Researchers have developed an agentic framework to assist blind and low-vision individuals with indoor navigation by parsing floor plans into a structured knowledge base. This system uses a multi-agent module for floor plan analysis and a path planner with a safety evaluator to generate navigation instructions. Tested on the UMBC Math and Psychology building, the framework achieved success rates of up to 92.31% for short routes, significantly outperforming baseline models like Claude 3.7 Sonnet. AI
影响 Provides a scalable, lightweight solution for accessible indoor navigation, potentially improving independence for visually impaired individuals.
排序理由 This is a research paper detailing a novel framework for indoor navigation using LLMs.
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
