SEAL: Towards Safe Autonomous Driving via Skill-Enabled Adversary Learning for Closed-Loop Scenario Generation

Benjamin Stoler, Ingrid Navarro, Jonathan Francis, Jean Oh
9/16/2024
cs.ROcs.AIcs.LG

Official code: cmubig/SEAL (24 stars, MIT, last commit 2025-04-16).

Abstract

Verification and validation of autonomous driving (AD) systems and components is of increasing importance, as such technology increases in real-world prevalence. Safety-critical scenario generation is a key approach to robustify AD policies through closed-loop training. However, existing approaches for scenario generation rely on simplistic objectives, resulting in overly-aggressive or non-reactive adversarial behaviors. To generate diverse adversarial yet realistic scenarios, we propose SEAL, a scenario perturbation approach which leverages learned objective functions and adversarial, human-like skills. SEAL-perturbed scenarios are more realistic than SOTA baselines, leading to improved ego task success across real-world, in-distribution, and out-of-distribution scenarios, of more than 20%. To facilitate future research, we release our code and tools: https://github.com/cmubig/SEAL

View on arXivView PDF

Code Implementations(1)

cmubig/SEALOfficial100%
MIT
241C, C++Sep 15, 20241 years agoMIT

Similar Papers

LLM-attacker: Enhancing Closed-loop Adversarial Scenario Generation for Autonomous Driving with Large Language Models

86% match2025

FREA: Feasibility-Guided Generation of Safety-Critical Scenarios with Reasonable Adversariality

85% match1 repo672024

SaFeR: Safety-Critical Scenario Generation for Autonomous Driving Test via Feasibility-Constrained Token Resampling

85% match2026

From Words to Collisions: LLM-Guided Evaluation and Adversarial Generation of Safety-Critical Driving Scenarios

84% match2 repos312025

ChatScene: Knowledge-Enabled Safety-Critical Scenario Generation for Autonomous Vehicles

84% match1 repo1852024

Stackelberg Driver Model for Continual Policy Improvement in Scenario-Based Closed-Loop Autonomous Driving

83% match1 repo312023

Cite this paper

@article{stoler2024seal,
  title  = {SEAL: Towards Safe Autonomous Driving via Skill-Enabled Adversary Learning for Closed-Loop Scenario Generation},
  author = {Benjamin Stoler and Ingrid Navarro and Jonathan Francis and Jean Oh},
  year   = {2024},
  eprint = {2409.10320},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2409.10320v3}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2409.10320.svg)](https://codeofpaper.com/paper/2409.10320)

Discussion