Accurate pedestrian trajectory prediction is critical for ensuring the safety of autonomous vehicles and advancing higher levels of driving automation. However, the complex interpersonal interactions and highly dynamic trajectory patterns in real-world scenarios pose significant challenges to achieving precise predictions. Recently, Transformers have shown remarkable success in pedestrian trajectory prediction, primarily due to their effective modeling of temporal and spatial dependencies via Multi-Head Self-Attention (MHA) mechanisms. Despite these advancements, existing self-attention methods often rely on Euclidean distance-based metrics and dot-product operations, which are inadequate for capturing interaction-induced trajectory curvatures. To address this limitation, we propose a novel hybrid Transformer architecture, FuseFormer, that incorporates Geodesic Self-Attention (GSA) mechanisms. GSA utilizes geodesic distances to characterize interaction features effectively, complementing MHA, which excels in capturing local features and maintaining temporal correlations. FuseFormer employs a gating network to adaptively combine GSA and MHA embeddings, leveraging their complementary strengths. Additionally, FuseFormer integrates a Transformer-based Neural Ordinary Differential Equation (ODE) decoder to model trajectory temporal dynamics. This design enables the generation of future trajectories that align closely with motion trends while adapting the network depth to input sequence lengths. Experimental results demonstrate that FuseFormer achieves state-of-the-art performance across widely used pedestrian trajectory prediction datasets, including ETH/UCY, SDD, and NBA. These results underscore the model’s effectiveness and generalization capability in capturing complex interaction patterns and handling diverse scenarios.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    FuseFormer: A Manifold Metric Fusing Attention for Pedestrian Trajectory Prediction


    Beteiligte:
    Zou, Yi (Autor:in) / Ko, Kohsin (Autor:in) / Yang, Jian (Autor:in) / Liu, Yingjie (Autor:in) / Li, Ke (Autor:in) / You, Xiong (Autor:in) / Mi, Jinpeng (Autor:in) / Tang, Xuan (Autor:in) / Chen, Mingsong (Autor:in) / Wei, Xian (Autor:in)

    Erschienen in:

    Erscheinungsdatum :

    01.08.2025


    Format / Umfang :

    4669954 byte




    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    SEQUENTIAL PEDESTRIAN TRAJECTORY PREDICTION USING STEP ATTENTION FOR COLLISION AVOIDANCE

    MASOUD NEDA / BANDEGI MAHDI / LULL JOSEPH et al. | Europäisches Patentamt | 2023

    Freier Zugriff

    A Pedestrian Trajectory Prediction Model Based on Interactive Attention Mechanism

    Dou, Wanqing / Lu, Lili / Huang, Yamei | Springer Verlag | 2024


    Pedestrian trajectory prediction method based on generative adversarial network and attention mechanism

    GAO JIANPING / DU GUOGUO / LIU PAN et al. | Europäisches Patentamt | 2023

    Freier Zugriff

    Visual end-to-end trajectory prediction system fusing historical prediction trajectory features

    WANG YAFEI / LIU XUANMING / LI ZEXING et al. | Europäisches Patentamt | 2025

    Freier Zugriff