AI-Driven Human-Robot Interaction for a 3D-Printed Robotic Arm Using Gemini AI and ROS

摘要

This paper presents a novel AI-driven control system for a 3D-printed robotic arm that leverages natural language processing (NLP) and the Robot Operating System (ROS) to enable intuitive human-robot interaction. The proposed system integrates Gemini AI, a state-of-the-art NLP model, to interpret user commands and translate them into executable ROS commands. These commands are transmitted via a ROSBridge WebSocket server to a controller that drives the robotic arm. A mobile application serves as the primary user interface, ensuring seamless, real-time communication between the operator and the robotic system. Furthermore, the arm’s movements are pre-validated in RViz via simulation to guarantee safety and accuracy. Experimental results demonstrate high command execution accuracy and low latency, thus contributing to scalable and user-friendly robotic control frameworks suitable for industrial automation, healthcare, and educational applications.