Human-robot interaction can be improved with teamwork, study finds

February 19, 2013

As more U.S. manufacturers embrace the principles of lean enterprise, there is a growing acceptance of the role that automation can have in the industry.

According to Wired, research conducted by the Massachusetts Institute of Technology has shown that while heavy-duty robots are usually confined to a separate part of the manufacturing facility, increased interaction between human and mechanical workers can increase productivity. With 225,000 robots now in active service in the U.S., the study, which concentrated on a practice known as “cross-training,”placed robots in a workplace environment and allowed workers to carry out simple tasks in tandem with the machine.

This technique has been used for many years in the military, with team members encouraged to engage in “reward motivator scenarios” with their teammates, thereby increasing levels of unspoken communication and coordination. While the battlefield and the assembly line may have different expectations of teamwork, the MIT research took the concept and used it to demonstrate how human to robot interaction could be used in simple tasks.

The task involved the human worker placing screws in one of three positions, with the robot then instructed to drill them into place. With the workers told to swap and change the position of the screws, the research team concluded that the robot would need to adjust quickly to a variety of challenges, especially as there were occasions when more than one screw was used.

Thirty-six human volunteers were involved in the study, which was originally conducted in a virtual reality scenario. Fifty percent of them merely encouraged the robot with positive comments, while the others swapped roles with the machine halfway through. The test was then taken into a real-world environment and repeated, with an emphasis on quality management in manufacturing..

The results showed that robot-human sychronicity was 71 percent higher in those that had swapped roles while idle time – the amount of time spent waiting for a co-worker to complete a task – dropped by 41 percent. The study also showed that the robot was able to anticipate what its human partner would do – in other words the machine was able to predict what it should be doing and when.

“People aren’t robots, they don’t do things the same way every single time,” said Julie Shah, lead researcher and an assistant professor of aeronautics and astronautics at MIT. “And so there is a mismatch between the way we program robots to perform tasks in exactly the same way each time and what we need them to do if they are going to work in concert with people.”