AI agents trained in simulations that differ from the environments where they are deployed sometimes perform better than agents trained and deployed in the same environment, MIT research shows.

Seoul National University College of Engineering announced that Professor Kyu-Jin Cho's research team from the Department of Mechanical Engineering has developed a "hyperelastic torque-reversal mechanism" capable of delivering rapid and powerful motions using soft, rubber-like materials. Inspired by the flexible joints of the mantis shrimp and fleas, this technology leverages the hyperelastic properties of elastomeric polymers, which stiffen as they compress. Based on this principle, the team designed a simple structure that delivers exceptional speed and force, and demonstrated applications including a soft gripper capable of grasping ping-pong balls and a robot that can move on rough terrains. This pioneering research, which opens new possibilities for soft robotics design and application, has been published in the prestigious international journal Science Robotics.