arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2304.02733cs.ROcs.LGcs.SYeess.SY
收藏

在基于视觉的端到端驾驶策略中学习稳定性注意力

Learning Stability Attention in Vision-based End-to-end Driving Policies

Tsun-Hsuan Wang, Wei Xiao, Makram Chahine, Alexander Amini, Ramin Hasani, Daniela Rus

AI总结:

本文提出利用控制 Lyapunov 函数(CLFs)为基于视觉的端到端自动驾驶策略赋予稳定性,并引入稳定性注意力(att-CLFs)与不确定性传播技术以应对环境变化、提升学习灵活性,经仿真与实车实验验证了其有效性。

摘要:

现代端到端学习系统能够学习直接从感知中推断控制。然而,由于这些系统常暴露于非结构化、高维且复杂的观测空间(例如从像素输入流中进行自动驾驶),因此很难保证其稳定性和鲁棒性。我们提出利用控制 Lyapunov 函数(CLFs)为基于视觉的端到端策略赋予稳定性属性,并引入 CLFs 中的稳定性注意力(att-CLFs)以应对环境变化并提升学习灵活性。我们还提出了一种与 att-CLFs 紧密集成的不确定性传播技术。我们通过在照片级逼真模拟器和真实全尺寸自动驾驶车辆上与经典 CLFs、模型预测控制及普通端到端学习进行比较,证明了 att-CLFs 的有效性。

英文摘要:

Modern end-to-end learning systems can learn to explicitly infer control from perception. However, it is difficult to guarantee stability and robustness for these systems since they are often exposed to unstructured, high-dimensional, and complex observation spaces (e.g., autonomous driving from a stream of pixel inputs). We propose to leverage control Lyapunov functions (CLFs) to equip end-to-end vision-based policies with stability properties and introduce stability attention in CLFs (att-CLFs) to tackle environmental changes and improve learning flexibility. We also present an uncertainty propagation technique that is tightly integrated into att-CLFs. We demonstrate the effectiveness of att-CLFs via comparison with classical CLFs, model predictive control, and vanilla end-to-end learning in a photo-realistic simulator and on a real full-scale autonomous vehicle.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑