摘要
域内异构多智能体的轨迹预测对确保自动驾驶车辆的安全高效运行至关重要。然而,在城市复杂交通路口场景中,多样化的信息交互、行为意图的连续性和时间的一致性为轨迹预测带来了巨大挑战。现有无锚查询生成方法常面临模式崩塌与训练不稳定,且在复杂路口难以保持意图的连续性与时间的一致性。为此,文章提出了一种两阶段异构边增强图注意力网络(HEGANet)用于动态轨迹预测。第一阶段通过时空上下文编码提取智能体动态特征,结合有向边特征的异构图建模智能体间的时空交互,并通过异构图注意力网络和动态特征门控机制优化特征权重,同时利用多尺度动态交互图注意力网络进一步建模智能体之间及其内部的动态交互。第二阶段通过多层感知机(MLP)对第一阶段生成的预测嵌入进行解码,并通过迭代优化生成精细化的多智能体多模态未来轨迹预测结果。在INTERACTION和Argoverse数据集上的实验验证表明,本文所提动态轨迹预测方法在轨迹预测的准确性上显著优于现有方法。
Trajectory prediction for heterogeneous multi-agent systems within local areas is essential for ensuring the safe and efficient operation of autonomous vehicles.However,complex urban traffic intersection scenarios present significant challenges due to diverse information exchanges,as well as the need for continuity and temporal consistency in behavioral intentions.Existing anchor-free query generation methods often suffer from mode collapse and training instability,and struggle to maintain intent continuity and temporal consistency at complex intersections.To address these challenges,this paper proposes a two-stage heterogeneous edge-augmented graph attention network(HEGANet)for dynamic trajectory prediction.The first stage performs spatiotemporal context encoding to extract dynamic agent features.These features are combined with heterogeneous graphs featuring directed edge attributes to model spatiotemporal interactions between agents.A heterogeneous graph attention network and a dynamic feature gating mechanism are introduced to optimize feature weighting,while a multi-scale dynamic interaction graph attention network is utilized to further model both inter-agent and intra-agent dynamic interactions.In the second stage,a multi-layer perceptron(MLP)decodes the predicted embeddings from the first stage,and iterative optimization is applied to refine multi-agent,multi-modal future trajectory predictions.Experimental results on the INTERACTION and Argoverse datasets demonstrate that the proposed model significantly outperforms existing methods in terms of trajectory prediction accuracy.
作者
陈琳
宁念文
田诗涵
周毅
CHEN Lin;NING Nianwen;TIAN Shihan;ZHOU Yi(School of Artificial Intelligence,Henan University,Zhengzhou,Henan 450046,China;Henan International Joint Laboratory of Cooperative Vehicular Networking,Zhengzhou,Henan 450046,China)
出处
《控制与信息技术》
2025年第5期15-23,共9页
Control and Information Technology
基金
国家重点研发计划政府间国际科技创新合作专项(2023YFE0112500)
国家自然科学基金项目(62176088)
河南省高等学校重点科研项目(22A120001)
融合知识图谱的城市时空大模型关键技术与应用(ghfund202407028284)
东濮老区采出水高效智能化水质调控技术研发(231220038300001)。
关键词
动态轨迹预测
多尺度动态交互图注意力网络
图神经网络
异构交互
dynamic trajectory prediction
multi-scale dynamic interaction graph attention network
graph neural network
heterogeneous interaction
