A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrates that large language models (LLMs) can significantly accelerate the process of robot motion planning, a traditionally computationally intensive task. The research shows that LLMs can break down complex navigation tasks into smaller, manageable subgoals described in natural language, which a simpler, specialized planner can then execute. This approach, termed ‘LLM-GROP,’ allows robots to navigate using more common-sense, intuitive descriptions of their environment rather than relying solely on complex technical maps and coordinates. In tests, this method proved faster and more effective than traditional planning systems, successfully guiding robots through obstacle-filled offices and new environments. The work suggests a promising path toward more adaptable and efficient robots for warehouses, homes, and other dynamic settings. Read the full article at: https://technologyreview.com/2024/07/12/1094756/llms-robot-navigation-planning/