SurrealDriver: Designing LLM-powered Generative Driver Agent Framework based on Human Drivers' Driving-thinking Data

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jin, Ye, Yang, Ruoxuan, Yi, Zhijie, Shen, Xiaoxi, Peng, Huiling, Liu, Xiaoan, Qin, Jingli, Li, Jiayang, Xie, Jintao, Gao, Peizhong, Zhou, Guyue, Gong, Jiangtao
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914878354096128
author Jin, Ye
Yang, Ruoxuan
Yi, Zhijie
Shen, Xiaoxi
Peng, Huiling
Liu, Xiaoan
Qin, Jingli
Li, Jiayang
Xie, Jintao
Gao, Peizhong
Zhou, Guyue
Gong, Jiangtao
author_facet Jin, Ye
Yang, Ruoxuan
Yi, Zhijie
Shen, Xiaoxi
Peng, Huiling
Liu, Xiaoan
Qin, Jingli
Li, Jiayang
Xie, Jintao
Gao, Peizhong
Zhou, Guyue
Gong, Jiangtao
contents Leveraging advanced reasoning capabilities and extensive world knowledge of large language models (LLMs) to construct generative agents for solving complex real-world problems is a major trend. However, LLMs inherently lack embodiment as humans, resulting in suboptimal performance in many embodied decision-making tasks. In this paper, we introduce a framework for building human-like generative driving agents using post-driving self-report driving-thinking data from human drivers as both demonstration and feedback. To capture high-quality, natural language data from drivers, we conducted urban driving experiments, recording drivers' verbalized thoughts under various conditions to serve as chain-of-thought prompts and demonstration examples for the LLM-Agent. The framework's effectiveness was evaluated through simulations and human assessments. Results indicate that incorporating expert demonstration data significantly reduced collision rates by 81.04\% and increased human likeness by 50\% compared to a baseline LLM-based agent. Our study provides insights into using natural language-based human demonstration data for embodied tasks. The driving-thinking dataset is available at \url{https://github.com/AIR-DISCOVER/Driving-Thinking-Dataset}.
format Preprint
id arxiv_https___arxiv_org_abs_2309_13193
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle SurrealDriver: Designing LLM-powered Generative Driver Agent Framework based on Human Drivers' Driving-thinking Data
Jin, Ye
Yang, Ruoxuan
Yi, Zhijie
Shen, Xiaoxi
Peng, Huiling
Liu, Xiaoan
Qin, Jingli
Li, Jiayang
Xie, Jintao
Gao, Peizhong
Zhou, Guyue
Gong, Jiangtao
Human-Computer Interaction
H.5.2
Leveraging advanced reasoning capabilities and extensive world knowledge of large language models (LLMs) to construct generative agents for solving complex real-world problems is a major trend. However, LLMs inherently lack embodiment as humans, resulting in suboptimal performance in many embodied decision-making tasks. In this paper, we introduce a framework for building human-like generative driving agents using post-driving self-report driving-thinking data from human drivers as both demonstration and feedback. To capture high-quality, natural language data from drivers, we conducted urban driving experiments, recording drivers' verbalized thoughts under various conditions to serve as chain-of-thought prompts and demonstration examples for the LLM-Agent. The framework's effectiveness was evaluated through simulations and human assessments. Results indicate that incorporating expert demonstration data significantly reduced collision rates by 81.04\% and increased human likeness by 50\% compared to a baseline LLM-based agent. Our study provides insights into using natural language-based human demonstration data for embodied tasks. The driving-thinking dataset is available at \url{https://github.com/AIR-DISCOVER/Driving-Thinking-Dataset}.
title SurrealDriver: Designing LLM-powered Generative Driver Agent Framework based on Human Drivers' Driving-thinking Data
topic Human-Computer Interaction
H.5.2
url https://arxiv.org/abs/2309.13193