AAAI 2026 | 时空数据(Spatial-temporal)论文总结[上]（时空预测，轨迹挖掘，自动驾驶等）

时空探索之旅

发布于 2026-03-10 14:45:51

1220

文章被收录于专栏：时空探索之旅时空探索之旅

AAAI 2026将在2026年1月20日到1月27日于新加坡（Singapore）举行。AAAI 2026会议主会共有23, 680篇论文投稿，其中4, 167 篇被接收，接收率为17.6%。

本文总结了2026 AAAI 上有关时空数据（Spatial-Temporal）相关论文。如有疏漏，欢迎大家补充。

时空数据Topic：时空预测，天气预报，城市区域表示，轨迹表示学习，相似度计算，轨迹预测，自动驾驶等。总计45篇，本文涉及24篇。

注：由于论文数目较多，分为上下篇，此为上篇，主要涵盖时空预测，轨迹数据挖掘，部分自动驾驶等的文章。下篇主要涵盖自动驾驶，城市区域表示学习，天气预报等。

1. Spatio-Temporal Hierarchical Causal Models2. Unlocking Dynamic Inter-Client Spatial Dependencies: A Federated Spatio-temporal Graph Learning Method for Traffic Flow Forecasting3. WaveDiST: A Wavelet Diffusion Transformer for Spatio-Temporal Estimation on Unobserved Locations4. Knowledge Graph Guided Heterogeneity-Informed Diffusion Model for Spatio-Temporal Generation5. Inter-Client Dependency Recovery with Hidden Global Components for Federated Traffic Prediction6. HyperD: Hybrid Periodicity Decoupling Framework for Traffic Forecasting7. Generalising Traffic Forecasting to Regions Without Traffic Observations8. UrbanPG: An Efficient Framework with Personalized Context and General Backbone Interaction for Urban Spatio-Temporal Learning9. STRIDE-QA: Visual Question Answering Dataset for Spatiotemporal Reasoning in Urban Driving Scenes10. DarkFarseer: Robust Spatio-Temporal Kriging Under Graph Sparsity and Noise11. Spatial-Temporal Feedback Diffusion Guidance for Controlled Traffic Imputation12. DyC-STG: Dynamic Causal Spatio-Temporal Graph Network for Real-time Data Credibility Analysis in IoT13. UrbanNav: Learning Language-Guided Embodied Urban Navigation from Web-Scale Human Trajectories14. ViTE: Virtual Graph Trajectory Expert Router for Pedestrian Trajectory Prediction15. TrajEvo: Trajectory Prediction Heuristics Design via LLM-driven Evolution16. TrajAgg: Dual-Scale Feature Aggregation with Hybrid Training for Trajectory Similarity Computation in Free Space17. Self-Supervised Cross-City Trajectory Representation Learning Based on Meta-Learning18. Region-Point Joint Representation for Effective Trajectory Similarity Learning19. MovSemCL: Movement-Semantics Contrastive Learning for Trajectory Similarity20. ReflexDiffusion: Reflection-Enhanced Trajectory Planning for High-lateral-acceleration Scenarios in Autonomous Driving21. Measuring What Matters: Scenario-Driven Evaluation for Trajectory Predictors in Autonomous Driving22. Intention-Aware Diffusion Model for Pedestrian Trajectory Prediction23. GeoPTH: A Lightweight Approach to Category-Based Trajectory Retrieval via Geometric Prototype Trajectory Hashing24. GeoGen: A Two-stage Coarse-to-Fine Framework for Fine-grained Synthetic Location-based Social Network Trajectory Generation

1 Spatio-Temporal Hierarchical Causal Models

链接：https://arxiv.org/abs/2511.20558

作者：Xintong Li; Haoran Zhang; Xiao Zhou

关键词：时空层次因果模型

2 Unlocking Dynamic Inter-Client Spatial Dependencies: A Federated Spatio-temporal Graph Learning Method for Traffic Flow Forecasting

链接：https://arxiv.org/abs/2511.10434

作者：Feng Wang; Tianxiang Chen; Shuyue Wei; QianChu; Yi Zhang; Yifan Sun; Zhiming Zheng

关键词：交通预测，联邦时空图

3 WaveDiST: A Wavelet Diffusion Transformer for Spatio-Temporal Estimation on Unobserved Locations

作者：Huiling Qin; Yuanxun Li; Weijia Jia

关键词：未观测位置时空估计，小波，Transformer

4 Knowledge Graph Guided Heterogeneity-Informed Diffusion Model for Spatio-Temporal Generation

作者：Zi'ang Wang; Lei Chen; Yuanchang Jin; Pan Deng; Shuangshuang Pang; Junting Liu; Yu Zhao

关键词：时空生成，知识图谱，扩散模型

5 Inter-Client Dependency Recovery with Hidden Global Components for Federated Traffic Prediction

作者：Hang Zhou; Wentao Yu; Yang Wei; Guangyu Li; Sha Xu; Chen Gong

关键词：联邦交通预测，隐私保护

6 HyperD: Hybrid Periodicity Decoupling Framework for Traffic Forecasting

链接：https://arxiv.org/abs/2511.09275

作者：Minlan Shao; Zijian Zhang; Yili Wang; Yiwei Dai; Xu Shen; Xin Wang

关键词：交通预测，混合周期解耦

7 Generalising Traffic Forecasting to Regions Without Traffic Observations

链接：https://arxiv.org/abs/2508.08947

作者：Xinyu Su; Majid Sarvi; Feng Liu; Egemen Tanin; Jianzhong Qi

关键词：无观测交通预测

8 UrbanPG: An Efficient Framework with Personalized Context and General Backbone Interaction for Urban Spatio-Temporal Learning

作者：Aoyu Liu; Yaying Zhang

关键词：高效时空学习

9 STRIDE-QA: Visual Question Answering Dataset for Spatiotemporal Reasoning in Urban Driving Scenes

链接：https://arxiv.org/abs/2508.10427

作者：Keishi Ishihara; Kento Sasaki; Tsubasa Takahashi; Daiki Shiono; Yu Yamaguchi nan

关键词：时空推理的视觉问答数据集

10 DarkFarseer: Robust Spatio-Temporal Kriging Under Graph Sparsity and Noise

链接：https://arxiv.org/abs/2501.02808

作者：Zhuoxuan Liang; Wei Li; Dalin Zhang; Ziyu Jia; Yidan Chen; Zhihong Wang; Xiangping Zheng; Moustafa Youssef

关键词：克里格，稀疏，噪声

11 Spatial-Temporal Feedback Diffusion Guidance for Controlled Traffic Imputation

作者：Xiaowei Mao; Huihu Ding; Yan Lin; Tingrui Wu; Shengnan Guo; Dazhuo Qiu; Feiling Fang; Jilin Hu; Huaiyu Wan

关键词：交通插补，扩散模型

12 DyC-STG: Dynamic Causal Spatio-Temporal Graph Network for Real-time Data Credibility Analysis in IoT

链接：https://arxiv.org/abs/2509.06483

作者：Guanjie Cheng; Boyi Li; Peihan Wu; Feiyi Chen; Xinkui Zhao; Mengying Zhu; Shuiguang Deng

关键词：因果时空图，IOT

13 UrbanNav: Learning Language-Guided Embodied Urban Navigation from Web-Scale Human Trajectories

作者：Yanghong Mei; Yirong Yang; Longteng Guo; Qunbo Wang; MingMing Yu; Xingjian He; Wenjun Wu; Jing Liu

关键词：城市导航，人类轨迹

14 ViTE: Virtual Graph Trajectory Expert Router for Pedestrian Trajectory Prediction

链接：https://arxiv.org/abs/2511.12214

作者：Ruochen Li; Zhanxing Zhu; Tanqiu Qiao; Hubert P. H. Shum

关键词：行人轨迹预测

15 TrajEvo: Trajectory Prediction Heuristics Design via LLM-driven Evolution

链接：https://www.arxiv.org/abs/2508.05616

作者：Zhikai Zhao; Chuanbo Hua; Federico Berto; Kanghoon Lee; Zihan Ma; Jiachen Li; Jinkyoo Park

关键词：轨迹预测，大模型，OOD

16 TrajAgg: Dual-Scale Feature Aggregation with Hybrid Training for Trajectory Similarity Computation in Free Space

作者：Xiao Zhang; Xingyu Zhao; Yuan Cao; Bin Wang; Guiyuan Jiang; Yanwei Yu

关键词：轨迹相似度计算

17 Self-Supervised Cross-City Trajectory Representation Learning Based on Meta-Learning

作者：Yanwei Yu; Hong Xia; Shaoxuan Gu; Xingyu Zhao; Dongliang Chen; Yuan Cao

关键词：轨迹表示学习，自监督，跨城市

18 Region-Point Joint Representation for Effective Trajectory Similarity Learning

链接：https://www.arxiv.org/abs/2511.13125

作者：Hao Long; Silin Zhou; Lisi Chen; Shuo Shang

关键词：轨迹相似度

19 MovSemCL: Movement-Semantics Contrastive Learning for Trajectory Similarity

链接：https://arxiv.org/abs/2511.12061

作者：Zhichen Lai; Hua Lu; Huan Li; Jialiang Li; Christian S. Jensen

关键词：轨迹相似度，对比学习

20 ReflexDiffusion: Reflection-Enhanced Trajectory Planning for High-lateral-acceleration Scenarios in Autonomous Driving

作者：Xuemei Yao; Xiao Yang; Jianbin Sun; Liuwei XIE; Shao Xuebin; Xiyu Fang; Hang Su; Yang Kewei

关键词：轨迹规划，自动驾驶

21 Measuring What Matters: Scenario-Driven Evaluation for Trajectory Predictors in Autonomous Driving

作者：Longchao Da; David Isele; Hua Wei; Manish Saroya

关键词：轨迹预测，自动驾驶

22 Intention-Aware Diffusion Model for Pedestrian Trajectory Prediction

链接：https://arxiv.org/abs/2508.07146

作者：Yu Liu; Zhijie Liu; Xiao Ren; Youfu Li; He Kong

关键词：行人轨迹预测，扩散模型

23 GeoPTH: A Lightweight Approach to Category-Based Trajectory Retrieval via Geometric Prototype Trajectory Hashing

链接：https://arxiv.org/abs/2511.16258

作者：Yang Xu; Zuliang Yang; Kai Ming Ting

关键词：轨迹检索（相似度），轻量化

24 GeoGen: A Two-stage Coarse-to-Fine Framework for Fine-grained Synthetic Location-based Social Network Trajectory Generation

链接：https://arxiv.org/abs/2510.07735

作者：Rongchao Xu; Kunlin Cai; Lin Jiang; Zhiqing Hong; Yuan Tian; Guang Wang

关键词：轨迹生成