The National Science Foundation is providing a research team at the University of Michigan with US $ 2 million so that robots can learn from human partners on construction sites.

Robots are expected to make the global construction industry safer and more attractive to workers and alleviate labor shortages in the United States.

Building has been one of the most dangerous and inefficient human endeavors for decades. It lags far behind other parts of the economy in productivity and struggles to attract workers to jobs that are often perceived as backbreaking work. In collaboration with the University of Florida and Washington State University, researchers from the College of Engineering and the A. Alfred Taubman College of Architecture and Urban Planning want to change that.

Carol Menassa, senior research director on the research team and associate professor of civil and environmental engineering at the College of Engineering, said the use of automation and robotics on construction sites is essential if construction is to capitalize on the productivity gains that have changed other industries, like this one Production.

A UM graduate operates a KUKA robot like the one used in Carol Menassa’s robot apprentice research. The $ 2 million project aims to enable robots to learn from and cooperate with human construction workers. (Photo by Robert Coelius, College of Engineering)

“Building is much more dynamic and unpredictable than an environment like a factory, so we’re working to redefine the balance between humans and robots,” she said. “Humans and robots have to coexist, and that is the premise of what we are doing right now.”

The three-year project will pair people with “interactive robot assistants” who can learn from people by watching and listening – similar to what human apprentices would do. The robots could someday make the construction work less dangerous and stressful for humans and still allow us to be in charge and solve problems.

By the end of the project, the team plans to deploy a machine learning system that enables this learning through natural interaction, as well as a set of freely available educational tools that train human workers to use these systems effectively. The project also includes contacting K-12 schools to raise awareness and interest in the new opportunities a redesigned construction site offers.

In this vision of the future, robots would take on physically demanding tasks such as lifting bricks or moving sheets of drywall, while humans could find the best way to complete a specific task or make adjustments if the built structure had to deviate from the original plan . The research team has already developed an experimental system that allows humans and robots to work together on simple tasks like laying drywall and ceiling tiles.

“In addition to the direct benefits to the construction industry, this research has the potential to have a broader impact on our built environment,” said Arash Adel, co-research director on the project and assistant professor of architecture and urban planning at Taubman University.

“By using the capabilities of the robots, such as high-quality novel architectures that are too expensive or not completely feasible with current building practices. “

The system uses a virtual reality copy of both the construction site and the robot (called a process-level digital twin in real time) through which the human operator interacts with an Oculus-style VR headset and a video game-like clone of. sees the work area.

With a joystick-like controller and a pointer, the person shows the system what needs to be done, for example picking up a sheet of drywall and aligning it with the stud wall. Based on these instructions, the robot develops the most efficient way to accomplish this and develops a sequence of actions called a movement plan. It then demonstrates its plan to the operator on the screen in the virtual copy of the construction site.

At this point, it is up to the human being to decide whether the proposed plan will get the job done. You can either approve the plan as it is, make changes, or request a completely new plan. When the plan is agreed, the human tells the robot to implement it and then watches in near real time how the robot does the work. The machine learning component enables the robot to memorize the corrections made by the human partner so that the next sheet of drywall is brought up faster.

The team works with industry partners including Barton Malow from Michigan. Barton Malow is already using simple robots to mortar and place bricks in the exterior walls of large buildings.

Daniel Stone, Barton Malow’s director of innovation, hopes that increased use of technology could help curb the long-standing labor shortage caused by the aging of baby boomers and the industry’s reputation for being dirty and dangerous.

“A job that in a few years will blow your shoulders lifting blocks from the wall is hard to sell,” he said. “We’re trying to attract a broader spectrum of people to the trade, people who are interested in technology. If you take the load off of employees taking off, we can attract new employees and enable our current employees to extend their careers. “

The construction company Barton Malow is already using simple robots on some construction sites, such as this MULE, which helps workers place retaining wall blocks. (Photo courtesy of Baurobotik)

For the educational component, Menassa’s team is working with construction unions, community colleges, and others to develop training courses that can teach workers how to work with robotic assistants.

“The future of construction in particular is only a win-win-win situation if it is a fruitful human-robot partnership and ongoing collaboration,” said Vineet Kamat, a co-research director on the project and professor of construction – and environmental engineering at the engineering college.

“That is why we are working on adapting current training methods that use a combination of face-to-face training and construction site training with a builder who trains in a practical manner.”

Kamat said the new training could include training in a classroom, on site and in a laboratory where people could interact with robots in a virtual reality environment. The virtual reality experience is key as humans and robots may not always be in the same physical space in the future. By combining on-site and virtual training, employees can learn how to translate their virtual commands into physical actions and how to work with a robotic assistant to improvise when things don’t go as planned.

The research is funded by the National Science Foundation Future of Work at Human-Technology Frontier program. Other lead researchers on the project include Joyce Chai, Professor of Electrical Engineering and Computer Science, and Wes McGee, Associate Professor of Architecture and Materials Performance.

Other researchers on the project include Honglak Lee, Associate Professor of Electrical Engineering and Computer Science; Jessie Yang, Assistant Professor of Business and Industrial Engineering; Curt Wolf, managing director of the Urban Collaboratory in the field of civil and environmental engineering; Denise Simmons from the University of Florida and Olusola Adesope from Washington State University.