{
  "paper_id": "8adf676b8868b6e73c72536903977595",
  "paper_title": "ELPTNet: An Efficient LiDAR-based 3D Pedestrian Tracking Network for Autonomous Navigation Social Robots",
  "paper_authors": [
    "Jinzheng Guang",
    "Zhenzhong Cao",
    "Yinuo Song",
    "Jingtai Liu"
  ],
  "publication_date": "",
  "publication_venue": "2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)",
  "abstract": "Autonomous navigation social robots need to track pedestrian movements in real-time with high precision to optimize path planning and avoid collisions. However, the main challenge of pedestrian tracking lies in the significant variations in human posture, which differ from rigid-body structures like vehicles. In this paper, we propose an Efficient LiDAR-based 3D Pedestrian Tracking Network (ELPTNet). First, our ELPTNet employs a 3D object detector to extract directional 3D pedestrian bounding boxes from LiDAR point clouds. Then, our ELPTNet employs a Constant Acceleration (CA) model and prediction confidence for target trajectory prediction. During the data association process, it integrates geometric, appearance, and motion features to enhance the robustness and real-time performance of 3D MOT when targets are temporarily occluded. Experimental results demonstrate that our ELPTNet achieves the highest ranking on the large-scale JRDB dataset for the 3D tracking task, outperforming previous state-of-the-art (SOTA) methods with improvements of 8.4% in MOTA and 6.6% in HOTA. Additionally, our ELPTNet attains an inference speed of 61 frames per second (FPS) on a single CPU. Therefore, our method enables accurate and real-time tracking of multiple pedestrians. The code is publicly available at <https://github.com/jinzhengguang/ELPTNet>.",
  "keywords": [
    "LiDAR-based 3D pedestrian tracking",
    "multi-object tracking",
    "constant acceleration model",
    "prediction confidence",
    "data association",
    "autonomous navigation social robots",
    "JRDB dataset"
  ],
  "paper_institutions": [],
  "output_language": "zh",
  "provider": "doubao",
  "model": "ark-code-latest",
  "prompt_template": "simple",
  "extracted_at": "2026-08-16T19:46:33.274522Z",
  "template_tag": "simple",
  "summary": "### 摘要\n本文针对自主导航社交机器人在复杂动态环境中对行人进行实时高精度3D多目标跟踪的问题，提出了一种基于LiDAR的高效3D行人跟踪网络ELPTNet。该网络首先利用3D目标检测器从LiDAR点云中提取带方向的3D行人检测框，然后采用恒定加速度（CA）模型结合预测置信度进行目标轨迹预测，并在数据关联阶段融合几何、外观和运动特征，以增强目标临时遮挡时的鲁棒性和实时性。实验在大规模JRDB数据集上进行，采用HOTA、MOTA、IDF1等多种指标评估，结果显示ELPTNet在3D跟踪任务中排名第一，相比现有最优方法SMAT，MOTA提升8.4%，HOTA提升6.6%，并且在单个CPU上达到61 FPS的推理速度。进一步的分析表明，该方法在不同室内外场景中表现稳定，但在高密度人群和严重遮挡条件下仍存在身份切换问题，未来工作将探索LiDAR与相机融合以提升复杂环境下的跟踪准确性。\n\n### 关键要点\n- ELPTNet是一个在线3D多目标跟踪框架，由3D检测、状态预测、数据关联和状态更新四部分构成，专门针对行人姿态多变和非刚体特性设计。\n- 状态预测采用恒定加速度（CA）模型，公式为 \\(\\boldsymbol{p}_{t}=\\boldsymbol{p}_{t-1}+\\boldsymbol{v}_{t-1}\\cdot\\boldsymbol{\\delta}+\\frac{1}{2}\\boldsymbol{a}_{t-1}\\cdot\\boldsymbol{\\delta}^{2}\\)，相比恒速模型能更好地适应加减速。\n- 提出预测置信度 \\(\\gamma_{t}=\\gamma_{t-1}-\\eta\\gamma_{t-1}\\)，随时间衰减，用于动态调整数据关联的搜索范围，并在检测到目标时通过 \\(\\gamma_{t}^{j}=\\gamma_{t}^{j}+\\eta c_{t}^{j}\\) 恢复。\n- 数据关联综合几何（G）、外观（A）和运动（M）代价，总代价 \\(C^{i},j_{t}=G^{i},j_{t}+A^{i},j_{t}+M^{i},j_{t}\\)，并用预测置信度加权的代价矩阵进行匹配。\n- 检测模块使用预训练的DCCLA网络，在JRDB数据集的3D检测任务上具有SOTA性能，输出包含中心坐标、尺寸、朝向角和置信度的边界框。\n- 实验基于JRDB数据集，包含54个校园场景，其中27个用于测试，采用HOTA、MOTA、IDF1、OSPA等多项指标评估。\n- 在JRDB测试集上，ELPTNet的HOTA达到25.538%，比SMAT高6.608%；MOTA达到32.653%，比SMAT高8.443%，均排名第一。\n- 消融研究显示，新目标数量设为15时HOTA和MOTA最优，最大预测新目标数设为25时HOTA和IDF1最佳，后处理分数阈值0.7时HOTA和IDF1达到峰值。\n- 延迟实验表明，实时在线跟踪（延迟0）时FPS为61.132，延迟1秒时MOTA和MODA显著提升但牺牲实时性，说明实时性与精度需要权衡。\n- 鲁棒性分析中，在场景“meyer-green-2019-03-16_0”获得最佳HOTA（34.224%）和IDF1（40.607%），IDSW仅65；但在高密度室内场景“tressider-2019-04-26_2”中IDSW高达2391，HOTA仅16.755%。\n- 方法的局限性在于密集人群和频繁遮挡下身份切换和轨迹断裂仍较严重，未来将通过融合相机与LiDAR数据进一步提升复杂环境中的跟踪性能。"
}