用逆强化学习发现自动驾驶异常行为，提升安全识别能力。
Anomalous Decision Discovery using Inverse Reinforcement Learning
- 通过逆强化学习从感知数据推断驾驶意图，隐式学习时间信用分配。
- 在1.4万条模拟轨迹上实现0.90 AUC与82.2% F1-score，召回率领先39%。
- 对噪声和未见异常类型有强鲁棒性，适合真实自动驾驶场景部署。
异常检测在自动驾驶中至关重要，能通过感知系统识别可能威胁安全的异常行为。现有方法多依赖预设阈值或监督学习，在面对未知场景、传感器噪声和遮挡时性能下降，且需大量标注数据，限制实际应用。为此，我们提出基于逆强化学习（IRL）的异常检测框架，通过序列感知数据推断潜在驾驶意图，实现稳健识别。具体提出轨迹-奖励引导自适应预训练（TRAP）框架，解决现有方法在噪声鲁棒性和未知场景泛化上的两大缺陷。核心创新在于通过奖励与最坏情况监督隐式学习时间信用分配，并采用可变时域采样预训练以最大化后果发生前的时间，实现早期行为偏离检测。在超过14,000条模拟轨迹上实验表明，该方法达到0.90 AUC和82.2% F1-score，相较同类监督与无监督基线，召回率提升39%，F1-score提升12%。同时在多种噪声类型下表现稳定，对未见异常类型具有良好泛化能力。代码将开源于https://github.com/abastola0/TRAP.git。
原文摘要 · Abstract (English)
Anomaly detection plays a critical role in Autonomous Vehicles (AVs) by identifying unusual behaviors through perception systems that could compromise safety and lead to hazardous situations. Current approaches, which often rely on predefined thresholds or supervised learning paradigms, exhibit reduced efficacy when confronted with unseen scenarios, sensor noise, and occlusions, leading to potential safety-critical failures. Moreover, supervised methods require large annotated datasets, limiting their real-world feasibility. To address these gaps, we propose an anomaly detection framework based on Inverse Reinforcement Learning (IRL) to infer latent driving intentions from sequential perception data, thus enabling robust identification. Specifically, we present Trajectory-Reward Guided Adaptive Pre-training (TRAP), a novel IRL framework for anomaly detection, to address two critical limitations of existing methods: noise robustness and generalization to unseen scenarios. Our core innovation is implicitly learning temporal credit assignments via reward and worst-case supervision. We leverage pre-training with variable-horizon sampling to maximize time-to-consequence, resulting in early detection of behavior deviation. Experiments on 14,000+ simulated trajectories demonstrate state-of-the-art performance, achieving 0.90 AUC and 82.2\% F1-score - outperforming similarly trained supervised and unsupervised baselines by 39\% on Recall and 12\% on F1-score, respectively. Similar performance is achieved while exhibiting robustness to various noise types and generalization to unseen anomaly types. Our code will be available at: https://github.com/abastola0/TRAP.git
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。