The word “automation” may conjure ideas of jobs being replaced and robots taking over, but manufacturers around the world have had a difficult time getting rid of humans in the assembly line.
Robots can work in conditions humans cannot due to extreme heat or radiation and perform tasks that require heavy lifting or extreme precision. On the other hand, human judgment and flexibility are not traditional functions of a robot.
Many manufacturers, like Tesla, have tried making their factories mostly automated, only to find that humans are necessary in the production line.
“Humans are underrated,” Tesla CEO Elon Musk said in an April 2018 tweet about excessive automation.
Collaboration between robots and humans requires both of them to know exactly what the other is going to do — a problem that currently stands in the way of more effective factories.
A team at Purdue is developing a platform called Ghost to solve this problem.
“Factories currently prefer for humans and robots to work separately because they would need an extremely robust method to sync a human with a robot. That’s not the case for current technology,” said Yuanzhi Cao, a postdoctoral student in the College of Engineering, in a Purdue press release.
Ghost is an augmented reality system that lets humans plan out what they’re going to do in relation to the robot, and then wirelessly send their movements to the robot so they both know what the other is doing, according to an abstract on the research.
With this system, the tasks of both the human and robot can be acted out and edited in space and time with just a headset and hand controllers. Humans can plan out their part of the task through an AR headset and then a “ghost” is created of their actions. With that, the user can plan out the robot’s task alongside the ghost of the human that they just recorded.
“It’s totally codeless. The user programs the robot by demonstrating in an AR environment how the two will work together,” said Karthik Ramani, a mechanical engineering professor.
This could potentially be scaled up to program the tasks of hundreds of robots at once, all without writing any code on the part of the user, according to a press release.
The researchers’ goal for this technology is to make it easy for robots to work alongside workers in joint assembly tasks, with the potential for application in countless other industries in the future.