Research on an Intent-Driven Autonomous Organizing Robot System for Home Environments

With the continuous advancement of robotics technology, household robots have become an integral part of daily life. To enhance household robots' responsiveness to high-level user commands and improve their environmental perception and autonomous decision-making capabilities. We propose an intention-driven task planning framework for household robots. This framework incorporates a semantic knowledge base and a task planner as core modules. Based on the ROS2 software architecture, it enables collaborative interaction between modules, allowing the robot to interpret ambiguous high-level user commands and autonomously complete sub-task planning, execution, and monitoring. The framework integrates a depth camera with the YOLO object detection algorithm, combined with SLAM environment mapping and localization technology, to achieve semantic understanding and task execution in complex domestic settings. Experiments conducted in various indoor settings verify the effectiveness of the framework. The robot autonomously performs a sequence of operations based on highlevel commands, including object recognition, sequence planning for grasping, and categorized placement. The system substantially enhances the autonomous operational capabilities of domestic robots.