ResAD：面向端到端自动驾驶的归一化残差轨迹建模
ResAD: Normalized Residual Trajectory Modeling for End-to-End Autonomous Driving
把轨迹预测从直接回归改为「惯性参考 + 残差」的建模方式，并用逐点归一化重新加权损失，可直接迁移到任何需要长时程、非均匀误差敏感度的轨迹/位姿序列回归任务（如行人轨迹预测、视频中的运动预测）。
原文依据 Instead of predicting the future trajectory directly, our approach reframes the learning task by predicting the residual from a deterministic inertial reference.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论端到端自动驾驶（E2EAD）系统直接从传感器数据学习预测未来轨迹，但其本质上受轨迹数据固有的时空不平衡问题困扰。这种不平衡带来显著的优化负担，使模型学到虚假相关而非稳健的驾驶逻辑，同时还会优先关注不确定的远距离预测，从而损害即时安全性。为解决这些问题，我们提出 ResAD，一种新颖的归一化残差轨迹建模框架。该方法不直接预测未来轨迹，而是通过预测相对于确定性惯性参考的残差来重构学习任务。该惯性参考作为稳定的物理先验，促使模型关注必要的、由上下文驱动的偏离（如交通规则与障碍物），而非简单的模式匹配。为缓解由不确定的长时程带来的优化不平衡，ResAD 进一步对预测残差引入逐点归一化。该做法对目标重新加权，避免远处不确定路点上的大误差主导学习信号。在 NAVSIM v1 与 v2 基准上，ResAD 仅用两步去噪即取得 88.8 PDMS 与 85.5 EPDMS 的最优结果，表明其简化了学习并提升了规划性能。
英文原文摘要 1394 字符
End-to-end autonomous driving (E2EAD) systems, which learn to predict future trajectories directly from sensor data, are fundamentally challenged by the inherent spatio-temporal imbalance of trajectory data. This imbalance creates a significant optimization burden, causing models to learn spurious correlations instead of robust driving logic, while also prioritizing uncertain, distant predictions, thereby compromising immediate safety. To address these issues, we propose ResAD, a novel Normalized Residual Trajectory Modeling framework. Instead of predicting the future trajectory directly, our approach reframes the learning task by predicting the residual from a deterministic inertial reference. This inertial reference serves as a stable physical prior, encouraging the model to focus on necessary, context-driven deviations (e.g., traffic rules and obstacles) rather than simple pattern matching. To mitigate the optimization imbalance caused by uncertain, long-term horizons, ResAD further incorporates Point-wise Normalization of the predicted residual. This re-weights the objective to prevent large errors at distant, uncertain waypoints from dominating the learning signal. On the NAVSIM v1 and v2 benchmarks, ResAD achieves state-of-the-art results of 88.8 PDMS and 85.5 EPDMS with only two denoising steps, showing that it simplifies learning and improves planning performance.