2026-09-01

7 arXiv categories· 96 journals· 2486 candidates — 7 worth your time· 100 without an open abstract

Headline

A platform-independent trajectory refinement framework for vehicle emission modeling

Today, together

tag shift
no tag ran above its 30-day average
shares references
no items share references today counted over the 7 papers published above
on the same shoulders
no foundation work was cited by four of the day's papers counted over the 98 of this day's 107 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
University of Virginia none of them on a paper we published in the 81 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.

To analyze speeding behavior at the OpenStreetMap way ID level, one study developed a scalable data pipeline that processes connected vehicle trajectory data and fills in missing speed limits. Another paper introduced a domain adaptation framework to estimate intersection turning movement counts by combining traffic-controller-event-based data, road infrastructure data, and point-of-interest data. Finally, researchers developed the Trajectory Refinement for Emission Modeling framework, which uses mixed-integer quadratic programming to post-process microsimulated vehicle trajectories and eliminate nonphysical acceleration and jerk fluctuations.

codedatapublished: Transportation Research Record Journal of the Transportation Research Board

Shaomin Qin, Haobing Liu, Daejin Kim, et al.
Transportation Research Record Journal of the Transportation Research Board

Tongji University

Vehicle emissions and performance · Advanced Aircraft Design and Technologies · Air Traffic Management and Optimization · Tongji University · Inha University

TRIM: Platform-Independent Framework for Speed Trajectory Refinement in Emission Modeling

The Trajectory Refinement for Emission Modeling (TRIM) framework post-processes microsimulated vehicle trajectories using mixed-integer quadratic programming to eliminate nonphysical acceleration and jerk fluctuations. Validated against high-resolution UAV-derived ground truth across three car-following models (Wiedemann 99, Krauss, and the Intelligent Driver Model), the optimization framework reduces total emission estimation errors to within 9% for nitrogen oxides and 7% for PM 2.5.

Why it matters — It provides a platform-independent, standardized method to correct systematic overestimation of vehicular emissions in traffic microsimulations without requiring modification of the underlying simulation engines.

Transportation Research Record Journal of the Transportation Research Board · doi

published: Transportation Research Record Journal of the Transportation Research Board

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

Pima County Health Department

Human Pose and Action Recognition · machine learning · United States · Pima County Health Department · University of Washington

Domain Adaptation Framework for Turning Movement Count Estimation with Limited Data

This research developed a domain adaptation framework to estimate intersection turning movement counts using traffic-controller-event-based data, road infrastructure data, and point-of-interest data. Tested across 30 intersections in Tucson, Arizona, the model achieved mean absolute errors of 12.29 for left turns, 34.39 for through movements, and 16.17 for right turns, outperforming existing state-of-the-art models.

Why it matters — The framework overcomes the drop in accuracy that machine learning models typically experience when applied to new, unseen intersections with different geometries, signal settings, and driver behaviors.

Caveat: The model was evaluated using data from only a single city, which may limit immediate generalizability to different urban layouts.

Transportation Research Record Journal of the Transportation Research Board · doi

published: Transportation Research Record Journal of the Transportation Research Board

Shi Feng, B. Brian Park, Andrew Mondschein
Transportation Research Record Journal of the Transportation Research Board · was arXiv

University of Virginia

Traffic and Road Safety · Traffic control and management · Traffic Prediction and Management Techniques · University of Virginia

Analyzing Residential Speeding Using Connected Vehicle Data: A Case Study in the Charlottesville, VA Area

This study developed a scalable data pipeline that processes connected vehicle trajectory data and fills in missing speed limits to analyze speeding behavior at the OpenStreetMap way ID level. Applying this framework to residential roads in the Charlottesville, Virginia area, the analysis revealed that 38% of road segments experienced at least one instance of aggressive speeding, while 20% had at least one instance of reckless speeding. The data also showed that point-based speeding rates on the vast majority of residential segments were higher during the night than during the day.

Why it matters — It demonstrates how telematics and connected vehicle data can be systematically processed to identify specific, high-risk residential road segments and temporal speeding patterns, offering a scalable alternative to traditional, localized speed-monitoring equipment.

Caveat: The empirical findings are limited to a single case study of residential roads in the Charlottesville, Virginia area.

Transportation Research Record Journal of the Transportation Research Board · doi

preprint

Andrew Renninger, Egor Kotov, Carmen Cabrera
arXiv physics.soc-ph

Human Mobility and Location-Based Analysis · Urban Transport and Accessibility · Impact of Light on Environment and Health · mobile phone data

The scales of urban mobility shape socioeconomic mixing

Using anonymized GPS traces from approximately 20 million devices across 226 functional urban areas in France, Germany, and the UK, this study decomposes daily mobility into proximal, medial, and distal scales. It finds that while distal trips make up 33% of all travel, they facilitate 42% of encounters between different socioeconomic groups. Reallocating these long-distance trips to local movement reduces experienced socioeconomic mixing by up to one third.

Why it matters — The study demonstrates a fundamental trade-off in urban planning, showing that shifting toward highly localized 'local-living' or polycentric models may inadvertently reduce socioeconomic integration by cutting off the longer trips that drive diverse social encounters.

Caveat: The analysis relies on GPS traces from mobile devices as a proxy for physical encounters and social exposure.

arXiv · pdf

preprint

Ben O'Brien, Lewis J. Lehe
arXiv stat.AP

A Tool for Reconstructing Transit Vehicle Trajectories: A Case Study at IndyGo

This study introduces an open-source R package built on open data standards to clean raw transit automatic vehicle location (AVL) data and reconstruct vehicle trajectories. The tool's performance was validated using a dataset of over 3,000 transit trips from Indianapolis, Indiana, achieving a root mean square error in position estimates under 10 meters at a 15-second polling frequency. The entire dataset was processed in approximately 3 minutes, and the resulting trajectories were used to estimate traffic signal performance metrics.

Why it matters — It provides the first open-source tool for transit practitioners to convert sparse and noisy AVL data into detailed stop-and-go trajectories, enabling them to calculate custom microscopic performance metrics without developing complex custom algorithms.

Caveat: The methodology and tool performance were demonstrated and validated using data from a single transit agency in Indianapolis.

arXiv · pdf

preprint

Luigi Russo, Anabella Ferral, Silvia Liberata Ullo, et al.
arXiv cs.CV

deep learning

Multi-Sensor Mapping of Vulnerable Urban Settlements Using SAR, Multispectral, and Hyperspectral Imagery: A Case Study in Córdoba, Argentina

This study developed a multi-sensor deep learning framework for mapping informal settlements in C3rdoba, Argentina, by fusing high-resolution PlanetScope multispectral imagery, COSMO-SkyMed Synthetic Aperture Radar (SAR) data, and medium-resolution PRISMA hyperspectral observations. Using the official ReNaBaP inventory as reference data across four geographically partitioned folds, the researchers compared early, middle, and late fusion strategies, finding that late fusion combined with hyperspectral data yielded the best classification performance. Thermal analysis during a heatwave event also revealed that these informal settlements exhibit significantly higher surface temperatures than their immediate surroundings.

Why it matters — It demonstrates that integrating hyperspectral data with SAR and multispectral imagery improves the detection of heterogeneous informal settlements, and shows that apparent false positives from the model often correspond to broader, unmapped vulnerable urban areas.

Caveat: The framework's performance and classification thresholds are validated against a single city's official inventory, which may limit its immediate generalisability to other regional urban contexts with different informal construction styles.

arXiv · pdf

preprint

Amadou Tidiane Kassa Diallo
arXiv cs.AI

machine learning · housing transactions

Predicting Residential Rents in Dakar Using Machine Learning

A machine learning pipeline was developed to predict residential rents in Dakar, Senegal, using a newly compiled dataset of 1,507 rental listings scraped from the web. The researchers engineered custom features, including a luxury score and a keyword-based quality score, and compared five models using target encoding for location. The optimized XGBoost model performed best, achieving an R² of 0.847, a Mean Absolute Error of 210,902 XOF, and a Root Mean Squared Error of 324,195 XOF.

Why it matters — It establishes an interpretable benchmark for a highly undocumented rental market where over half of the households rent, while demonstrating that SHAP values and native XGBoost gain yield highly divergent rankings for the importance of target-encoded location variables.

Caveat: The predictive models rely on a relatively small dataset of 1,507 scraped online listings, which may not fully represent the informal rental market in Dakar.

arXiv · pdf

Still cited

Pablo Álvarez López, Evamarie Wießner (2018), Microscopic Traffic Simulation using SUMO 1 of today's items cite it · 12 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.