Urban Currents

Built Environment Typologies and their Associations with Children’s Spatial Behavior. A Methodological Comparison of Hierarchical Cluster, Principal Component, Latent Class and Latent Profile Analysis

This study compares four data reduction methods—hierarchical cluster analysis, principal component analysis, latent profile analysis, and latent class analysis—to construct built environment typologies and test their influence on children's independent mobility along school routes. Using regression models, it evaluates how the choice of method and sampling bias affect the resulting typologies and their associations with children's travel behavior.

Why it matters — It demonstrates that while general trends remain consistent, the specific relationships between built environment types and travel behavior depend heavily on the chosen statistical method. This reveals a methodological vulnerability in how planners use data-driven typologies for resource allocation, showing that robust results require spatially definite sampling units without prior aggregation.

Caveat: The findings are based on a specific use case of children's school routes, where parental attitudes and distance had a much stronger influence on mobility than any of the built environment typologies.

Applied Spatial Analysis and Policy · pdf · doi

A CityGML extension for 3D modeling of pedestrian mobility space

The study introduces the Pedestrian ADE, an Application Domain Extension that extends the CityGML 3.0 standard to model the physical, legal, and regulatory dimensions of pedestrian environments in 3D. The extension adds new feature classes for vertical infrastructure like stairs and elevators, alongside pedestrian-specific attributes for existing classes. The model's utility was validated through a real-world 3D visualization case study in Melbourne, Australia, mapping where different pedestrian groups are permitted, restricted, or physically able to walk.

Why it matters — It provides a standardized, interoperable framework to integrate detailed pedestrian mobility constraints—including vertical transitions and legal access rules—directly into 3D city models, a capability previously missing from standard urban data pipelines.

Caveat: The paper demonstrates the extension's modeling and visualization capabilities in a single city, but does not evaluate its performance in routing algorithms or large-scale computational simulations.

Cities · doi

Aligning transit provision with urban morphology: A diagnostic assessment of Porto and Cagliari

The study introduces a data-efficient framework combining Space Syntax angular segment analysis, 300-meter transit catchments, and population-weighted exposure to measure how well transit stops align with highly integrated street corridors. This framework, featuring a Gap Index and an Equity Ratio, was applied to compare the Mediterranean port cities of Porto and Cagliari. The analysis revealed that Porto has complete transit coverage along its primary structural corridors, whereas Cagliari has a significant deficit, leaving many highly integrated corridors without nearby transit access.

Why it matters — It demonstrates that differences in transit accessibility between morphologically similar cities can stem from routing design rather than population distribution or street layouts, providing planners with a method to detect structural service gaps that standard supply-side metrics miss.

Caveat: The empirical application is limited to a comparative analysis of two specific Mediterranean port cities, which may limit the direct generalizability of the specific routing patterns to other urban typologies.

Cities · pdf · doi

Measuring lockdown travel rebound: Experimental insights from the Hong Kong COVID-19 omicron outbreak

This study analyzed anonymized mobile positioning data from 117,370 residents to measure the mobility impacts of Hong Kong's localized Restriction-testing Declaration (RTD) lockdowns during the Omicron outbreak. Using a Propensity Score Matching and Difference-in-Differences framework, the researchers tracked changes in the distance to maximum reach position (DMRP) over a two-week period. The localized lockdowns triggered an average daily DMRP increase of 1.20 km, showing an inverted-U-shaped travel rebound that lasted up to 11 days when implemented during the rising phase of the epidemic, compared to a shorter 5-day rebound during the peak phase.

Why it matters — It provides the first empirical measurement of how short-term, localized lockdowns trigger unintended travel rebounds once lifted, demonstrating that the timing of a lockdown within an epidemic's cycle dictates the duration and intensity of subsequent resident mobility.

Caveat: The findings are based specifically on public housing estates in Hong Kong, which may limit generalizability to other housing types or lower-density urban contexts.

Cities · doi

Examining dual exposure to traffic congestion and in-vehicle crowding among senior and regular cardholders using smart card data

Using bus smart card data from Beijing, this study measures concurrent exposure to traffic congestion and in-vehicle crowding, inferring passenger home locations to analyze demographic disparities. The researchers applied statistical and random forest models to compare exposure patterns between senior and regular cardholders across urban, suburban, and rural zones. The analysis reveals that while congestion peaks in urban centers and suburban hubs during rush hours, crowding persists all day along suburban-urban corridors, leaving suburban and rural residents with higher crowding exposure.

Why it matters — It demonstrates that senior cardholders face shorter travel durations but experience higher rates of adverse travel conditions than regular cardholders, providing empirical evidence of age-based transit disparities to help planners target localized interventions like congestion relief in urban centers and crowding reduction on suburban routes.

Caveat: The findings rely on smart card data to infer both passenger home locations and exposure levels, which may not capture the subjective physical discomfort or health impacts experienced by individual riders.

Cities · doi

COVID-19 as a spatial shock: Uneven land market dynamics in metropolitan areas

This study analyzed land transaction data across metropolitan areas in the Czech Republic using a Bayesian Vector Autoregression with exogenous variables (BVARX) model. The analysis reveals that the COVID-19 pandemic acted as a demand shock that accelerated suburbanization, with large metropolitan areas experiencing strong, persistent land market pressures that outlasted the pandemic, while mid-sized cities showed much more muted responses.

Why it matters — It demonstrates that pandemic-induced spatial shifts in land markets are not temporary disruptions but lasting structural transformations, and that a city's specific size and regional integration dictate how vulnerable its suburban fringe is to sudden demand shocks.

Caveat: The findings are based on metropolitan areas within a single country, the Czech Republic, which may limit direct generalizability to countries with different land planning systems or suburban dynamics.

Cities · doi

A Predictive–Prescriptive Analytics Framework forRoute Planning via Discrete Time–SpaceGraph and ST-GCN-GRU Model

This study develops a predictive-prescriptive routing framework that combines a spatiotemporal graph convolutional network with gated recurrent units (ST-GCN-GRU) for multihorizon link speed forecasting, embedding these predictions into a discrete time-space network. The framework was validated using a real-world OpenStreetMap subnetwork of Nanjing to test routing decisions under a unified replay protocol, alongside controlled experiments on the Sioux Falls benchmark network.

Why it matters — It establishes a computationally efficient, polynomially solvable method for time-dependent shortest-path routing that accounts for dynamic traffic evolution rather than static costs, demonstrating near-oracle routing performance in realistic settings.

Caveat: The empirical validation is limited to a subnetwork of a single city and a standard synthetic benchmark.

Journal of Transportation Engineering Part A Systems · doi

Bus Lane Research: A Comprehensive Bibliometric and Literature Review

This paper presents a bibliometric and systematic literature review of bus lane research spanning from 1969 to 2024. It maps out publication distributions, key contributors, and prominent themes, categorizing the existing body of work into four core domains: design and planning, operational efficiency evaluation, broader societal and environmental impacts, and integrated operational strategies.

Why it matters — It synthesizes over five decades of fragmented transportation research into a structured taxonomy, establishing a baseline of historical trends and current knowledge boundaries for transit planners and researchers.

Journal of Transportation Engineering Part A Systems · doi

Impact of Perceived Accident Risk on Automated Driving Technology Acceptance: Roles of Functional, Operational, and Human–Machine Collaboration Risks

This study quantified perceived accident risk into functional, human operational, and human-machine collaboration risks, and analyzed their impact on automated driving technology acceptance using structural equation modeling. The analysis was based on survey data collected from 1,642 drivers. The results show that functional risk influences behavioral intention through attitude and subjective norms, while operational and collaboration risks act through perceived behavioral control, with collaboration risk showing a stronger mediating effect.

Why it matters — It establishes that drivers feel a significantly weaker ability to manage and control human-machine collaboration risks compared to purely human operational risks, providing a specific target for interventions designed to increase technology acceptance.

Caveat: The study relies on self-reported survey data to measure perceived risks and behavioral intentions rather than observing actual driver behavior or responses during real-world automated driving failures.

Journal of Transportation Engineering Part A Systems · doi

The Research into the Impact of Anchoring Prestressed Reinforcement on the Stress–Strain State and Crack Resistance of Reinforced Concrete Models of Sleepers

This study evaluated the structural performance of prestressed reinforced concrete railway sleepers by comparing physical and finite-element models of specimens with anchored versus non-anchored reinforcement. Physical models were fabricated and subjected to full-scale loading tests up to failure to measure crack resistance and strength across the under-rail and mid-section areas.

Why it matters — The findings demonstrate that anchoring reinforcement does not increase the strength or crack resistance of sleepers, and actually causes a minor reduction in these properties. This refutes the assumption that anchoring improves structural performance, though it remains useful as a technological measure to protect rebar ends from environmental exposure and leakage currents.

Caveat: The conclusions are based on laboratory physical models and computational finite-element simulations rather than long-term field testing under actual railway operating conditions.

Future Transportation · doi

Robust Visual SLAM with Multi-Level Adaptive Image Enhancement

This paper introduces a visual simultaneous localization and mapping (VSLAM) method that combines dynamic brightness compensation, an adaptive-threshold CLAHE algorithm based on local statistical characteristics, and a YOLOv5 object detection thread integrated into the ORB-SLAM3 framework. Tested on the public EuRoC and TUM datasets, the system reduces the root mean square error of absolute trajectory error by an average of 29.60% compared to standard ORB-SLAM3, and by up to 97.85% on highly dynamic sequences.

Why it matters — It provides a reliable visual localization method for autonomous systems operating in environments with highly variable lighting and moving obstacles, overcoming the failure modes of traditional VSLAM algorithms that rely on static lighting and fixed thresholding.

Caveat: The performance improvements are demonstrated on standard public datasets rather than in real-world, real-time deployment scenarios.

ISPRS International Journal of Geo-Information · pdf · doi

The Rise of the Smart Compound: Privately Governed Urban Intelligence and Its Research Agenda

This paper conceptualizes the 'smart compound', a real estate model where private developers integrate smart city technologies like home automation, app-mediated access control, and centralized resource management into gated residential and mixed-use developments. It analyzes the distinct data governance challenges that arise when urban digital infrastructure is privately owned and operated outside of municipal oversight, and establishes a structured research agenda to guide future studies.

Why it matters — It bridges the previously separate literatures on smart cities and gated communities, offering a framework to analyze how private capital is quietly absorbing the costs, control, and data governance of local urban infrastructure.

Caveat: This is a conceptual and agenda-setting paper rather than an empirical study with a specific localized dataset.

arXiv · pdf