2026-05-05

研究雷达

近期值得关注的方向趋势、核心观点与证据汇总。

实际更新时间:2026/05/05 06:45:54
当前焦点方向
轨迹数据与城市交通研究
直接查看

近期研究重心正从单一模态轨迹建模转向多源异构数据的系统级整合与统一表征,方法上强调跨数据集可复现性、物理一致性建模及面向真实政策/工程问题的闭环评估。

Ozone平台提出覆盖硬件到原型的五层统一架构,将NGSIM、highD等四个轨迹数据集映射至规范格式
TrajFlow采用flow matching生成全国尺度伪GPS轨迹,以缓解真实轨迹数据的隐私与获取瓶颈
活跃方向
4
最强信号:轨迹数据与城市交通研究
证据条目
32
主方向覆盖 1 个来源
最新更新

全量最新条目

按入库时间从新到旧查看资料库更新,支持筛选论文、资讯、数据集、项目与工具。

筛选
当前展示 24 条,共 4357
最新更新列表
论文
Transactions in GIS
PublisherJournal
GeoAI
A GCN‐Transformer Based Spatial–Temporal Data Prediction Method for Location‐Fixed Heterogeneous Monitoring Sensor Network: A Case Study of Dam Seepage Safety Monitoring

Transactions in GIS, Volume 30, Issue 3, May 2026.

入库时间:2026/05/05 18:12
作者:Zhongyuan Wang, Jinghai Xu, Jingli Zhang, Xiaoming Hu, Yannan Shi, Qiang Peng, Yuchun Cao, Zhihao Xu
资讯
Cities Today
UrbanPlanning
UrbanComputing
How cities are using data to analyse the impact of mega-events

Cities have long relied on forecasts to justify the cost and complexity of... The post How cities are using data to analyse the impact of mega-events appeared first on Cities Today .

入库时间:2026/05/05 18:12
作者:Jonathan Andrews
数据集
USDOT Open Data
Dataset
OpenData
公路性能监测系统(HPMS)——2011年密西西比州数据

HPMS 汇集了关于公路路网规模、使用情况、状况及性能的数据。该系统由一个具备地理空间功能的数据库构成,用于生成报告并提供数据分析工具。HPMS 数据被美国交通部(U.S. DOT)、行政机构、国会以及交通运输界众多利益相关方所使用。

入库时间:2026/05/05 12:10
作者:USDOT Open Data
论文
Transactions in GIS
PublisherJournal
GeoAI
Enhancing the Understanding of Urban Street Perception With LLMs and Street View Imagery

Transactions in GIS, Volume 30, Issue 3, May 2026.

入库时间:2026/05/05 18:12
作者:Xin Han, Yu Zhu, Lei Wang, Zikun Guo
论文
GIScience & Remote Sensing
PublisherJournal
Temporal behavior of major crops with SWOT Ka-band SAR observations

Accurate monitoring of crop development is critical for understanding production drivers and implementing effective agricultural policies. In this study, we propose a novel application of the surface water and ocean topography (SWOT) Ka-band synthetic aperture radar (SAR) sensor by exploring the temporal evolution of its backscattering signature across various winter and summer crops. The analysis assessed the incidence angle dependency of SWOT backscatter for each crop type, the relationship between backscatter dynamics and soil moisture variation, and the influence of the parcel slope on near-nadir SWOT backscattering. Finally, the study provides a comprehensive evaluation of the temporal backscatter behavior of SWOT at key phenological stages of each crop compared to that obtained from the C-band Sentinel-1 time series. The results revealed a typical SAR incidence-angle dependency in SWOT for each crop type despite having a small range of incidence angles between 0.7° and 4°. With near vertical looking, SWOT backscattering dynamics followed the soil moisture variation regardless of vegetation attenuation. For near-nadir geometric acquisition, steeper slopes generally resulted in lower backscatters, except for wheat, barley, and grassland, which showed increased backscatters for steeper slopes during peak vegetation stages. Concerning SWOT temporal behavior, gramineous crops exhibited similar Ka- and C-band patterns from sowing to postgermination until the heading phase. The temporal profiles of maize and rapeseed in the Ka-band aligned with previously observed X-band behaviors, whereas sunflower showed a Ka-band backscattering pattern comparable to that of the C-band. With its 21-day revisit cycle, the SWOT Ka-band time series demonstrates promising potential for operational crop growth monitoring.

入库时间:2026/05/05 18:12
作者:Henri Bazzi Nicolas Baghdadi Cecile Cazals Saeideh Maleki Mehrez Zribi Frédéric Frappart a UMR TETIS, University of Montpellier, AgroParisTech, INRAE, CIRAD, CNRS, Montpellier, Franceb CS Group France, Toulouse, Francec Université de Toulouse, CESBIO, CNES, CNRS, INRAE, IRD, Toulouse, Franced ISPA, UMR1391, INRAE, Bordeaux Science Agro, Villenave d'Ornon, France
资讯
Cities Today
UrbanPlanning
UrbanComputing
Google Earth adds global cycling data layer

Google Earth has launched a new data layer showing how much of a... The post Google Earth adds global cycling data layer appeared first on Cities Today .

入库时间:2026/05/05 18:12
作者:Jonathan Andrews
数据集
USDOT Open Data
Dataset
OpenData
月度交通量趋势报告——2005年9月

《交通量趋势》月度报告是一份国家级数据报告,为各州所有道路提供经质量控制的车辆行驶里程(VMT)数据。

入库时间:2026/05/05 12:10
作者:USDOT Open Data
论文
GIScience & Remote Sensing
PublisherJournal
Mapping Coastal Redwoods (Sequoia sempervirens) across their natural range: an updateable and field-validated distribution map using Sentinel satellite data and cloud computing

Coast redwood (Sequoia sempervirens) is a uniquely-tall and long-lived tree species that occupies a narrow fog-belt along the Pacific coast of North America. Despite its ecological and conservation significance, existing maps of redwood distribution remain limited in spatial resolution, accuracy, and timeliness. In this study, we present an updatable and field-validated distribution map of S. sempervirens across its entire native range—from southwestern Oregon to central California—developed using freely available Sentinel-2 multispectral and Sentinel-1 SAR data. We compiled a georeferenced canopy classification dataset of 960 points, combining field surveys conducted in October 2024 with field-based ground truth points collected in 2017 from a prior mapping study, and externally sourced field-based redwood presence records. This dataset was used to train machine learning models (Random Forest and Gradient Boosted Trees) within a cloud computing framework to classify redwood presence and absence at 10 m spatial resolution. Binary classification models achieved high predictive performance, with the best model yielding over 88% overall accuracy and an AUC of 0.92 on a 30% hold-out validation set. Ten-fold cross-validation further confirmed model consistency (TPR = 0.89; FPR = 0.17). To assess geographic transferability, spatial block cross-validation yielded a mean overall accuracy of 0.78, demonstrating robust performance across the species' range despite reduced accuracy in under-sampled southern regions. A secondary multi-class model differentiated between redwood-dominated and mixed-conifer forest types, achieving an overall accuracy of 73.82%. Comparison with previous redwood distribution datasets revealed substantial agreement but also significant discrepancies, with the new model suggesting redwood presences in previously unmapped redwood fragments and absences in locations previously mapped as redwood. Validation against field data suggested higher accuracy in the new map. The resulting range-wide redwood map offers a current, field-validated, and updateable platform for conservation planning, habitat monitoring, and ecological research. It also establishes a high-confidence baseline for tracking redwood distribution dynamics under ongoing climate and land-use change.

入库时间:2026/05/05 18:12
作者:Subham Banerjee Colin N. Mast Emily J. Francis a Department of Forest and Rangeland Stewardship, Warner College of Natural Resources, Fort Collins, USA
论文
International Journal of Geographical Information Science
PublisherJournal
A stochastic approach to integerize floating-point estimates in gridded population mapping

Gridded population datasets are increasingly relied upon for spatial planning, resource allocation, and disaster response due to their flexible integration with other spatial data layers. These datasets are typically produced by disaggregating population counts from administrative units into grid cells, yielding non-integer values that preserve overall counts. However, floating-point cell values are often difficult for users to interpret, and standard rounding approaches may introduce aggregation errors at administrative levels that affect planning decisions. Here, we present a stochastic integerisation method that preserves total population and demographic proportions, and compare it with existing approaches. The method separates the value of each cell into integer and decimal parts, and probabilistically allocates the residual based on decimal magnitudes. Applying the method to gridded population data shows that it effectively reduces unrealistic population predictions in uninhabited areas. The results also demonstrate that the new integerisation method can effectively convert floating-point population estimates into integers while preserving both spatial distribution and demographic proportions, such as age-sex structures. These findings highlight the performance of the proposed integerisation method to generate reliable gridded population distribution datasets across diverse contexts that are easier to interpret, particularly for areas with sparse populations or complex geometries of underlying administrative units.

入库时间:2026/05/05 18:12
作者:Wen-Bin Zhang Alessandro Sorichetta Charlie Frye Natalia Tejedor-Garavito Weixuan Fang Duygu Cihan Dorothea Woods Gregory Yetman Jason Hilton Andrew J. Tatem Maksym Bondarenko a WorldPop, School of Geography and Environmental Science, University of Southampton, Southampton, United Kingdomb Dipartimento di Scienze della Terra ‘Ardito Desio’, Università degli Studi di Milano, Milan, Italyc Environmental Systems Research Institute, Inc, Redlands, CA, USAd The Center for Integrated Earth System Information, Columbia Climate School, Columbia University, NY, USAe Department of Social Statistics and Demography, Faculty of Social Sciences, University of Southampton, Southampton, United KingdomWen-Bin Zhang is a Research Fellow at WorldPop, University of Southampton. His research interests include spatial modelling and geospatial big data analysis. He contributed to the conceptualization, methodology, formal analysis, visualization, original draft and review of this article.Alessandro Sorichetta is an Associate Professor in Applied Geospatial Research at the Dipartimento di Scienze della Terra “A. Desio”, Università degli Studi di Milano. His research focuses on spatial modelling, geospatial data integration, and applied Earth observation. He contributed to the conceptualization, methodology design, supervision, and review of this article.Charlie Frye is Chief Cartographer at Esri. His expertise lies in cartographic design, geospatial data visualization, and applied GIS methods. He contributed to the conceptualization, validation of geospatial methods, and review of this article.Natalia Tejedor-Garavito is a Principal Enterprise Fellow at WorldPop, University of Southampton. Her research focuses on geospatial modelling, health geography, and population distribution mapping. She contributed to data curation, analysis, and manuscript review.Weixuan Fang is a Research Assistant at WorldPop, University of Southampton. Her research interests include spatial data analysis, urban studies, and policy applications of geospatial modelling. She contributed to data processing, visualization, and the preparation and review of the manuscript.Duygu Cihan is a Research Fellow at WorldPop, University of Southampton. Her research focuses on population estimation, statistical modelling, and geospatial applications for development. She contributed to the methodology, validation, and review of the manuscript.Dorothea Woods is an Enterprise Fellow at WorldPop, University of Southampton. Her research interests include geospatial analysis, environmental change, and demographic modelling. She contributed to data curation, and review of the article.Gregory Yetman is Associate Director for Geospatial Applications at the Center for International Earth Science Information Network (CIESIN), Columbia University. His expertise includes global geospatial datasets, population distribution modelling, and environmental applications. He contributed to data provision, validation, and manuscript review.Jason Hilton is an Associate Professor in Demography and Social Statistics at the University of Southampton. His research focuses on demographic modelling, population projections, and quantitative social science methods. He contributed to demographic modelling, interpretation, and review of the manuscript.Andrew J. Tatem is Director of WorldPop and Professor of Spatial Demography and Epidemiology at the University of Southampton. His research interests span population distribution modelling, infectious disease dynamics, and applications of spatial demography. He contributed to conceptualization, funding acquisition, project supervision, and review of the article.Maksym Bondarenko is a Principal Enterprise Fellow at WorldPop, University of Southampton. His research focuses on geospatial data processing, machine learning, and population mapping. He contributed to data processing, validation, and review of the manuscript.
论文
arXiv
SpatialIntelligence
Trajectory
MiniVLA-Nav v1:面向语言条件机器人导航的多场景仿真数据集

我们提出 MiniVLA-Nav v1,一个面向语言条件物体接近(Language-Conditioned Object Approach, LCOA)导航任务的仿真数据集:给定一段简短的自然语言指令,NVIDIA Nova Carter 差分驱动机器人需在四个逼真的 Isaac Sim 环境(办公室、医院、完整仓库、含多层货架的仓库)中导航至指定物体,并在其 1 米范围内停止。全部 1,174 个 episode 均配有一条指令,以及同步采集的 640×640 RGB 图像、度量深度图(float32,单位:米)和实例分割掩码;同时提供由基于视觉的比例控制器生成的连续控制信号(线速度 v 和角速度 ω)及 7×7 tokenized 专家动作标签,采样频率为 60 Hz。轨迹多样性通过三类起始距离层级(近:1.5–3.5 米;中:3.5–7.0 米;远:全局人工筛选点位;起始距离与轨迹长度的 Pearson 相关系数 r = 0.94)、12 类物体、18 个训练模板及 12 个分布外(OOD)同义改写模板予以保障。本数据集包含五个评估划分,分别用于评估分布内准确性、模板-改写鲁棒性及分布外物体类别泛化能力。该数据集已公开发布于 https://huggingface.co/datasets/alibustami/miniVLA-Nav

入库时间:2026/05/05 08:11
作者:Ali Al-Bustami, Jaerock Kwon
资讯
ArchDaily
UrbanComputing
68号别墅 / MAS建筑事务所

68号别墅位于越南胡志明市一处新兴城区。项目源于一个年轻家庭希望在城市中安家的需求,同时兼顾私密性、便利性及开放的生活环境。基地面积为500平方米,地处道路交叉口,周边规划有大型公寓楼及公共配套设施。尽管住宅可从多个方向便捷抵达,但如何营造私密性、规避噪声污染并实现内部空间的个性化设计,构成了显著挑战。

入库时间:2026/05/05 12:10
作者:Miwa Negoro
数据集
USDOT Open Data
Dataset
OpenData
非重大安全与安保事件(仅限美国联邦铁路管理局监管的通勤铁路)

非重大安全与安保事件数量由各交通运营机构按交通方式逐月向国家公共交通数据库(National Transit Database, NTD)报送。此类事件包括需灭火处置的交通设施内轻微火灾、未导致伤员转运就医的交通从业人员遭袭事件,以及其他未达到重大事件报告阈值因而不构成重大事件报告要求的安全事件(重大事件清单详见“安全与安保事件”数据集)。本文件所含事件数据涵盖通勤铁路(Commuter Rail, CR)和阿拉斯加铁路(Alaska Railroad, AR)模式,以及以下机构所报的重载铁路(Heavy Rail, HR)与混合铁路(Hybrid Rail, YR)服务:港务局跨哈德逊河运输公司(NTD ID: 20098)、俄勒冈州三县大都会交通区(NTD ID: 00008)、德顿县交通局(NTD ID: 60101)及奥斯汀首都大都会交通局(NTD ID: 60048)。由于上述服务受美国联邦铁路管理局(Federal Railroad Administration, FRA)安全监管,相关机构无需通过NTD向美国联邦公共交通管理局(Federal Transit Administration, FTA)报送安全事件(如碰撞、脱轨等)。但上述实体在自有交通设施内发生的安保事件仍须向NTD报送;为保障其他数据集的完整性,此类安保事件被排除于其他文件之外。本文件为完整起见单独列出该类事件,任何试图全面了解公共交通运营机构所报告可报安保事件范围与规模的用户均应参考本文件。

入库时间:2026/05/05 10:11
作者:USDOT Open Data
论文
arXiv
SpatialIntelligence
Trajectory
GSDrive:基于三维高斯泼溅环境的多模态轨迹探测强化驾驶策略

端到端(E2E)自动驾驶提供了一种将感知输入直接映射为驾驶动作的有前景方法。然而,高昂的标注成本与随时间推移而加剧的数据质量退化,阻碍了其在真实场景中的长期部署。尽管结合模仿学习(IL)与强化学习(RL)是提升策略性能的常用策略,但传统RL训练依赖延迟的、基于事件的奖励——策略仅从碰撞等灾难性结果中学习,易导致过早收敛至次优行为。为克服上述局限,我们提出GSDrive框架,利用三维高斯泼溅(3DGS)实现可微分、基于物理的奖励塑形,以改进E2E驾驶策略。本方法在3DGS仿真器中嵌入基于流匹配的轨迹预测器,支持多模态轨迹探测:对候选轨迹进行前向 rollout 以评估其预期奖励。该设计通过将奖励函数锚定于物理仿真的交互信号,在IL与RL之间建立双向知识交换,从而提供即时、稠密的反馈,而非稀疏的灾难性事件反馈。在重建的nuScenes数据集上开展的闭环实验表明,本方法性能优于现有基于仿真的RL驾驶方法。代码开源地址:https://github.com/ZionGo6/GSDrive。

入库时间:2026/05/05 08:11
作者:Ziang Guo, Chen Min, Xuefeng Zhang
论文
Sustainable Cities and Society
PublisherJournal
基于XGBoost-SHAP模型识别的中国新型生产力提升建设用地利用效率潜力预测

出版日期:2026年5月4日在线发表;来源:《可持续城市与社会》(Sustainable Cities and Society);作者:叶云、刘宇、佟彤、吴群

入库时间:2026/05/05 06:11
作者:Sustainable Cities and Society
论文
Sustainable Cities and Society
PublisherJournal
UrbanComputing
面向能源贫困评估的多尺度城市能源数据集成:基于语义方法的荷兰实证研究

出版日期:2026年5月3日在线发布;来源:《可持续城市与社会》;作者:刘璇、桑德·德梅伊、亚历克斯·唐克斯、杨杜娟

入库时间:2026/05/05 06:11
作者:Sustainable Cities and Society
资讯
ArchDaily
AI
Northview Point 公寓楼 / Brooks Scarpa Huber

Northview 公寓楼摒弃了典型社区中以实墙与围栏为特征的防御式公寓建筑模式,转而采用内凹式中央庭院设计,使之成为社区中的一个标志性场所,通过弱化私密空间来凸显社交空间的价值。经过策略性布置的窗户、有明确功能定位的外部交通流线,以及沿建筑最外缘布置的单元,共同将67套低收入住宅单元朝向彼此在空间上分离、却经由建筑群中央的社交空间实现视觉连通的公共区域。

入库时间:2026/05/05 06:11
作者:Susanna Moreira
数据集
USDOT Open Data
Dataset
OpenData
月度交通量趋势——2004年4月

《交通量趋势》月度报告是一份国家级数据报告,为各州所有道路提供经质量控制的车辆行驶里程(VMT)数据。

入库时间:2026/05/05 02:12
作者:USDOT Open Data
论文
Sustainable Cities and Society
PublisherJournal
UrbanComputing
气候情景下城市尺度气温相关死亡率影响的预测:方法学综述

出版日期:2026年5月4日在线发布;来源:《可持续城市与社会》(Sustainable Cities and Society);作者:Kedi Liu、Ranran Wang、Arnold Tukker、Samir KC、Rutger Hoekstra

入库时间:2026/05/05 06:11
作者:Sustainable Cities and Society
论文
Sustainable Cities and Society
PublisherJournal
UrbanComputing
揭示城市碳排放与城市热岛耦合驱动模式:基于局地气候区(LCZ)的可解释机器学习方法

出版日期:2026年5月3日在线发表;来源:《可持续城市与社会》(Sustainable Cities and Society);作者:刘旺银、张子琦、陈宇轩、王思源、刘婉

入库时间:2026/05/05 06:11
作者:Sustainable Cities and Society
论文
arXiv
Trajectory
Mobility
面向先进空中交通(AAM)运行的监视系统可靠性、鲁棒性与韧性建模

在低空空域中保障先进空中交通(Advanced Air Mobility, AAM)的安全高效运行,需依赖一套可靠、鲁棒且具韧性的监视系统,该系统须能在正常及非标工况下持续完成航空器的探测、识别与跟踪任务。为满足此需求,本研究构建了一个涵盖可靠性(reliability)、鲁棒性(robustness)与韧性(resilience)的综合3R建模框架,面向先进空中交通监视系统(Surveillance for Advanced Air Mobility, SAM),重点支撑多类型传感器网络的最优设计与运行。在正常运行条件下,可靠性模型确定满足监视覆盖与探测要求所需的基准传感器类型、数量及布设位置;针对外部扰动(如恶劣天气或AAM交通需求突增),鲁棒性模型识别为维持系统性能所必需的额外传感器配置;对于因主用传感器失效导致的监视中断,韧性模型则制定备用传感器部署与调度策略,以提供临时监视覆盖、最小化运行中断,并保障AAM作业的安全延续。

入库时间:2026/05/05 00:11
作者:Esrat Farhana Dulia, Caleb Adams, Syed Arbab Mohd Shihab
资讯
ArchDaily
GeoAI
Casa Luce / HK Associates Inc

Casa Luce 是对位于卡塔利娜山麓的一座20世纪中期现代主义住宅的大规模翻新工程。该住宅最初由图森现代主义建筑师 Tom Gist 设计并建造,以其标志性的烧制土坯墙体支撑低矮屋檐为特征。尽管这座建于20世纪60年代的住宅具有鲜明的建筑特色,但仍亟需全面更新。本次改造的目标是重新定义住宅的本质——la essenza。

入库时间:2026/05/05 06:11
作者:Susanna Moreira
数据集
USDOT Open Data
Dataset
OpenData
A&I - 公交/客运承运商信息 - 数据挖掘工具

本网站旨在向客运承运商及公众介绍美国联邦机动车承运商安全管理局(FMCSA)的客运承运商项目及相关法规。

入库时间:2026/05/05 02:12
作者:USDOT Open Data
论文
arXiv
SpatialIntelligence
Trajectory
尺度感知的对抗性分析:面向多尺度复杂系统的生成式人工智能诊断框架

从超音速湍流到宇宙大尺度结构,复杂的物理系统均由连续的多尺度动力学所支配。尽管现代机器学习架构在映射此类系统的高维可观测量方面表现优异,但尚不清楚其是否内化了基本物理定律,抑或仅在离散统计相关性之间进行插值。标准可解释人工智能(XAI)架构——尤其是基于扰动和梯度显著性的方法——依赖于像素级扰动,这会产生非物理的人工伪影,并将输入推离有效的经验分布。为解决此问题,我们提出一种由约束扩散分解(Constrained Diffusion Decomposition, CDD)驱动的诊断框架;CDD 是一种基于扩散的多尺度数据分解算法,支持通过尺度感知的修改实现物理约束下的数据生成与模型评估。我们将该框架应用于去噪扩散概率模型(Denoising Diffusion Probabilistic Model, DDPM),在连续的、基于 CDD 的尺度空间中执行确定性干预。结果表明,在适度的物理扰动下,无约束生成模型表现出局域结构冻结与非线性失稳,而非类似偏微分方程(PDE)的连续响应。该网络无法维持跨尺度连续性,导致生成轨迹在进入未见物理状态时发生偏离。通过合成一系列物理自洽的状态 continuum,这一尺度感知方法构建了一个受控测试环境,用于评估算法脆弱性,并为未来架构尊重自然宇宙中多尺度因果关系提供了严格的物理约束基础。

入库时间:2026/05/05 00:11
作者:Mengke Zhao, Guang-Xing Li, Duo Xu
论文
arXiv
RemoteSensing
EarthObservation
SARU:一种面向遥感图像的阴影感知与去除统一框架及新基准数据集

阴影是遥感影像(RSI)中普遍存在的问题,会降低视觉质量,并严重制约目标检测、语义分割等下游任务的性能。现有大多数方法将阴影检测与阴影去除视为彼此分离、级联执行的任务,导致流程繁琐且易产生误差累积。此外,许多深度学习方法依赖成对的含阴影与无阴影图像进行训练,而此类配对数据在实际应用中往往难以获取。为应对上述挑战,我们提出阴影感知与去除统一框架(SARU),一种结构紧凑的两阶段框架:首先,其双分支检测模块(DBCSF-Net)融合多色彩空间特征与语义特征,生成高保真阴影掩膜,有效区分阴影区域与暗色物体;随后,基于该掩膜,一种新颖的无需训练的物理算法(N²SGSR)利用单幅输入图像中邻近非阴影区域的光照特性,实现光照恢复。为支持严格评估并推动后续研究,我们还构建了两个新基准数据集:遥感影像阴影检测数据集(RSISD)与单图像阴影去除基准(SiSRB)。大量实验表明,SARU 在公开 AISD 数据集及我们新构建的基准上均达到当前最优性能。通过整体整合阴影检测与去除以抑制误差传播,并消除对配对训练数据的依赖,SARU 构建了一个鲁棒且实用的遥感影像分析框架。源代码与数据集已公开发布于:https://github.com/AeroVILab-AHU/SARU-Framework。

入库时间:2026/05/05 00:11
作者:Zi-Yang Bo, Wei Lu, Hongruixuan Chen
1 / 182