LLM-Grounder: Open-Vocabulary 3D Visual Grounding with Large Language Model as an Agent | IEEE Conference Publication | IEEE Xplore