Researchers use table tennis to understand human-robot dynamics in an agile environment

Researchers point to potential areas where a robot can work closely with human partners to complete tasks.

The introduction of collaborative robots promises to increase productivity and improve safety. These robots are expected to collaborate with human workers, unlike the caged robots typically seen in today’s manufacturing industry.

A team of researchers at Georgia Tech uses table tennis to demonstrate that humans may not always trust a robot’s explanations of its intended actions. They have developed an agile robot called a “cobot” that uses table tennis to demonstrate possible areas where a robot can work closely with human partners to complete tasks.

The robot is a Barrett WAM arm equipped with a camera and a paddle. It was trained using a machine learning process called imitation learning. The team developed a system that gives the robot positive reinforcement for successful volleys and negative reinforcement for unsuccessful flights.

“We have also trained our robot to be a safe table tennis partner” said Matthew Gombolay, assistant professor in the School of Interactive Computing and director of Georgia Tech’s Cognitive Optimization and Relational (CORE) Robotics Laboratory. “We’re leveraging previous table tennis work and ‘learning from demonstration techniques’ where a human can teach a robot a skill, such as hitting a table tennis shot, or simply have a human demonstrate a task to the robot.

The project shows that robots can work in close cooperation with humans physically and socially, which is a significant step forward in collaborative robotics. The development of intelligent systems that cooperate with people has numerous applications, from manufacturing and healthcare to education.

During the demonstrations, the researchers observed in most cases a lack of trust in the human participants based on the explanations provided by the robot, and therefore did not cooperate with it. A possible reason is a lack of trust that a robot might not have the same goals or motivations as a human partner. Participants research they were more likely to trust the robot’s explanations when they felt the robot shared their goals and motivations.

Gombolay emphasizes the importance of developing robots that can effectively communicate with humans in a way that builds trust and confidence. This is important in environments where the consequences of a mistake or miscommunication can be serious, such as in healthcare or emergency situations.

“While the choice to work with a robot is ultimately an objective behavior and may vary depending on the context or riskiness of the interaction, ultimately this factor of trust is a key driver of your decision-making and behavior.” said Gombolay. “In practice, we often find that people design and implement impressive robot solutions, but the robot is not designed to generate enough trust from human end users.”

The researchers suggest that a possible solution could be to design robots that are more transparent in their decision-making processes. By providing human partners with a clear understanding of how the robot arrived at a particular decision or action, it may be possible to build trust in the robot’s capabilities.

“The larger goal of the project is to understand the design of robots for fast-paced, close interaction in production, logistics warehouses, restaurant kitchens and even homes. Of course, we need to know how to design physically secure systems, but we also need to know what users find intuitive and reliable – what makes users feel safe. said Gombolay. “Otherwise, these robots will never make it out of the lab to live side by side with humans. I believe our work answers key questions that help design robots to interact with humans, especially how robots communicate their intentions to their human counterparts. But of course, the research opens up even more exciting possibilities than before.”