LASER: Script Execution by Autonomous Agents for On-demand Traffic Simulation

Hao Gao1, Jingyue Wang1, Wenyang Fang1, Jingwei Xu1, Yunpeng Huang1, Taolue Chen2,
Xiaoxing Ma1
1 State Key Lab of Novel Software Technology, Nanjing University, China
2 School of Computing and Mathematical Sciences, Birkbeck, University of London, UK
Code arXiv

Abstract

Autonomous Driving Systems (ADS) require diverse and safety-critical traffic scenarios for effective training and testing, but the existing data generation methods struggle to provide flexibility and scalability. We propose LASER, a novel framework that leverage large language models (LLMs) to conduct traffic simulations based on natural language inputs. The framework operates in two stages: it first generates scripts from user-provided descriptions and then executes them using autonomous agents in real time. Validated in the CARLA simulator, LASER successfully generates complex, on-demand driving scenarios, significantly improving ADS training and testing data generation.

Framework

To achieve on-demand and interactive traffic simulation, we propose a framework called LASER.
LASER consists of two stages, implemented by two modules respectively, i.e., script writer and LASER-Agent. Unlike the previous learning based methods that conduct generation and simulation simultaneously, our LASER framework first generates scripts that define logic-chained behaviors (LCB) with natural language instructions from the user requirements. It then executes the script by the real-time cooperation of LASER-Agents.

Result:

  1. Select the Scene, and you will see the corresponding script.
  2. Click Run, and you will see the playback of the actual simulated recorded video.

Scene Script


        

        

          
Result Video

          

            
          

          This is the playback of the actual simulated recorded video