Generating and Evolving Reward Functions for Highway Driving with Large Language Models
This addresses the complex, manual process of reward function design in autonomous driving, offering potential productivity gains, though it is incremental as it applies existing LLM capabilities to a specific domain.
The paper tackles the problem of manually crafting reward functions for autonomous driving by introducing a framework that uses Large Language Models to generate and evolve these functions, achieving a 22% higher average success rate compared to expert handcrafted ones in highway driving simulations.
Reinforcement Learning (RL) plays a crucial role in advancing autonomous driving technologies by maximizing reward functions to achieve the optimal policy. However, crafting these reward functions has been a complex, manual process in many practices. To reduce this complexity, we introduce a novel framework that integrates Large Language Models (LLMs) with RL to improve reward function design in autonomous driving. This framework utilizes the coding capabilities of LLMs, proven in other areas, to generate and evolve reward functions for highway scenarios. The framework starts with instructing LLMs to create an initial reward function code based on the driving environment and task descriptions. This code is then refined through iterative cycles involving RL training and LLMs' reflection, which benefits from their ability to review and improve the output. We have also developed a specific prompt template to improve LLMs' understanding of complex driving simulations, ensuring the generation of effective and error-free code. Our experiments in a highway driving simulator across three traffic configurations show that our method surpasses expert handcrafted reward functions, achieving a 22% higher average success rate. This not only indicates safer driving but also suggests significant gains in development productivity.