ADV-0: Closed-Loop Min-Max Adversarial Training for Long-Tail Robustness in Autonomous Driving

• URL: http://arxiv.org/abs/2603.15221v1
• Date: Mon, 16 Mar 2026 12:58:31 GMT
• Title: ADV-0: Closed-Loop Min-Max Adversarial Training for Long-Tail Robustness in Autonomous Driving
• Abstract summary: This paper presents ADV-0, a closed-loop min-max optimization framework that treats the interaction between driving policy (defender) and adversarial agent (attacker) as a zero-sum Markov game.<n>To make this tractable, we cast dynamic adversary evolution as iterative preference learning, efficiently approximating this optimum and offering an algorithm-agnostic solution to the game.<n> Experiments indicate that it effectively exposes diverse safety-critical failures and greatly enhances the generalizability of both learned policies and motion planners against unseen long-tail risks.
• Score: 63.980630608984605
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Deploying autonomous driving systems requires robustness against long-tail scenarios that are rare but safety-critical. While adversarial training offers a promising solution, existing methods typically decouple scenario generation from policy optimization and rely on heuristic surrogates. This leads to objective misalignment and fails to capture the shifting failure modes of evolving policies. This paper presents ADV-0, a closed-loop min-max optimization framework that treats the interaction between driving policy (defender) and adversarial agent (attacker) as a zero-sum Markov game. By aligning the attacker's utility directly with the defender's objective, we reveal the optimal adversary distribution. To make this tractable, we cast dynamic adversary evolution as iterative preference learning, efficiently approximating this optimum and offering an algorithm-agnostic solution to the game. Theoretically, ADV-0 converges to a Nash Equilibrium and maximizes a certified lower bound on real-world performance. Experiments indicate that it effectively exposes diverse safety-critical failures and greatly enhances the generalizability of both learned policies and motion planners against unseen long-tail risks.

Related papers

• World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning [60.88859946006306]
  Adversa World Modeling is a theoretically grounded multi-agent self-play fine-tuning framework.<n>In inner minimization, the planner's predictive world model is converted into a role-conditioned adversary that learns sparse, scene-adaptive attack coalitions.<n> Experiments on the nuPlan and InterPlan benchmarks demonstrate that our method generates transferable adversarial interactions.
  arXiv  Detail & Related papers  (2026-07-12T07:44:16Z)
• From Attacks to Curricula: Learnability-Guided Adversarial Training for Safe Autonomous Driving [56.30087557121323]
  AlignADV is a learnability-guided closed-loop adversarial training framework.<n>It converts adversarial scenarios into resolvable and capability-aligned curricula.<n> Experiments show it reduces training steps by up to 40.6 percent.
  arXiv  Detail & Related papers  (2026-06-12T02:13:55Z)
• EvoDrive: Pareto Evolution for Safety-Critical Autonomous Driving via Self-Improving LLM Agents [59.44069233970625]
  Existing methods usually manage this trade-off with handcrafteds and overlooking underexplored patterns.<n>Here we present EvoDrive, the first automated, LLM-based agentic evolution framework for scenario generation.<n>EvoDrive employs a simulator-grounded actor-critic architecture where a memory-driven actor iteratively proposes improvements to the generators and critics filter out implausible candidates.
  arXiv  Detail & Related papers  (2026-06-02T14:01:23Z)
• Steerable Adversarial Scenario Generation through Test-Time Preference Alignment [58.37104890690234]
  Adversarial scenario generation is a cost-effective approach for safety assessment of autonomous driving systems.<n>We introduce a new framework named textbfSteerable textbfAdversarial scenario textbfGEnerator (SAGE)<n>SAGE enables fine-grained test-time control over the trade-off between adversariality and realism without any retraining.
  arXiv  Detail & Related papers  (2025-09-24T13:27:35Z)
• Robust Policy Switching for Antifragile Reinforcement Learning for UAV Deconfliction in Adversarial Environments [6.956559003734227]
  An unmanned aerial vehicles (UAVs) has been exposed to adversarial attacks that exploit vulnerabilities in reinforcement learning (RL)<n>This paper introduces an antifragile RL framework that enhances adaptability to broader distributional shifts.<n>It achieves superior performance, demonstrating shorter navigation path lengths and a higher rate of conflict-free navigation trajectories.
  arXiv  Detail & Related papers  (2025-06-26T10:06:29Z)
• An Optimistic Algorithm for online CMDPS with Anytime Adversarial Constraints [7.275101606364466]
  Online safe reinforcement learning (RL) plays a key role in dynamic environments, with applications in autonomous driving, robotics, and cybersecurity.<n>The objective is to learn optimal policies that maximize rewards while satisfying safety constraints modeled by constrained Markov decision processes (CMDPs)<n>Existing methods achieve sublinear regret under constraints but often fail in adversarial settings, where constraints are unknown, time-varying, and potentially adversarially designed.<n>We propose the Optimistic Mirror Descent Primal-Dual (OMDPD) algorithm, the first to address online CMDPs with anytime adversarial constraints.
  arXiv  Detail & Related papers  (2025-05-28T00:16:34Z)
• Plan-R1: Safe and Feasible Trajectory Planning as Language Modeling [74.41886258801209]
  We propose a two-stage trajectory planning framework that decouples principle alignment from behavior learning.<n>Plan-R1 significantly improves planning safety and feasibility, achieving state-of-the-art performance.
  arXiv  Detail & Related papers  (2025-05-23T09:22:19Z)
• Efficient Adversarial Training in LLMs with Continuous Attacks [99.5882845458567]
  Large language models (LLMs) are vulnerable to adversarial attacks that can bypass their safety guardrails. We propose a fast adversarial training algorithm (C-AdvUL) composed of two losses. C-AdvIPO is an adversarial variant of IPO that does not require utility data for adversarially robust alignment.
  arXiv  Detail & Related papers  (2024-05-24T14:20:09Z)
• Uniformly Safe RL with Objective Suppression for Multi-Constraint Safety-Critical Applications [73.58451824894568]
  The widely adopted CMDP model constrains the risks in expectation, which makes room for dangerous behaviors in long-tail states. In safety-critical domains, such behaviors could lead to disastrous outcomes. We propose Objective Suppression, a novel method that adaptively suppresses the task reward maximizing objectives according to a safety critic.
  arXiv  Detail & Related papers  (2024-02-23T23:22:06Z)
• Stackelberg Driver Model for Continual Policy Improvement in Scenario-Based Closed-Loop Autonomous Driving [5.765939495779461]
  adversarial generation methods have emerged as a class of efficient approaches to synthesize safety-critical scenarios. We tailor the Stackelberg Driver Model (SDM) to accurately characterize the hierarchical nature of vehicle interaction dynamics. Our algorithm exhibits superior performance compared to several baselines especially in higher dimensional scenarios.
  arXiv  Detail & Related papers  (2023-09-25T15:47:07Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.