根据模型内部信号动态选最优轨迹预测器，提升自动驾驶安全性。
Dynamic Model Selection for Trajectory Prediction via Pairwise Ranking and Meta-Features
- 用模型自身稳定性与不确定性等元特征做专家选择决策
- 在nuPlan-mini上实现2.567米的最终位移误差，比GameFormer低9.5%
- 适合关注自动驾驶可靠性与模型自适应的工程研究者
近期深度轨迹预测模型（如Jiang et al., 2023；Zhou et al., 2022）虽平均表现良好，但在复杂长尾驾驶场景中仍不可靠。这暴露出‘一模型通吃’范式的缺陷，尤其在安全关键的城市环境中，简单物理模型有时反而优于先进网络（Kalman, 1960）。为此，我们提出一种动态多专家门控框架，基于样本级自适应选择物理感知LSTM、Transformer和微调GameFormer中的最优预测器。方法利用模型内部信号（元特征），如稳定性和不确定性（Gal and Ghahramani, 2016），证明其比几何场景描述符更具信息量。据我们所知，这是首个将轨迹专家选择建模为基于内部信号的成对排序问题的工作（Burges et al., 2005），直接优化决策质量，无需事后校准。在nuPlan-mini数据集（Caesar et al., 2021）上评估，1,287个样本下，增强型三专家门控系统实现2.567米的最终位移误差（FDE），较GameFormer的2.835米降低9.5%，达到理论最优性能的57.8%。开环仿真中，经轨迹时序对齐后，左转场景下FDE进一步下降约10%，表明在离线验证与开环评估中均具一致性改进。结果表明，自适应混合系统能显著提升安全关键自动驾驶中的轨迹可靠性，为超越静态单模型范式提供了实用路径。
原文摘要 · Abstract (English)
Recent deep trajectory predictors (e.g., Jiang et al., 2023; Zhou et al., 2022) have achieved strong average accuracy but remain unreliable in complex long-tail driving scenarios. These limitations reveal the weakness of the prevailing "one-model-fits-all" paradigm, particularly in safety-critical urban contexts where simpler physics-based models can occasionally outperform advanced networks (Kalman, 1960). To bridge this gap, we propose a dynamic multi-expert gating framework that adaptively selects the most reliable trajectory predictor among a physics-informed LSTM, a Transformer, and a fine-tuned GameFormer on a per-sample basis. Our method leverages internal model signals (meta-features) such as stability and uncertainty (Gal and Ghahramani, 2016), which we demonstrate to be substantially more informative than geometric scene descriptors. To the best of our knowledge, this is the first work to formulate trajectory expert selection as a pairwise-ranking problem over internal model signals (Burges et al., 2005), directly optimizing decision quality without requiring post-hoc calibration. Evaluated on the nuPlan-mini dataset (Caesar et al., 2021) with 1,287 samples, our LLM-enhanced tri-expert gate achieves a Final Displacement Error (FDE) of 2.567 m, representing a 9.5 percent reduction over GameFormer (2.835 m), and realizes 57.8 percent of the oracle performance bound. In open-loop simulations, after trajectory horizon alignment, the same configuration reduces FDE on left-turn scenarios by approximately 10 percent, demonstrating consistent improvements across both offline validation and open-loop evaluation. These results indicate that adaptive hybrid systems enhance trajectory reliability in safety-critical autonomous driving, providing a practical pathway beyond static single-model paradigms.
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。