arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2606.07464cs.ROcs.AIcs.CV
收藏

面向长上下文自动驾驶的规划对齐令牌压缩

Planning-aligned Token Compression for Long-Context Autonomous Driving

NVIDIA Research(NVIDIA研究);School of Computing and Data Science, The University of Hong Kong(计算与数据科学学院,香港大学)
Zhixuan Liang, Yuxiao Chen, Yurong You, Peter Karkus, Wenhao Ding, Boyi Li, Alexander Popov, Yan Wang, Maximilian Igl, Yiming Li, Danfei Xu, Nikolai Smolyanskiy… 展开作者

Zhixuan Liang, Yuxiao Chen, Yurong You, Peter Karkus, Wenhao Ding, Boyi Li, Alexander Popov, Yan Wang, Maximilian Igl, Yiming Li, Danfei Xu, Nikolai Smolyanskiy, Boris Ivanovic, Ping Luo, Marco Pavone

AI总结:

提出COMPACT-VA框架,基于条件VQ-VAE将长上下文压缩为有界表示,通过规划对齐实现决策关键信息保留,在动态场景中成功率提升超6%,速度提升3.3倍。

摘要:

整体视觉-动作模型代表了自动驾驶中的一种新兴范式。然而,这种架构在编码用于复杂交互的扩展时间上下文时,会产生迅速超过实时计算预算的令牌序列。虽然线性变换器和外部记忆等方法试图使上下文轻量化,但令牌压缩与架构最为兼容,因为它不需要修改主干网络。然而,现有的压缩采用基于规则的启发式方法(如时间衰减),与规划解耦,存在丢失决策关键信息的风险。我们提出COMPACT-VA,一种基于条件VQ-VAE的规划对齐工作记忆框架,将扩展上下文压缩为有界表示。压缩条件同时基于历史轨迹和学习的规划意图,其中后验编码器在训练期间从未来轨迹中提炼规划意图,而先验编码器学习从压缩观测中预测它。压缩记忆与预测的潜在变量拼接,输入策略进行端到端优化,从而在保留决策关键信息的情况下进行规划。我们在历史上下文对行为正确性(如停车、让行或前行)最关键的高信号动态场景中进行评估,并相应地设计了行为指标。在可比的令牌预算下,我们在成功率上实现了超过6%的提升(68.3%),且各项指标一致提升。消融实验验证了规划对齐耦合的有效性。闭环评估证实,与未压缩处理相比,COMPACT-VA在保持一般驾驶性能的同时实现了3.3倍的速度提升和2.7倍的内存减少。

英文摘要:

Monolithic vision-action models represent an emerging paradigm in autonomous driving. However, this architecture produces token sequences that quickly exceed real-time computational budgets when encoding extended temporal context for complex interactions. While approaches like linear transformers and external memory try to make the context lightweight, token compression is most compatible with the architecture as it requires no backbone modifications. Yet existing compression adopts rule-based heuristics like temporal decay, decoupled from planning, risking loss of decision-critical information. We propose COMPACT-VA, a planning-aligned working memory framework built on conditional VQ-VAE, compressing extended context into bounded representations. Compression is conditioned on both historical trajectory and a learned planning intent that the posterior encoder distills from future trajectories during training, while the prior encoder learns to predict it from compressed observations. The compressed memory, concatenated with the predicted latent, feeds the policy for end-to-end optimization, planning with retained decision-critical information. We evaluate on high-signal dynamic scenarios where historical context is most critical for behavior correctness (e.g., stop, yield, or proceed), and accordingly design behavioral metrics. Under comparable token budgets, we achieve $>$6% improvement (68.3%) on success rates with consistent gains across metrics. Ablations validate planning-aligned coupling effectiveness. Closed-loop evaluation confirms that COMPACT-VA maintained general driving performance with 3.3* speedup and 2.7* memory reduction over uncompressed processing.

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。

↑