juanfra DeVillena, PhD, senior vice president of quality assurance and food safety at Oakwood, Ga.-based Wayne-Sanderson Farms presented his perspective with a wish list in the poultry industry. After covering technologies the poultry processing industry currently has at its disposal, he went on to talk about the things he’d like to see adopted. Scanners for the litter in chicken houses to detect Campylobacter or Salmonella and automated bird hanging were just two of the opportunities he identified. He also mentioned that an in-plant system capable of detecting an undercooked bird immediately out of the oven as another opportunity.

“Cameras or whatever you want to call it right after the ovens, that would tell me that piece right there is undercooked and I’m going to kick it out, eject it,” DeVillena said. “I don’t care what you do with it, but if we can solve that we will be solving the third-largest reason for our recalls.”

Brian McFarlane, Windsor, Colo.-based McFarlane Enterprises USA, and red meat processing industry veteran spoke to the technical challenges facing automation in the red meat processing industry. At the top of the list, as with almost every processing plant regardless of species, were food and worker safety, but other difficulties include size variation of carcasses, facility age and structure, and technology needs, as well as costs/margins and optimizing yields.

Cutting and trimming

Session two at IFAN appealed to meat processors interested in implementing robotics within the processing plant. While incorporating robots into processing lines in the form of split saws and in packaging is not new, the presenters at IFAN spoke about applications for robots to be used to debone and trim.

Kazuhiro Hattori, PhD, division manager of R&D for Mayekawa Mfg. Co., Ibaraki, Japan, with plants and offices across the globe, showcased Mayekawa’s CELLDAS concept. The CELLDAS presentation focused mainly on the fully robotic deboning of pork primals.

CELLDAS is a system consisting of a recognition unit and multiple cells. Each cell contains one or more multiple axis (depending on designated process) robotic arms. Each cell is responsible for a series or range of processing. Flexibility allows increases in capacity by increasing the number of cells depending on the necessary capacity required.

The recognition unit identifies and classifies the primal and communicates the necessary information and work type to the robots. It contains an X-ray, 3D scanner and detects feature points of the primal and processes through AI. Each CELLDAS unit consists of programmable equipment including a 3D scanner, multiple robot arms with tools such as a knife, a chuck, etc., and a conveyor to move raw materials through. Through collaborative development with Chiba University, each cell incorporates collision avoidance technology to ensure robotic arms do not interfere with one another during processing.

Mayekawa sees a future in which large plants incorporate multiple robotic CELLDAS systems where the processing floor operates a line to remove the tail and hipbone, with applications beyond that to include hock leg, shank leg, belly ribs and the forequarter.

“More applications can be realized by changing programs and effect, and this technology was developed for CELLDAS but, we believe that we can adapt such technologies for several applications,” Hattori said.

Ai-Ping-Hu, researcher at the Georgia Tech Research Institute (GTRI), continues to work on automated poultry deboning. He described the difficulties, such as the way humans excel in the deboning process compared to any currently available automation and robotic technologies, but he believes the research on automated deboning is promising, yet still in the future. One promising concept is based on Leonardo DaVinci’s drawing, “Vitruvian Man.” The research works on the problem with the theory that if one knew several key dimensions of a body you could predict height, weight, etc.

Ping-Hu uses “stereo” cameras (two cameras side by side) and gets two images that are slightly different. From there, researchers can reconstruct 3D information to use with a robot. The knife on the robot will cut the bird in a way that yields the most value.

“Like a human, there’s ball-and-socket joint,” Ping-Hu said. “The shoulder joint is essentially several bones meeting at a juncture, and then there are ligaments and tendons holding them together, and what we’re trying to do is just basically separate those tendons, to be able to separate those bones and to be able to remove the wings and also the breast meat from the carcass.”

Colin Usher, senior research scientist at GTRI, showcased a robot for chicken houses that is not harmful to the chickens and has the potential to lessen the need for human labor, as well as provide valuable data. (Source: ©JANONKAS - STOCK.ADOBE.COM)

Chicken house robot

Commercial chicken production houses present a number of challenges to operators. Thousands of chickens require intensive labor daily and turnover can be high. Issues such as equipment assessment, the removal of floor eggs, quality analysis and cross-contamination need constant attention. Senior research scientist, Colin Usher, at GTRI, suggested robotics could hold a place in future chicken house labor.

The robotic farm assistant for commercial chicken farms project began with a feasibility study in 2013. It used manually controlled robots with broiler chickens and concluded that robot interaction with chickens was not detrimental to the chickens’ well-being.

In 2015, Usher and GTRI developed the first-generation robot with an ability to actively sense chickens on the chicken house floor. The next step was developing a problem for the robot to solve, creating a need for the robot in commercial chicken houses.

Commercial broiler flocks require large spaces to breed in with raised nests for hens to lay eggs, but not all eggs end up in nests. The robot was developed further to continuously remove those floor eggs rather than a paid employee walking the floor to remove them.

By 2016-2017, the robot used object detection to identify chicken, eggs, feeders and drinkers, as well as a 3D location map to know where it had been and where it needed to go to continuously, and repeatedly cover the entire house. Also during this time period, a robotic arm with a suction end picker was installed and finely tuned for the actual pick up of eggs.

In 2018-2019 GTRI produced the second-generation robot with the expanded capabilities of smart navigation, safe chicken interaction, search execution and enhanced sensing and egg picking. COVID-19 and high-pathogenic avian influenza (HPAI) caused in-the-field testing to halt and research then went inside.

Today, GTRI continues to refine the third-generation robot and has added an ultrasonic localization system, full battery management with a magnetic, external charging docking station, new motor controls and robot controller from Nvidia. The latest features include temperature, humidity, gas and light level sensing within the house, all adding value to the robot. 3D vision-based sensing also adds value by assessing a given chicken’s physical health relative to its appearance in areas such as gait scores. 

Useful waste

Bluera, Gran Santiago, Región Metropolitana de Santiago, Chile, is an environmental startup dedicated to advancing the circular economy and waste-to-energy conversion through proprietary technologies. The company partners with governments, municipalities and industries to implement sustainable waste management solutions that drive environmental and economic impact.

Jean Benoit, chief innovation officer at Bluera SpA and Christobal Hurtado, business manager, presented the company’s Bluebox concept. The Bluebox converts waste to energy onsite through patent pending anoxic thermo dissociation. The process is a “low pressure hyperfast pyrolysis [rapidly heating biomass to very high temperatures in the absence of oxygen]” facility working in partial vacuum conditions, without the presence of air and oxygen, leading to non fossil fuels production.

The example of a poultry processing configuration would have the Bluebox onsite at the plant. Garbage, tissues, paper, cardboard, polystyrene, bags, plastic films, gloves, boots, masks, greases, etc., would be shredded, while residues, feathers, dead birds, remains of pulp and discarded meat, viscera, blood, bones, sludge, etc., go through the process of high temperature torrefaction (a thermal treatment process that involves heating biomass in an oxygen-deprived environment to temperatures of up to 300°C) yielding water and a dry powder of the biomass.

The materials after those processes enter the anoxic thermo dissociation unit (Bluebox) and come out in the forms of useable steam, heating, cooling and electricity.