Manufacturing USA update: Institute also focused on—believe it or not—seafood handling
With a single example, Ira Moskowitz makes the case for why the organization he leads may be critical for advancing manufacturing in the United States.
Moskowitz, CEO of the Advanced Robotics for Manufacturing (ARM) Institute since June 2020 after decades in private industry, cites the projects ARM has funded for robotic assembly and installation of wire harnesses, the so-called nervous system of any large-scale electromechanical system. The task is one that is well suited for a robot because it is difficult for humans to accomplish without error and is physically and ergonomically demanding.
Results overall have reached the point where defense contractors involved in the ARM-sponsored projects—along with other manufacturers, universities and non-profits—intend to deploy robots in their plants to connect the countless wires and connectors that make up wire harnesses.
“Once this has matured, it’s going to make a huge impact in every dimension of industry that deals with wire harnesses, which is virtually anything that is a large-scale electromechanical system,” Moskowitz said. “It’s a really great example of why a large ecosystem like this is required to accomplish something that’s hard for individual entities to do on their own.”
Like the wire harness, ARM’s projects generally focus on technology that is far-enough developed that practical applications of their use can result from the project’s activity. Other projects have focused on robotic composites layup, inspection, assembly of garments, packaging, post-processing, and more.
The goal during a project’s duration, typically six months to one year, is to deliver solutions to member organizations, which can then further customize and use in their own manufacturing.
If it is any indicator of importance, of all the projects ARM has funded, the wire harness project has captured significant interest among members, said Suzy Teele, head of marketing and communications.
A pop-up boot camp leads to employment
Those members, 252 at last count, include:
- Manufacturers from large multi-nationals to small companies;
- Established to startup suppliers;
- Systems integrators, implementers and consultants;
- Major universities, community colleges and trade schools;
- Economic development organizations and associations and
- Government organizations and laboratories.
They pay membership fees of $0, for non-profits, up to $100,000 for the opportunity to not only participate in and/or lead ARM’s projects but also have a say in what those projects should be in order to advance the use of robotics in American industry. All members are responsible for contributing cost-share, with minimums ranging from $2,000-5,000 per year.
The Pittsburgh-based, non-profit ARM is one of 14 such institutes, each with a unique technological concentration, that make up Manufacturing USA, a national initiative to accelerate advanced manufacturing. Carnegie Mellon University established ARM in 2017 with an $80 million, seven-year grant from the U.S. Department of Defense. The two entities share space in a building that was once a steel mill on a brownfield and has since been redeveloped as part of Hazelwood Green, an R&D hub, although the institute is independent of the school. ARM has since received additional funding from the DoD and Department of Commerce for COVID-19-related projects.
ARM focuses on robotics in the aerospace, automotive, composites, electronics, food and beverage, logistics, pharmaceuticals and textiles and apparel industries. Of 59 total projects ARM has funded to date, 34 are technology-related, six are COVID-19-related and 19 are workforce development-related.
For workforce projects, ARM focuses on areas such as inspiring the next generation of engineers and inventors, up-skilling an existing workforce to work with robotics, pre-apprenticeship and apprenticeship programs, defining standards around robotics career pathways and providing advanced manufacturing skills to the non-traditional employee, such as veterans, single mothers, homeless people and former prisoners.
For these non-traditional manufacturing employees, ARM is providing funding to the AmSkills Manufacturing Career Discovery Workshop and Bootcamp in Tampa, Fla. AmSkills, the project’s principal investigator, is a three-county group in the Tampa Bay area that uses apprenticeships for workforce training. The one-day workshop and two-week bootcamp are pop-ups located at a local community center in a low-income community. The program centers on industry-identified skills and uses 80 percent hands-on activities and projects to demonstrate ability.
Participants in the free program are guaranteed a job interview. During a bootcamp in February, 12 of 20 participants were hired for entry-level positions.
The hired group included a 59-year-old homeless veteran, two participants with criminal backgrounds and a wife-and-husband pair who both accepted jobs after years of sending unsuccessful applications, according to ARM’s website. After a second bootcamp in July, 12 of 16 participants were invited back by eight companies for a second interview.
Other workforce-related projects, both online and in-person, have provided training to almost 1,600 people nationwide through the first quarter of 2020, according to an ARM spokesperson.
In addition to the bootcamp, ARM has figured out other ways to help employers and workers alike. In the works is an endorsement program to give a stamp of approval to educational programs that employ best practices in training. Yet another program is Internet-based and focuses on career development.
Online resource helps spell out careers
Because robotics career pathways are still undefined, and training and education are so varied among universities, community colleges, robotics vendors, professional organizations and the like, employers, workers and students can easily be confused about what a certain certificate or diploma means.
To lend some order to the process, ARM recently funded a project for mapping training resources for robotics career pathways. The efforts in the American Northwest are being led by Xu Chen, assistant professor of mechanical engineering at the University of Washington. Other teams are focusing on other areas in the United States.
“It’s a big problem for companies, because when they want to hire someone for robotics in manufacturing, they find it difficult to match their specific needs to an applicant’s credentials,” he said.
“There’s a lack of standardization in the skill set for employees to manage robots,” Teele said.
A similar problem exists for an employer who wants to up-skill his workers and needs information on where to send his team for the best possible training, Chen said. The website for robotic training resources and career pathways, once its operational, is meant to give that employer vetted information to help guide his decision.
Chen, who earned master’s and doctoral degrees at the University of California, Berkeley, in the early 2010s, said he turned to U.S. News & World Report’s best colleges database to help him decide which program was the best choice for his education and career goals. He expects ARM’s pilot career pathways website to be another resource for students in robotics for manufacturing.
“If we can accomplish that, that’ll really move the needle forward in terms of automation robotics training,” Moskowitz said.
An inspector that never gets tired
The mechanical engineering professor is also a principal investigator on a technical project through ARM whose goal is the automation of visual inspection during the manufacturing of complex metallic parts. His partners on the project are Alexander Strzelecki and Andreas Andersson, engineers with GKN Aerospace, a multi-technology Tier 1 aerospace supplier.
For a robot to visually inspect the thousands of components present in an airliner engine and collect and process the data from the inspection is a major challenge, Chen said. That’s due to the components’ complex geometries, reflective metallic surfaces and extremely small tolerances.
“Surface imperfections, like small scratches and pitting, can lead to an imbalance in the airflow for the engine as well as future catastrophic failure,” Strzelecki said. “It’s very important to find these problems early in the process.”
The team aims to combine artificial intelligence with industrial robotics to create an inspection robot. Unlike a human inspector, it will never get tired, bored, or have its attention drawn elsewhere.
While the University of Washington/GKN Aerospace project focuses on aerospace applications, the technology could be used in the automotive, agricultural and consumer goods markets.
Teaching a robot to FISH
Industries not traditionally associated with robotics are also receiving ARM’s attention. One such project could result in the reshoring of a task that some may find surprising to have been outsourced in the first place; seafood processing.
“I don’t think many people are aware that most of the seafood caught off the New England shores is shipped off to Asia for processing and then it comes back, as insane as that sounds,” Moskowitz said.
In 2017, the U.S. imported $21.5 billion in seafood, ARM said. Much of the “imported” seafood was caught domestically but sent abroad for processing due to a labor shortage in the U.S. seafood industry.
Also surprising may be the thought that a robot can work on something as soft and supple as a fish and not mangle it beyond practical use—something Moskowitz characterizes as a “significant technical advantage.”
“But the other aspect is that, as you know, there are concerns about robotics and automation and the fear of job replacement,” he said. “In general, my experience is that most of the folks doing robotics and automation are actually trying to advance companies and wind up with more employees.”
If the Collaborative Robotics to Foster Innovation in Seafood Handling (FISH) project is successful in automating the manual early steps in seafood processing, it may bring that job back to domestic companies, which he believes will increase their revenue and recreate jobs that had been outsourced, Moskowitz said.
Northeastern University, the principal investigator for FISH, got a subsequent grant from the National Science Foundation in 2019 for $2.5 million to continue the work with a consortium of seafood processors.
More work ahead for humans and robots
If robots can do so much in automotive and other industrial assembly, where they’ve been widely used for many years, and can potentially do dangerous jobs like processing fish or tedious jobs like inspecting metallic parts, why aren’t they more widely used?
Despite their advantages, Moskowitz said there are major barriers, including:
- Trying to integrate robotics into a process that wasn’t set up for automation.
- Overcoming robot-human communication and co-working safety problems.
- Making it possible for the software that controls the robot to communicate with other software, standards and process flows in a factory.
- Instilling agility and reconfigurability into robots so they can change as existing factory lines are reconfigured and evolve.
- Designing and engineering robots so they’re more capable of understanding and reacting to their environment and interfacing with machine learning and other technology.
“Those are the major areas for improvement that need to be developed and continue to evolve for robots to continue to add more and more value,” Moskowitz said.
If so, he just explained the rough outline for work ARM has ahead if it and its members will continue to advance manufacturing in the United States.