This research introduces a novel technique (LLM-GP-BT) that combines large language models with genetic programming to automatically generate robust Behavior Trees for robot task planning.

• Integrates LLMs to understand task requirements and generate initial control policies
• Uses genetic programming to optimize and refine the solutions
• Employs Behavior Trees framework for modular, flexible robot control systems
• Reduces need for domain expertise in creating reliable autonomous systems

This innovation addresses a critical engineering challenge: developing reliable control policies for autonomous systems in factories, aerospace, and other domains where precise task execution is essential.

Automatic Robot Task Planning by Integrating Large Language Model with Genetic Programming