2026-09-13

96 journals· 534 candidates — 3 worth your time· 201 without an open abstract

no arXiv arXiv answered and returned nothing this day, so this issue covers the rest. The papers below are real — what is missing is the other half of the scope, not the reading.

Headline

Built environment thresholds for active travel in Mumbai, modelled via machine learning

Today, together

tag shift
no tag ran above its 30-day average
shares references
no items share references today counted over the 3 papers published above
on the same shoulders
5 of the day's papers cite Travel and the Built Environment (Reid Ewing, Robert Cervero 2010); 4 cite Travel demand and the 3Ds: Density, diversity, and design (Robert Cervero, Kara M. Kockelman 1997); 4 cite Social infrastructure and the public life of cities: Studying urban sociality and public spaces (Alan Latham, Jack Layton 2019) counted over the 117 of this day's 204 named papers whose reference lists we have — standing on the same older work is not the same as being about the same thing
first appearance
Global Environment Facility; Humanitarian OpenStreetMap Team; Inner Mongolia University; and 5 more none of them on a paper we published in the 93 days before this issue

Two counts, not one counted twice: every row here but “on the same shoulders” counts the papers published above, and that one counts every paper this day named — those and the ones under Also published today, whose abstracts we could not read.

datapublished: Journal of Urban Management

Ali Shkera, Domokos Esztergár‐Kiss
Journal of Urban Management

Siksha O Anusandhan University

Urban Transport and Accessibility · Transportation Planning and Optimization · Traffic and Road Safety · regression · machine learning · Mumbai

A comparative machine learning approach to analyze the non-linear dynamics of the built environment in active travel behavior

This study evaluated a logistic regression model alongside seven machine learning algorithms to analyze how built environment characteristics influence active travel choices in the Mumbai Metropolitan Region. The Gradient Boosting Decision Trees model performed best, achieving 64% accuracy and an AUC of 0.70 for work trips, and revealed that built environment features hold more predictive power (38% for work and 48% for non-work trips) than socio-demographic factors. Non-linear analysis identified specific threshold ranges where active travel likelihood peaks, including proximity within 200 meters of commercial areas, 500 meters of bus stops, intersection densities exceeding 400 per square kilometer, and a land-use entropy index between 0.7 and 0.8.

Why it matters — It establishes precise, non-linear threshold bands for built environment design in a high-density Global South megacity, replacing general planning assumptions with actionable, locally calibrated spatial metrics for walkability.

Journal of Urban Management · doi

datapublished: Transportation Research Part C Emerging Technologies

Qian Pu, Kun Xie, Yuan Zhu, et al.
Transportation Research Part C Emerging Technologies · was arXiv

National University of Singapore

Autonomous Vehicle Technology and Safety · Human-Automation Interaction and Safety · Traffic and Road Safety · National University of Singapore · Southwest Jiaotong University · Old Dominion University

Generating realistic safety–critical scenarios for vehicle–pedestrian interactions

The researchers developed a three-stage framework combining real-world data pre-training with online reinforcement learning in the CARLA simulator to generate realistic vehicle-pedestrian conflict scenarios. Using a multi-agent state-space Transformer-enhanced DDPG (MA-SST-DDPG) model, they generated the Vehicle-Pedestrian Safety-Critical Interaction (VPSCI) dataset containing over 198,000 high-resolution interaction episodes across eight intersection scenarios. The refined model reproduced human-like evasive behaviors with low trajectory errors (average displacement error of 0.072 m and final displacement error of 0.142 m), matching real-world statistical distributions of conflict severity.

Why it matters — It provides a validated method to generate high-fidelity, safety-critical training and testing scenarios at scale, overcoming the dual limitations of scarce real-world crash data and unrealistic default simulator behaviors for autonomous vehicle validation.

Transportation Research Part C Emerging Technologies · doi

published: Computers Environment and Urban Systems

Steffen Knoblauch, R. Muthusamy, Luís M. A. Bettencourt, et al.
Computers Environment and Urban Systems · was arXiv

Heidelberg University

Municipal Solid Waste Management · Impact of Light on Environment and Health · Advanced Neural Network Applications · deep learning · Heidelberg University · Santa Fe Institute

Open-access model for detecting openly dumped dispersed municipal solid waste from crowdsourced UAV imagery in Sub-Saharan Africa

Summary pending review.

Computers Environment and Urban Systems · doi

Still cited

Reid Ewing, Robert Cervero (2010), Travel and the Built Environment 1 of today's items cite it · 103 of 3240 in the archive stand on it

Also published today

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