RoboNav-Arm: Agentic AI-Driven Navigation and Obstacle Avoidance for Robotic Manipulator in Cluttered Environments

arXiv:2607.09716v1 Announce Type: new Abstract: Robotic manipulators operating in unstructured environments face significant challenges in safely executing goal-directed tasks due to dynamic and unforeseen obstacles, while traditional methods rely on prior knowledge or fixed perception pipelines, limiting adaptability. We propose a framework for safe task execution with effective obstacle avoidance. The environment module performs real-time obstacle detection, 3D localization, and ground surface...