Urban Currents

2026-08-19

7 arXiv categories· 96 journals· 2405 candidates — 10 worth your time· 121 without an open abstract

Headline

An anchor-sparse lane detection model for resource-constrained hardware and extreme weather

Today, together

tag shift
no tag ran above its 30-day average
canon
no foundational work is cited twice today
coupling
no items share references today
institutions
Tongji University on 20 papers in 30 days; Hong Kong Polytechnic University on 19 papers in 30 days; University of Hong Kong on 16 papers in 30 days; Peking University on 16 papers in 30 days

Three of today's papers carry the tag "Land Use and Ecosystem Services," focusing on stakeholder perceptions of tourism and urban growth in the Yucatan Peninsula, green space extraction and thermal comfort simulation using high-resolution remote sensing, and the 40-year relationship between urban heat islands and surface water dynamics in Gauteng Province. Outside of this group, one paper introduces SDLane, a lane detection model that uses sparse anchors and dynamic fusion to match the actual number of lane instances. Another study analyzes popular entrepreneurship as a cultural grammar based on four years of ethnographic fieldwork in the urban peripheries of São Paulo. Finally, a third paper examines how street painters in Bandung, Indonesia, perceive generative artificial intelligence and evaluates the corresponding local government policy responses.

codedatapublished: Transportation Research Record Journal of the Transportation Research Board

Zhiyong Yang, Ziqiang Zhang, Zhou Yu, et al.
Transportation Research Record Journal of the Transportation Research Board

Chongqing Vocational Institute of Engineering

Autonomous Vehicle Technology and Safety · Advanced Neural Network Applications · Automated Road and Building Extraction · Chongqing Vocational Institute of Engineering · Chongqing Normal University

SDLane: Efficient Lane Detection via Sparse Anchors and Dynamic Fusion

The researchers developed SDLane, a lane detection model that reduces the required number of line anchors to a quantity matching the actual lane instances. Tested on four datasets, including a newly created heavy rain dataset called RainyLane, the model achieves an F1-score of 79.76% at a processing speed of 486 frames per second on the CULane test set using only five anchors.

Why it matters — It eliminates the need for computationally heavy feature extraction and fusion mechanisms like Feature Pyramid Networks or attention modules, enabling high-speed lane detection on resource-constrained hardware even in extreme weather conditions.

Transportation Research Record Journal of the Transportation Research Board · doi · code

published: International Journal of Urban and Regional Research

Henrique Costa
International Journal of Urban and Regional Research · was arXiv

Migration, Ethnicity, and Economy · Digital Economy and Work Transformation · Global trade, sustainability, and social impact · São Paulo

POPULAR ENTREPRENEURSHIP: A Grammar of Moral Economy in Contemporary Urban Peripheries

Based on four years of ethnographic fieldwork, street-level commerce observations, entrepreneurship events, and in-depth interviews in São Paulo, this study analyzes how popular entrepreneurship functions as a cultural grammar in urban peripheries. It traces how these entrepreneurial practices are shaped by long-term family trajectories, moral values, and precarious work experiences rather than just recent labor market shifts.

Why it matters — It demonstrates that peripheral entrepreneurship is not merely an imposed neoliberal ideology, but an active, ambivalent strategy for social reproduction where workers negotiate survival, autonomy, and social mobility under highly precarious conditions.

Caveat: The findings are based on qualitative ethnographic data from a single metropolitan periphery, which may limit direct generalizability to urban contexts with different socioeconomic structures.

International Journal of Urban and Regional Research · pdf · doi

published: International Journal of Urban and Regional Research

Ardelia Indriani, Adiwan Fahlan Aritenang
International Journal of Urban and Regional Research · was arXiv

Cultural Industries and Urban Development · Art History and Market Analysis · Artistic and Creative Research

FROM PAVEMENT TO PIXELS: Street Artists’ and Urban Policy Responses to Generative AI in Bandung's Creative Economy

This study examines how street painters on Braga Street in Bandung, Indonesia, perceive generative artificial intelligence, and evaluates the local government's policy response. Using stakeholder interviews and a SWOT analysis, the research documents a divide between artists who view AI as a creative tool and those who perceive it as a threat, while identifying a complete lack of national legal frameworks regulating ethical AI use in Indonesia's creative industries.

Why it matters — It documents the immediate economic and psychological impacts of generative AI on informal, street-level creative economies in the Global South, demonstrating that local artistic and stakeholder networks are currently filling the regulatory void left by national governments.

Caveat: The empirical findings are limited to a specific group of informal street painters in a single Indonesian city, which may not represent the dynamics of formal creative sectors or other geographic regions.

International Journal of Urban and Regional Research · doi

published: Transportation Research Record Journal of the Transportation Research Board

Xiaobo Ma, Hyunsoo Noh, James Tokishi, et al.
Transportation Research Record Journal of the Transportation Research Board · was arXiv

Pima County Health Department

Traffic Prediction and Management Techniques · Pima County Health Department

Network-Level Vehicle Delay Estimation at Heterogeneous Signalized Intersections

The study introduces Gradient Boosting with Balanced Weighting (GBBW), a domain adaptation framework designed to estimate vehicle delays at diverse signalized intersections by reweighting source-domain data using a small labeled subset from a target intersection. The model was evaluated using traffic data from 57 heterogeneous intersections in Pima County, Arizona, outperforming eight standard machine-learning regression models and seven instance-based domain adaptation methods.

Why it matters — It overcomes the poor generalization of traditional machine-learning models across intersections with different geometries, signal timings, and driver behaviors, enabling reliable delay estimation and traffic-signal optimization in new locations without requiring extensive local datasets.

Caveat: The framework's performance was validated using data from a single county, which may not capture the full diversity of driving behaviors and intersection designs found in other regions.

Transportation Research Record Journal of the Transportation Research Board · doi

published: Journal of Housing and the Built Environment

Weichun Kong, Adelle Mansour, Ang Li, et al.
Journal of Housing and the Built Environment

The University of Melbourne

Assistive Technology in Communication and Mobility · Building Energy and Comfort Optimization · Urban Transport and Accessibility · United States · United Kingdom · The University of Melbourne

Impacts of energy efficiency retrofit interventions on the health and wellbeing of older people who rent their homes: a literature review

This literature review synthesizes findings from seven studies across the United States, the United Kingdom, and Spain to evaluate how energy efficiency retrofits affect the health of renters aged 60 and older. The analyzed interventions spanned weatherization programs and window or door replacements, showing modest improvements in mental health but inconsistent impacts on cardiovascular health. Notably, installing cavity wall insulation without proper ventilation was linked to declines in respiratory, mental, and general health.

Why it matters — It establishes that energy efficiency upgrades do not universally benefit vulnerable populations, demonstrating that specific physical interventions like unventilated insulation can actively harm the health of older tenants.

Caveat: The synthesis is based on a very small pool of only seven studies and relies on narrative synthesis due to the high heterogeneity of the underlying data.

Journal of Housing and the Built Environment · pdf · doi

published: Applied Spatial Analysis and Policy

Paulina Vázquez-Martínez, Gabriela Mendoza‐González, Blanca Vázquez, et al.
Applied Spatial Analysis and Policy

Universidad Nacional Autónoma de México

Environmental and Cultural Studies in Latin America and Beyond · Conservation, Biodiversity, and Resource Management · Land Use and Ecosystem Services · Universidad Nacional Autónoma de México · Autonomous University of Yucatán · Universidad Autónoma de Chile

Socioecological Trade-offs in the Yucatan Peninsula: Stakeholder Perceptions of How Urban and Tourism Development Reshape Nature’s Contributions to People

This study applied a multi-criteria analysis to map and evaluate how institutional stakeholders perceive nature's contributions to people (NCP) across 12 distinct ecosystems in the Yucatan Peninsula, including coastal dunes, mangroves, cenotes, and tropical forests. The analysis evaluated stakeholder priorities regarding ecosystem threats, vulnerability, and the trade-offs between regulatory contributions like water purification and material contributions like food production across different states.

Why it matters — It maps how regional urbanization and tourism trajectories create distinct localized threats to ecosystems, demonstrating that current development models systematically prioritize short-term economic gains over critical regulatory functions like extreme weather protection.

Caveat: The findings are based on the subjective perceptions of institutional stakeholders rather than direct biophysical measurements of ecosystem service degradation.

Applied Spatial Analysis and Policy · pdf · doi

published: Computational Urban Science

Yifei Jiang, Yuanyuan Hao, Lina Wang, et al.
Computational Urban Science

Macau University of Science and Technology

Urban Heat Island Mitigation · Urban Green Space and Health · Land Use and Ecosystem Services · Macau University of Science and Technology · Jiangsu University of Technology

Green space pattern and thermal comfort simulation based on high-resolution remote sensing images

The study developed a two-stage green space extraction method using high-resolution remote sensing images, combining a feature-based extraction model with an InternImage-based noise reduction model, and simulated the urban thermal environment using ENVI-met. Applied to Hangzhou, China, from 2010 to 2024, the extraction methods achieved accuracies of 95.15% and 94.28% respectively, revealing that area F experienced the highest green space growth rate of 3.93% alongside the most favorable thermal comfort levels.

Why it matters — It provides a highly accurate, automated workflow for mapping urban vegetation and shadows from high-resolution imagery, establishing a direct empirical link between long-term green space expansion and localized microclimate improvements.

Caveat: While the extraction models demonstrate high technical accuracy, the thermal comfort analysis is based on ENVI-met simulations rather than in-situ microclimate measurements.

Computational Urban Science · pdf · doi

published: Computational Urban Science

Tamiranashe Kuwaza, Paidamwoyo Mhangara, Eskinder Gidey
Computational Urban Science

University of the Witwatersrand

Urban Heat Island Mitigation · Land Use and Ecosystem Services · Plant Water Relations and Carbon Dynamics · geemap · University of the Witwatersrand · Mekelle University

Impacts of land system change− driven urban heat islands on surface water dynamics and moisture conditions in Gauteng Province, South Africa

This study analyzed the 40-year relationship between urban heat islands, vegetation cover, and surface water dynamics in Gauteng Province, South Africa, from 1984 to 2024 using Google Earth Engine. The analysis revealed that the average urban heat island intensity rose from -2.18°C to 0.25°C, while the average frequency of heatwaves increased from one per year (1984-2009) to four per year (2014-2024). Correlation analysis showed a negative relationship between land surface temperature and both the Modified Normalised Difference Water Index (ρ = -0.40) and NDVI (ρ = -0.29), alongside a declining trend in the Water Ratio Index.

Why it matters — It quantifies the direct, long-term connection between urban warming and declining surface moisture in a water-scarce region, providing empirical evidence to support targeted green infrastructure and water resource protection policies.

Caveat: The study relies on satellite-derived vegetation and water indices as proxies for actual surface water volume and soil moisture conditions.

Computational Urban Science · pdf · doi

preprint

Ying Yao, Nan Zhang, Daniel J. Graham
arXiv stat.AP

regression

Quantifying the Causal Operational Determinants of Service Reliability in Urban Rail Transit: Evidence from Panel Double/Debiased Machine Learning

This study analyzed 30 years of panel data from 1994 to 2024 across 46 international metro operators in the CoMET benchmarking database to quantify how operational factors causally affect transit reliability. Using a Double/Debiased Machine Learning (DML) framework adapted for panel data, the researchers controlled for over 90 confounding variables across technical, financial, and macroeconomic domains. The model revealed that a 1% increase in passenger demand intensity increases incident rates by 0.38%, while higher capacity utilization increases incidents by 0.49%; conversely, greater fleet supply adequacy and car-based operational intensity reduce incident rates by 0.52% and 0.80% respectively.

Why it matters — It establishes a causal link between service reliability and the balance of supply and demand, moving beyond simple correlations to prove that service disruptions are driven by how well service provision scales alongside passenger volumes.

arXiv · pdf

preprint

Mattis thor Straten, Yannick Wolker, Steffen Strohm, et al.
arXiv cs.LG

graph neural network · traffic sensor data · street network data

General Semantic Knowledge Infusion for Spatio-Temporal Traffic Forecasting

This study develops a spatio-temporal prediction framework that extracts semantic subgraphs from Wikidata to generate knowledge graph embeddings for traffic sensors. These embeddings, which capture nearby points of interest, administrative hierarchies, and functional roles, are fused with conventional sensor graphs to create semantic adjacency matrices. Evaluated across multiple established Graph Neural Network architectures, the framework demonstrates that incorporating external semantic knowledge improves traffic forecasting accuracy over models relying solely on physical road-network topology.

Why it matters — It establishes that general-purpose, external knowledge graphs can systematically enhance traffic prediction models without requiring custom neural network architectures, proving that semantic context contains predictive information missing from physical street networks.

arXiv · pdf · doi

Also published today

These appeared today in journals we track. Their abstracts are not openly available, so we cannot summarise them.