A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrates a significant advancement in robotic manipulation. Researchers have developed a system that allows robots to perform complex, dexterous tasks—like assembling furniture or manipulating small objects—by using a combination of visual perception and sophisticated planning algorithms. The system, which learns from both simulation and real-world data, enables robots to handle objects with irregular shapes and adapt to unforeseen obstacles, moving beyond simple, repetitive motions. This progress addresses a major challenge in robotics, bringing machines closer to performing delicate, human-like tasks in unstructured environments. For the full details, read the complete article at https://technologyreview.com/2024/05/15/robotic-manipulation-advance.