Urban Currents

Evaluating the Impact of CAV-Dedicated Lanes on Freeway Capacity and Safety: Findings from Simulation Experiments

This study built a modified simulation environment in SUMO to evaluate dedicated lane (DL) policies for connected and automated vehicles (CAVs) on a 102 km bidirectional eight-lane freeway modeled after the Beijing-Xiong'an Freeway. Using LandScan population density data to generate traffic demand, the researchers simulated five experimental scenarios to measure traffic efficiency, capacity, and safety under different CAV penetration rates and lane placements. The results show that a single dedicated lane is sufficient when CAV penetration is under 20%, and that placing the DL on the outermost lane is optimal when direct routes make up less than 60% of total demand, whereas placing it on the innermost lane during emergencies causes congestion that reduces average speeds by 30.77%.

Why it matters — It provides specific, spatial-configuration rules for freeway planners, demonstrating that the optimal placement of automated vehicle lanes depends directly on origin-destination distributions and emergency vulnerability rather than just penetration rates.

Caveat: The findings are derived entirely from simulation experiments modeled on a single, specific 102 km freeway corridor, which may not fully capture the driving behaviors and geometry of other highway networks.

Journal of Transportation Engineering Part A Systems · doi

Assessing urban ecological sensitivity in an eastern Indian industrial urban agglomeration using DEMATEL-VWC model

This study evaluates the urban ecological sensitivity of the Asansol Urban Agglomeration in eastern India using an integrated DEMATEL-VWC framework to analyze 14 ecological variables. The model identifies built-up density, land surface temperature, mining intensity, and industrial concentration as the primary risk factors, classifying 9.82% of the area as extremely sensitive and 32.79% as highly sensitive. Ward-level analysis across 106 wards reveals that 14 wards fell into the extremely sensitive category in 2020, with vulnerability concentrated in Kulti, Jamuria, Raniganj, and Burnpur due to mining, deforestation, and industrial pollution.

Why it matters — The study establishes a dynamic weighting method that captures the cause-effect relationships of environmental stressors, providing a localized diagnostic benchmark for integrating ecological sensitivity into land-use zoning and climate-resilience planning in rapidly industrializing regions.

Caveat: The assessment is highly specific to the industrial and mining context of the Asansol Urban Agglomeration, which may limit the direct transferability of the specific variable weights to non-industrial urban areas.

Discover Cities · pdf · doi

Integrating geo-spatial analysis and micro-level land use typologies for sustainable urban planning in Nigeria

This study integrated micro-level land use typologies with geospatial analytics to evaluate urban structure and zoning conformity across 21 political wards in Ile-Ife, Nigeria. Using high-resolution IKONOS imagery and Object-Based Image Analysis, the researchers classified built-up areas into six categories and calculated spatial metrics including patch density, Shannon's diversity index (which averaged 0.62), and Global Moran's I. The analysis revealed high residential aggregation (above 85%), significant spatial fragmentation and sprawl in peripheral wards, and highly dispersed industrial and recreational zones.

Why it matters — It provides a replicable, high-resolution spatial monitoring workflow for secondary African cities where rapid, uncoordinated expansion outpaces traditional planning systems. By quantifying specific zoning deviations and fragmentation patterns, the method establishes an empirical baseline for local authorities to enforce land-use regulations and target infrastructure interventions.

Caveat: The study relies on high-resolution IKONOS satellite imagery, which may be cost-prohibitive or difficult to acquire regularly for resource-constrained planning agencies in similar developing regions.

Discover Cities · doi

Evaluating stealth bio thermal wall skins for mould mitigation and hygrothermal performance in urban housing

This study evaluated a prefabricated, bio-coated thermal wall skin designed for low-disruption retrofitting using a 1D transient heat and moisture transport model simulated in GNU Octave. The system's performance was tested under real climate conditions across five African cities representing diverse urban climates. The retrofit reduced annual heat loss by 42% to 61% and decreased mould-risk hours (relative humidity above 80%) by up to 55% compared to baseline masonry walls.

Why it matters — It demonstrates that a single, modular exterior intervention can simultaneously mitigate indoor mould risks and improve energy efficiency in resource-constrained, ageing urban housing without requiring invasive structural reconstruction.

Caveat: The findings are based on a simplified 1D simulation model that does not account for airflow dynamics, material ageing, or physical experimental validation.

Discover Cities · doi

From global typologies to place-based indicators: An applied geographical approach to TOD in Ho Chi Minh city

The study presents a four-step geographical framework that uses machine learning and a Random Forest model to translate global Transit-Oriented Development (TOD) typologies into local planning indicators, using the Phuoc Long metro station in Ho Chi Minh City as a case study. The model identified key spatial features to generate specific land use recommendations, including eliminating industrial land from 39.4% to 0%, increasing mixed-use and green spaces to 15% each, and establishing a target population density of 150-190 persons per hectare.

Why it matters — It provides a quantitative, transferable method for planners in rapidly developing cities to convert abstract global TOD benchmarks into concrete, context-specific spatial indicators and policy recommendations.

Caveat: The framework's practical recommendations are demonstrated using only a single metro station case study.

Applied Geography · doi

Dispatch optimization under inventory-driven delays: A case study using Approximate Dynamic Programming

This study models dispatch optimization under warehouse space constraints by introducing inventory-driven delays, where delivery vehicles wait at their destination until storage space clears. Formulated as a discrete-time Markov Decision Process with inventory-dependent service times, the model is solved using Approximate Dynamic Programming and tested on operational data from an Ethiopian brewery across four replenishment strategies.

Why it matters — It demonstrates that allowing vehicles to wait at constrained destinations can recover one full truckload of unmet demand for every ten cumulative days of delay, proving that deliberate delivery delays can act as a strategic buffer in capacity-constrained logistics systems.

Caveat: The model is evaluated using a single case study of an Ethiopian brewery and assumes static order generation and parameter certainty.

Case Studies on Transport Policy · doi

A weighted KSA framework for measuring ship navigator competence in loss-of-ship-control scenarios

This study develops and validates a systematic Knowledge, Skills, and Attitude (KSA) framework to evaluate navigator performance during simulator-based loss-of-ship-control scenarios. Content validation was conducted with 65 experienced maritime professionals, and an Analytic Hierarchy Process with 27 participants established weighted priorities across 11 sub-criteria and 60 competencies. The resulting weights allocated 40% of competency priority to Knowledge, 33% to Skills, and 27% to Attitude, with diagnostic skills, rules knowledge, and equipment knowledge ranking highest.

Why it matters — It provides a weighted, scenario-specific evaluation framework that reduces subjectivity and inconsistency in maritime simulator assessments, establishing a structured method to identify competence gaps during high-risk, loss-of-control events.

Caveat: The framework's weights and priorities rely on the subjective consensus of a relatively small panel of 27 maritime professionals for the Analytic Hierarchy Process.

Case Studies on Transport Policy · doi

Reassessing travel time perception through ICT use: Evidence from commuter behaviour in the Madrid Region

A survey of 923 commuters in the Madrid Region analyzed how engaging in ICT-related activities (categorized into entertainment, social interactions, and productivity) affects the difference between actual and perceived travel times. Using generalized ordinal logistic regression and CHAID decision tree analysis, the study found that ICT-related behaviors significantly reduce perceived travel time compared to non-digital activities, with public transport users experiencing the greatest reductions. The effect was weaker for older commuters and statistically insignificant for medium-to-high income individuals, while gender and tech-savviness showed no influence.

Why it matters — The study demonstrates that digital tool usage directly mitigates the subjective burden of travel time, suggesting that transit agencies can increase rider satisfaction and encourage modal shifts by prioritizing onboard amenities like Wi-Fi and power outlets rather than focusing solely on physical speed improvements.

Caveat: The findings rely on self-reported estimates of actual and perceived travel times from a single metropolitan region, which may introduce recall bias.

Case Studies on Transport Policy · doi

An interpretable machine-learning mode choice model incorporating bicycle route quality

An eXtreme Gradient Boosting (XGBoost) mode choice model was developed by integrating bicycle route characteristics generated from a multicriteria routing model. Both models were trained on revealed-preference data from GPS-tracking-based travel diaries to capture how different types of bicycle infrastructure influence travel decisions, with model plausibility verified using SHAP interpretability analysis and simulated infrastructure interventions.

Why it matters — The model establishes a direct, data-driven link between specific bicycle infrastructure quality and overall mode choice, allowing planners to simulate how physical street interventions will shift commuter behavior toward active transportation.

Caveat: The study relies on GPS-tracking travel diaries from a single unspecified region, which may limit how well the routing preferences generalize to other urban contexts with different cycling cultures.

Journal of Cycling and Micromobility Research · doi

Beyond station: Rethinking built environment effects on metro-integrated electric bike-sharing from an origin–destination perspective

This study analyzes how built environment characteristics at both metro stations and non-station trip ends, alongside travel distance, influence metro-integrated electric bike-sharing (EBS) usage. Using EBS trip data from Hefei, China, the researchers applied an eXtreme Gradient Boosting (XGBoost) model combined with SHapley Additive exPlanations (SHAP) to capture nonlinear and interaction effects. The analysis reveals that travel distance is the primary determinant of usage, with station-level attributes dominating residential-related trips and grid-level attributes dominating workplace-related trips.

Why it matters — It shifts the analytical framework of first- and last-mile transit research from a station-centric catchment area perspective to an origin-destination perspective. This new approach captures how the built environment at farther, non-station trip ends influences electric bike-sharing, which operates over longer distances than conventional bikes.

Caveat: The findings are based on a single case study of Hefei, China, which may limit their generalizability to cities with different urban densities, transit networks, or micromobility regulations.

Journal of Transport Geography · doi

Causal machine learning uncovers thresholds and seasonal reversal in landscape-meteorology-driven urban Heat-PM₂.₅ coupling

This study combines XGBoost-SHAP machine learning with structural causal modeling to analyze how blue-green space patterns and meteorological factors nonlinearly influence land surface temperature and PM2.5 concentrations. Using the Yangtze River Delta urban agglomeration as a case study, the analysis identifies specific threshold effects, such as a 40-42% vegetation coverage threshold for land surface temperature and a 79-80% water aggregation threshold for PM2.5, while confirming a stable positive causal pathway where PM2.5 pollution exacerbates urban warming.

Why it matters — It establishes the specific physical thresholds at which urban landscape interventions reverse their influence on other environmental factors, moving beyond linear assumptions to prove that pollution directly intensifies urban heat.

Caveat: The empirical thresholds and seasonal dynamics are derived specifically from the Yangtze River Delta and may vary in different climatic zones.

Sustainable Cities and Society · doi

Bridging the digital divide in urban mobility: a latent class choice analysis of older adults’ preferences for autonomous vehicles in China

This study analyzed preferences for private and shared autonomous vehicles (AVs) among 996 older adults in Wuhan, China, using a mixed-mode survey that integrated a stated preference experiment, Theory of Planned Behavior indicators, and a five-domain digital capability scale. Using a sequential hybrid modeling framework that combined Structural Equation Modeling with a Latent Class Choice Model, the researchers identified two distinct segments: a public-transport-preferring group (54.82%) characterized by high cost sensitivity and lower digital skills, and an AV-preferring group (45.18%) with higher digital communication capabilities and a greater willingness to pay for time savings.

Why it matters — It demonstrates that trust in automation is not a primary driver of AV adoption among older adults once usability and digital capability constraints are accounted for, shifting the policy focus from building trust to addressing specific digital skill barriers and pricing structures.

Caveat: The empirical findings are based on a stated preference survey in a single Chinese city, which may limit generalizability to regions with different public transit infrastructures or cultural attitudes toward aging and technology.

Transportation Research Part A Policy and Practice · doi

Curbing the car or charging at the curb?

This study examines whether the expansion of curbside electric vehicle charging undermines policies aimed at reducing overall driving. The researchers surveyed residents in Los Angeles and conducted interviews with local officials across multiple North American cities to assess public support and municipal prioritization of curb space.

Why it matters — It demonstrates that EV infrastructure expansion does not act as a policy substitute or erode political will for sustainable transit alternatives, as city officials consistently prioritize bus lanes, bike lanes, and green space over dedicated EV parking.

Caveat: The findings are based on a sample of early-adopter cities and may not reflect political dynamics in municipalities with different planning priorities or lower initial support for active transportation.

Transportation Research Part D Transport and Environment · doi

From queueing performance evaluation to policy guidance: A new modeling framework for escalator systems

This study introduces a queueing and simulation framework to evaluate passenger boarding delays on transit escalators under different behavioral rules, using a metric called Average Delay Before Escalator (ADBE). The model tests Poisson and Single Batch arrival processes alongside three boarding policies: all-standing, walk-left-stand-right, and a hybrid real-world strategy, while accounting for four distinct passenger profiles based on urgency and spacing preferences. Simulation results show that the all-standing policy significantly reduces ADBE and increases passenger throughput during congested periods.

Why it matters — It provides transit operators with a quantitative, behavior-sensitive tool to evaluate escalator capacity, proving that informal 'walk left, stand right' norms actively degrade system efficiency during peak hours.

Caveat: The findings are based on simulated arrival and boarding processes rather than empirical validation with real-world passenger tracking data.

Transportation Research Interdisciplinary Perspectives · doi

‘ If I wouldn’t have had them, I would’ve been lost ’: Tenant Right to Counsel in Cleveland

This study evaluates Cleveland's Right to Counsel program, one of the oldest in the United States, using qualitative interviews with tenants, attorneys, and key program partners. The research analyzes how the program's specific design features and the local eviction ecosystem shape both individual tenant experiences and broader systemic outcomes.

Why it matters — It provides qualitative evidence on the operational processes and contextual factors that drive the success or failure of tenant legal aid programs, moving beyond existing quantitative metrics of representation and eviction rates.

Caveat: The findings are based on a single-city case study of Cleveland and may not directly generalize to jurisdictions with different legal frameworks or housing market dynamics.

Housing Studies · doi

UrbanWell: Benchmarking Multimodal Large Language Models for Spatio-Temporal Urban Wellbeing Analytics

The researchers developed UrbanWell, a large-scale benchmark spanning 38 cities over multiple years to evaluate how 15 multimodal large language models perform on urban wellbeing analytics. The benchmark aligns satellite and street-view imagery at the grid level across 14 indicators, including environmental metrics, spatial accessibility, urban form, vitality, and subjective perceptions. It tests zero-shot performance on both static spatial predictions and temporal tasks like future value forecasting and trend classification.

Why it matters — It establishes a standardized, multi-city testbed that reveals the current limitations of multimodal models in integrating heterogeneous spatial and temporal data, showing they struggle to consistently analyze complex urban indicators despite capturing basic spatial and perceptual cues.

arXiv · pdf · code