DELTA introduces a novel decomposed planning approach that leverages Large Language Models to make robot task planning more effective and efficient.

• Addresses LLM hallucination issues by decomposing complex long-term tasks into manageable subtasks
• Significantly improves plan feasibility in robotics applications
• Enhances computational efficiency through targeted replanning strategies
• Validates effectiveness through experiments in both simulated and real-world factory environments

This research advances engineering capabilities for industrial automation by enabling robots to handle complex, long-term tasks with greater reliability and contextual awareness, potentially reducing implementation costs and expanding automation possibilities.

DELTA: Decomposed Efficient Long-Term Robot Task Planning using Large Language Models