RADE: Learning Risk-Adjustable Driving Environment via Multi-Agent Conditional Diffusion

TL;DR

This paper introduces RADE, a multi-agent diffusion-based simulation framework that generates realistic, risk-adjustable traffic scenarios for autonomous vehicle testing, improving realism and scalability over traditional adversarial methods.

Contribution

The paper presents RADE, a novel multi-agent diffusion model that conditions traffic scene generation on risk levels, maintaining realism and physical plausibility for AV safety evaluation.

Findings

RADE preserves statistical realism across risk levels

Increases likelihood of safety-critical events with higher risk settings

Demonstrates effectiveness on real-world dataset

Abstract

Generating safety-critical scenarios in high-fidelity simulations offers a promising and cost-effective approach for efficient testing of autonomous vehicles. Existing methods typically rely on manipulating a single vehicle's trajectory through sophisticated designed objectives to induce adversarial interactions, often at the cost of realism and scalability. In this work, we propose the Risk-Adjustable Driving Environment (RADE), a simulation framework that generates statistically realistic and risk-adjustable traffic scenes. Built upon a multi-agent diffusion architecture, RADE jointly models the behavior of all agents in the environment and conditions their trajectories on a surrogate risk measure. Unlike traditional adversarial methods, RADE learns risk-conditioned behaviors directly from data, preserving naturalistic multi-agent interactions with controllable risk levels. To ensure physical plausibility, we incorporate a tokenized dynamics check module that efficiently filters generated trajectories using a motion vocabulary. We validate RADE on the real-world rounD dataset, demonstrating that it preserves statistical realism across varying risk levels and naturally increases the likelihood of safety-critical events as the desired risk level grows up. Our results highlight RADE's potential as a scalable and realistic tool for AV safety evaluation.