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

TL;DR

LASER is a framework that uses large language models to generate and execute traffic scenarios from natural language descriptions, enhancing the flexibility and scalability of autonomous driving system testing.

Contribution

LASER introduces a novel two-stage framework leveraging LLMs for on-demand traffic scenario generation and execution in simulation environments.

Findings

Successfully generates complex traffic scenarios in CARLA simulator.

Improves flexibility and scalability of traffic data generation.

Enhances training and testing of autonomous driving systems.

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 frame-work 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.