A domestic research team has enabled robots to perform difficult tasks without the need for expert programming.

The Electronics and Telecommunications Research Institute (ETRI) announced on the 27th that it has developed a technology that enables high-difficulty tasks requiring control of speed, acceleration, and contact force without expert programming training using a robot arm.

It uses a portable teaching device that works with the Wizard app, a robot teaching app. The intuitive user interface of the Wizard app allows the device to demonstrate robot movements.

The researchers said, “Robot tasks that previously required weeks or months of development can now be developed in just a few minutes, even by first-time users.”

He also explained, “It provides the flexibility to re-teach at any time even if the robot’s teaching results are not satisfactory to the user or an error occurs during the process, and it is versatile because it can be applied to industrial robots that only have position control without direct force control by connecting a force/torque sensor to the end tool position.”

ETRI demonstrated the robot's sanding process at a test bed built at Neuromeka Pohang Branch in Heunghae-eup, Pohang-si, Gyeongsangbuk-do. When the user uses a teaching device to specify only four points in space where the robot wants to work, the robot automatically finds the surface to be polished and proceeds with the work.

As a result of repeated test evaluations, the robot was able to successfully perform sanding tasks while maintaining performance at more than 90% of the target value when applying a weak or strong force in the vertical direction of the surface in the range of 10 to 30 N compared to wiping.

According to the research team, the developed robot teaching technology has been successfully applied to tasks other than sanding, such as opening/closing doors and fastening/separating parts.

This was revealed to be the result of the integration of △human-robot interface technology, △composite control technology that allows the robot to move its position while maintaining contact force, and △technology that automatically recognizes the vertical direction of the contact surface in real time based on the sensor's measurement values and controls the posture.

The research team expressed their expectations, saying, “It is expected that this will significantly reduce development time as it can be applied to process facilities such as smart factories in the future.”

ETRI plans to continue expanding the usability and application areas of Wizard based on the results of user research conducted on general users at the Usability Evaluation Research Center of Keimyung University and Georgia Tech in the United States.

He added that in addition to research on intelligent robot tasks, they are also conducting various research in the field of robotics, including autonomous driving and human following of mobile robots, environmental perception of robots, and reinforcement learning.

ETRI Robot IT Convergence Lab Principal Researcher Kang Dong-yeop said, “Currently, research is underway on task teaching for mobile manipulator robots, and in the future, we plan to develop teaching and control technologies for various types of robots, such as dual-arm robots and quadruped robots.”