A model that could improve robots' ability to grasp objects

When completing missions and tasks in the real-world, robots should ideally be able to effectively grasp objects of various shapes and compositions. So far, however, most robots can only grasp specific types of objects. Researchers at the Chinese Academy of Sciences and Peking University have recently developed a new machine learning model that could help to enhance the grasping abilities of robots. This model, presented in IEEE Transactions on Circuits and Systems for Video Technology, is specifically designed to predict grasps for objects in a robot's surroundings, so that they can devise optimal strategies for grasping these objects. "In real-world applications, such as intelligent manufacturing, human-machine interaction and domestic services, robotic grasping is becoming more and more essential," Junzhi Yu, one of the researchers who carried out the study, told Tech Xplore. "Grasp detection, a critical step of robotic grasping, entails finding the best grasp for a target object.