AI enables robots to feel human touch without artificial skin – 2024-09-04 09:20:39 – 2024-09-04 09:23:55 – 2024-09-04 09:25:38 – 2024-09-04 09:26:33 – 2024-09-04 09:28:45 – 2024-09-04 09:31:41 – 2024-09-04 09:34:56

A team of German researchers has developed a technological breakthrough that allows robots to have a sense of touch without the need for artificial skin or tactile instrumentation. The system will open the door to a future range of possibilities for physical interaction between humans and robots.

The new approach, which was developed by Maged Iskandar of the Institute of Robotics and Mechatronics at the German Aerospace Center DLR, allows the robot to sense and interpret human touch without the need to cover its surface with expensive biomimetic skins or special sensors.

“The intrinsic sense of touch we propose in this work may serve as a basis for an advanced category of human-robot physical interaction that has not yet been possible, enabling a shift from conventional modalities toward adaptability, flexibility, and intuitive handling,” the authors write.

The details of the study were published in the journal Science Robotics.

Sensors and artificial intelligence

To physically interact with humans, robots must be equipped with sensitive but durable sensors that can detect applied force, which can be expensive and complicated when dealing with large or curved robotic surfaces.

To overcome these challenges, Iskandar’s team used the instrumentation already integrated into the Safe Autonomous Robotic Assistant (SARA) system, a robotic arm with high-resolution ‘force and torque’ sensors in its joints that, in addition to recording the force applied to them, measure position and guide movement.

Thanks to sensors and artificial intelligence, the robot can detect where and in what order it is touched by a human, thereby sensitively perceiving the environment around it and accurately locating the tactile trajectories applied to its surface in time and space.

The researchers combined this capability with multiple learning algorithms to interpret the applied touch and demonstrated that the robot could recognize numbers or letters drawn on its surface using neural networks.

Thus, if a human draws the number six on the robot, the technology is able to interpret that the number drawn is indeed a six.