DART-LLM is a novel framework that enables multiple robots to understand natural language instructions and execute complex tasks with dependencies.

• Uses Directed Acyclic Graphs (DAGs) to model task dependencies
• Integrates Question-Answer modules to decompose high-level instructions
• Enables robots to work in a coordinated fashion rather than independently
• Demonstrates significant improvements in task completion efficiency

This research advances engineering applications by solving a critical challenge in multi-robot systems: coordinating actions when tasks have complex dependencies and prerequisites. The framework allows industrial robots to work together seamlessly on assembly lines, construction sites, and in other collaborative environments.

DART-LLM: Dependency-Aware Multi-Robot Task Decomposition and Execution using Large Language Models