‘Pain-feeling’ robot under work in Germany

© Wolfgang Rattay
German scientist are edging closer to creating a robotic “nervous system” that will imitate a pain response in machines to enable them to quickly react thus avoiding potentially harmful situations that will hinder the robot's performance.

The work on an “artificial robot nervous system to teach robots how to feel pain” was recently unveiled at the IEEE International Conference on Robotics and Automation (ICRA) in Stockholm, by researchers from Leibniz University of Hannover.

Part of the research involved placing a BioTac sensor that was built to mimic a human fingertip on a Kuka German-made industrial robot arm, to analyze how much pain the robot should feel when resistance is applied.

In their research, Johannes Kuehn and Professor Sami Haddadin focus on the formalization of robot pain, based on insights from human pain research. Experiments show that under gentle pressure from human touch the robot gracefully retracts the arm until the contact is over.

Once greater force is applied, the robot pulls back its arm much quicker and at a greater distance. When more pressure is brought to bear, Kuka is designed to go into a “passive behavior” mode to avoid being hurt.

“Pain is a system that protects us,” Johannes Kuehn, a researcher involved in the project told the conference last week. “When we evade the source of pain, it helps us not get hurt.”

Since robatization of workforces is constantly on the rise, and robots and humans share the same workspace where undetected robotic failure can lead to accidents, researchers argue that by protecting robots from damage, their system will be protecting humans as well.

“A robot needs to be able to detect and classify unforeseen physical states and disturbances, rate the potential damage they may cause to it, and initiate appropriate countermeasures, i.e., reflexes,” Kuehn and Haddadin wrote in their ICRA paper.