It’s possible that I shall make an ass of myself. But in that case one can always get out of it with a little dialectic. I have, of course, so worded my proposition as to be right either way (K.Marx, Letter to F.Engels on the Indian Mutiny)
Sunday, November 08, 2026
Open Access Week 2026: Lowering the Cost of Knowledge
Preprints.org brought together speakers from libraries, scholarly infrastructure organizations, preprint platforms, and the Preprints.org research community for a webinar on 29 October 2026.
Every October, researchers, librarians, publishers, universities, and Open Science advocates around the world come together for International Open Access Week—a global annual initiative established by SPARC to promote openness in research that encourages stakeholders to explore ways to make knowledge more accessible.
The theme of Open Access Week 2026 (October 19-25), “The Cost of Knowledge”, continued discussions from recent years, including “Who Owns Our Knowledge?” in 2025 and “Community over Commercialization” in 2023 and 2024.
This year, it invited the global academic and research community to consider the cost of creating, sharing, and sustaining knowledge, going beyond financial expenditure and expanding the discussion to human and environmental costs, particularly when these costs are driven by private rather than public interests.
Preprints.org Webinar
For OA Week 2026, Preprints.org brought together speakers from libraries, scholarly infrastructure organizations, preprint platforms, and the Preprints.org research community for a webinar on 29 October 2026.
It aimed to answer an often-overlooked question: what does it cost when knowledge takes too long to reach those who need it? Panelists explored these costs and the role that open and timely research sharing can play in addressing the issues.
The webinar featured four speakers, chaired by Dr. Ioana Craciun, scientific advisor at Preprints.org:
Ms. Melissa Chim, scholarly communications librarian at Excelsior University, discussed preprints in the world of scholarly communication.
Mr. Lloyd Shu, product manager of Preprints.org, spoke about how preprints can reduce the cost of keeping knowledge waiting.
Ms. Carolina Tangushi, SciELO Preprints platform coordinator, addressed the role of SciELO Preprints in building trust and accelerating research in Latin America.
Dr. Martyn Rittman, program lead at Crossref, talked about the role of metadata in knowledge sharing.
A recording of the event is available here. The conversation also continues on SciProfiles, where the unaddressed questions from the webinar are archived for further discussion, with the speakers joining the conversation to share their responses.
Preprints matter in the discussion
In 2023 and 2024, the theme “Community over Commercialization” highlighted preprint servers as examples of community-oriented infrastructure that can better serve the interests of the research community and the public. Even earlier, the 2016 topic “Open in Action” encouraged researchers to take actions toward openness, including posting preprints to make their work more accessible.
These discussions highlight the role preprints can play in making research more accessible to the wider community. Preprints.org is founded on the principle that openness, transparency, and timely research communication can benefit not only the research community but also the wider society.
“Research creates values when it can be seen, discussed, challenged, and built on,” said Lloyd Shu, product manager of Preprints.org. “Preprinting is not only about sharing the paper earlier, it is about starting the scholarly conversation sooner, and helping knowledge go further.”
This year’s OA Week provided an important opportunity to reflect on the role of preprints in open scholarly communication. Since 2025, Preprints.org has hosted OA Week webinars to create a space for discussion of preprints alongside the challenges and opportunities in open science, joined by key stakeholders from Open Science organizations and community initiatives working to advance transparent, inclusive, and reproducible research.
Lower the cost of knowledge with preprints
Open Access is not only about making research more accessible, but also about lowering the barriers to sharing, sustaining, and disseminating knowledge. Preprints offer a way to address these challenges.
INRS and the Pasteur Network bring together scientists from 25 countries in Québec to strengthen the fight against pandemics
Held in North America for the first time, the Pasteur Network Annual Meeting highlighted how artificial intelligence and emerging technologies can advance global health preparedness and response
Institut national de la recherche scientifique - INRS
Held in North America for the first time, the international gathering provided a unique forum for scientific leaders, researchers, and public health experts to explore how artificial intelligence and emerging technologies can strengthen pandemic preparedness and response. Over three days, participants exchanged knowledge, shared best practices, and reinforced collaborations to address current and future global health challenges.
“INRS was honoured to co-organize and host this major international gathering in Québec. Bringing together scientists from around the world in Laval demonstrated the importance of international collaboration in advancing knowledge and developing innovative solutions to better protect populations against future health threats. The event also showcased Québec's expertise and leadership in research, innovation, and global health,” said Luc-Alain Giraldeau, Chief Executive Officer of INRS.
Over three days, ten sessions and four roundtables explored how AI and emerging technologies can help detect, interpret, and respond more rapidly to infectious disease threats. Discussions focused on epidemic intelligence, pathogen genomics, the accelerated development of vaccines, therapeutics, and diagnostics, and the complex interactions between human, animal, and environmental health that underpin the One Health approach.
Participants also examined the ethical, governance, and equity dimensions of the growing use of AI in research and public health. The meeting provided an opportunity to strengthen existing collaborations and foster new partnerships among members of the Pasteur Network and the broader international scientific community.
PNAM 2026: Key outcomes
Scientists, public health experts, and research leaders from 25 countries gathered in Laval, Québec;
Ten scientific sessions and four roundtables explored artificial intelligence and emerging technologies for global health;
Participants shared advances in epidemic intelligence, pathogen genomics, vaccine development, therapeutics, and diagnostics;
Discussions addressed the ethical, governance, and equity implications of AI in public health and biomedical research;
Early-career researchers presented innovative projects and engaged with international scientific leaders;
New opportunities for scientific collaboration and training were identified across the Pasteur Network and its partner institutions;
The meeting reinforced the role of international cooperation in strengthening preparedness for future health threats.
“The Pasteur Network brings together institutions that share a common ambition: putting science to work for global health. PNAM 2026 offered a unique opportunity to harness the expertise across our network to accelerate innovation and strengthen our collective capacity to address global health challenges, from infectious diseases to emerging health threats and other pressing public health priorities,” said Mario Moreira, President of the Pasteur Network.
The meeting also highlighted Québec's growing role within the Pasteur Network and its contribution to global health research and innovation. Participants included leading scientists, public health experts, institutional leaders, and members of the next generation of researchers committed to advancing scientific solutions to pressing health challenges.
The Pasteur Network is an alliance of 33 research, innovation, and public health institutions across five continents. Many of its members play a key role in infectious disease surveillance, biomedical research, epidemic preparedness, and public health response.
Hosting PNAM 2026 marks an important milestone in the development of the Pasteur Network in Québec. A member of the Network since 2005 through its Armand-Frappier Santé Biotechnologie Research Centre, INRS expanded its membership in 2026 to include all of its research centres.
“The discussions held throughout PNAM 2026 underscored the importance of combining expertise from across disciplines to address increasingly complex health challenges. The participation of all INRS research centres in the Pasteur Network positions us to contribute new perspectives and innovative approaches to this global effort,” said Isabelle Delisle, Scientific Director of INRS.
The same year, the Faculty of Medicine at Université de Montréal joined the Network as an Associate Member under INRS sponsorship, further strengthening the Pasteur Network's presence in North America.
– 30 –
About INRS
The Institut national de la recherche scientifique (INRS) is a university dedicated exclusively to research and graduate training in strategic sectors in Québec. Since its founding in 1969, it has actively contributed to Québec’s economic, social, and cultural development. INRS ranks 1st in Québec for research intensity. It is made up of five interdisciplinary research and training centres, located in Québec City, Montréal, Laval, Varennes, and Charlevoix, which focus their activities on strategic sectors: Eau Terre Environnement, Énergie Matériaux Télécommunications, Urbanisation Culture Société, Armand-Frappier Santé Biotechnologie, and Ruralités durables. Its community includes nearly 1,500 students, postdoctoral fellows, faculty members, and staff.
About the Pasteur Network
The Pasteur Network is an alliance of more than 30 organizations that plays a crucial role in tackling global health challenges through science, innovation, and public health. Its strength lies in the diversity and broad geographic reach of its members, spread across 25 countries on five continents and united by shared values and missions in service of populations, fostering a dynamic community of scientific knowledge and expertise that is both locally rooted and globally connected (pasteur-network.org).
Can LLMs help judge how intelligently autonomous vehicles behave in traffic?
The framework converts autonomous-vehicle interaction data into structured language for LLM-based, expert-like assessment, then maps five-dimensional objective metrics to intelligence scores through an attention-based neural network. The five dimensions are safety, comfort, efficiency, social interaction and impact on surrounding traffic.
Self-driving vehicles are moving from pilot projects toward wider commercial deployment, but a vehicle that performs well on conventional tests may still behave in ways that feel hesitant, abrupt or disruptive in real traffic. Researchers at Tongji University have developed a new framework designed to evaluate the overall “intelligence” of high-level autonomous driving systems, rather than judging isolated functions such as braking or lane keeping alone.
The study, “Towards the Intelligence Evaluation of High-level Autonomous Vehicles: A Subjective-Objective Mapping Method via LLM,” introduces a Subjective-Objective Mapping Evaluation (SOME) framework. It combines measurable driving data with expert-style assessments generated by a large language model (LLM), then trains an attention-based neural network to learn how objective indicators relate to the broader judgments that human experts make when observing vehicle behavior.
“A high-level autonomous vehicle should not be evaluated only by whether it avoids a crash,” says Peng Hang, corresponding author of the study and a research professor at Tongji University. “Its behavior should also be judged by whether it moves smoothly, completes tasks efficiently, interacts reasonably with other road users and avoids unnecessarily disrupting surrounding traffic.”
From raw trajectories to expert-like evaluation
The framework evaluates five complementary aspects of autonomous driving behavior: safety, comfort, efficiency, social interaction and impact on surrounding traffic. It uses interaction segments from the Waymo Open Dataset, including following, merging and crossing situations. To improve coverage of rare but important poor-performance cases, the researchers also augmented the dataset with physically feasible low-score interaction samples. In total, the study constructed 5,200 valid interaction segments for training and evaluation.
Because language models are not naturally suited to interpreting raw vehicle trajectories and map data, the researchers developed a Data-Language Mapping Module. This component converts key interaction events, vehicle states and abnormal motions into structured natural-language descriptions. The descriptions are combined with autonomous-driving knowledge, traffic rules and driving-behavior guidance, enabling the LLM to assess the interaction in an expert-like manner.
The LLM-based judgments are then used as supervisory labels for an attention-based neural network. The network learns nonlinear relationships among ten quantitative metrics instead of relying on a fixed set of manually assigned weights. This allows the model to account for the fact that vehicle safety, motion smoothness, travel efficiency and social behavior can influence one another in complex ways.
Closer agreement with expert reasoning
To test whether the LLM-generated evaluations resembled expert reasoning, the team conducted a Turing Test with ten doctoral students specializing in autonomous driving. Across 66 interaction samples, the LLM-based evaluations achieved an average passing rate of 86.75 percent, meaning the reviewers frequently judged them to be human-generated or could not reliably distinguish their source.
In additional experiments, the proposed LLM evaluation strategy achieved the closest overall agreement with human expert assessments among the compared settings. It obtained a mean absolute error of 0.763, a Pearson correlation of 0.863 and the lowest score-distribution divergence from expert ratings. The downstream attention-based evaluation model achieved a mean absolute error of 0.068 and an R² value of 0.880 when predicting the LLM-derived intelligence scores.
“The goal is not to replace human experts in every decision,” Hang says. “Rather, the framework provides a scalable and consistent way to translate high-dimensional driving behavior into a comprehensive evaluation that can support scenario-based testing and system improvement.”
Why fixed weights can miss important behavior
The researchers applied their method to five autonomous-driving algorithms in six scenario-based tests from the OnSite Autonomous Driving Challenge. The rankings produced by the SOME framework were generally consistent with those of human experts. Compared with an analytic hierarchy process (AHP) method, the proposed model was better able to identify cases in which strong safety scores masked weaknesses in comfort, efficiency or interaction quality.
For example, a fixed-weight method can assign a high overall score to a vehicle with strong safety indicators even when it causes excessive delay or makes uncomfortable maneuvers. By learning relationships among metrics, the new model can capture these trade-offs more directly. The analysis also found that safety and comfort were particularly important for distinguishing different levels of autonomous-driving intelligence, while each of the five dimensions contributed useful information to the final assessment.
The authors note that the current framework focuses mainly on trajectory-level interactions and high-level behavioral indicators. Future work will incorporate richer multimodal information, such as video, perception outputs and more detailed environmental context, while testing the framework across additional datasets, regions, traffic cultures and evaluation platforms.
Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University, and Professor Xiaopeng (Shaw) Li from University of Wisconsin–Madison. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.
It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan PhaseⅡ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in "TRANSPORTATION" category. In 2026, its 2025 IF was announced as 12.7, maintaining the Top1 position (1/66, Q1) in the same category.
From Volume 6 (2026), Communications in Transportation Research will be published by Tsinghua University Press on the SciOpen platform with the official journal website at https://www.sciopen.com/journal/2097-5023. We kindly request that all new manuscript submissions be made through the journal’s submission system at https://mc03.manuscriptcentral.com/commtr. For any submission-related inquiries, please contact the Editorial Office at commtr_e@mail.tsinghua.edu.cn.
Starting from textual traffic descriptions and structural road priors, the framework first synthesizes a controllable roadside-view image through a conditional Diffusion Transformer (cDiT) backbone. The generated image then serves as a shared scene anchor for subsequent multimodal generation tasks, including traffic-video synthesis, safety-critical event generation, depth prediction, and LiDAR generation.
To overcome the scarcity of diverse roadside traffic data, researchers developed the Traffic World Model, or TWM, a unified framework that generates controllable roadside images, traffic videos, safety-critical events, and LiDAR-style point clouds.
Large-scale, accurately annotated traffic data are essential for autonomous driving and intelligent transportation systems. However, collecting multimodal data from infrastructure-mounted cameras and LiDAR sensors is expensive and time-consuming. Moreover, safety-critical situations, such as road debris, lane-occupying construction equipment, accidents, heavy fog, and snow, occur infrequently and are difficult to collect at scale.
“Most existing driving-generation models focus on the ego-vehicle perspective. Roadside traffic generation is more challenging because infrastructure sensors observe wider areas, more complex road structures, and traffic participants at substantially different distances and scales,” says Chunmian Lin, a researcher at Beihang University. “We therefore designed TWM specifically for roadside-view multimodal data generation. The framework first creates a controllable roadside image and then uses that image as a common scene anchor for video, safety-critical event, depth, and LiDAR-style point-cloud generation.”
A shared image-centered framework for four generation tasks
At the core of TWM is a conditional Diffusion Transformer, or cDiT. The model receives three complementary types of input: textual descriptions of environmental conditions, projected three-dimensional bounding boxes describing traffic participants, and map-derived priors representing road topology. The textual descriptions specify background characteristics such as the time of day, weather, illumination, and atmospheric conditions. Object boxes control the position, category, and spatial distribution of vehicles and other traffic participants, while the map prior constrains the structure of roads and lanes. Together, these inputs allow the model to control not only the appearance of the generated scene but also its spatial organization. After generating a roadside image, TWM uses it as the visual foundation for three additional generation branches.
For video generation, the researchers introduced traffic-aware prompt expansion, which supplements the original text prompt with information about fixed-camera viewpoints, traffic-flow directions, lane-following behavior, and temporal-consistency requirements. The video model is further adapted to roadside-surveillance footage using low-rank adaptation, or LoRA.
For safety-critical event generation, the researchers formulate the problem as a conditional image-editing task. Event-related objects are extracted from traffic images, semantically annotated, quality-filtered, reconstructed into coarse three-dimensional assets, and inserted into target scenes at specified locations, poses, and scales. This approach allows the model to create rare traffic situations involving road debris, traffic cones, warning signs, construction equipment, accident-related vehicles, and other potential hazards while maintaining consistency with the original road environment.
For LiDAR generation, TWM first predicts a dense pseudo-depth map from the input image. Sparse LiDAR measurements and monocular-depth estimates are combined through outlier filtering, global scale alignment, and edge-aware residual propagation. The predicted depth is then projected into three-dimensional space and transformed into the LiDAR coordinate system to create LiDAR-style point clouds.
Improved performance across four multimodal generation tasks
The researchers evaluated TWM using several vehicle–infrastructure datasets. The experiments assessed four tasks: roadside image generation, traffic-video generation, safety-critical event generation, and image-to-LiDAR-style point-cloud generation.
Compared with the strongest baseline for each task, TWM reduced the Fréchet Inception Distance by 39% for general traffic-image generation, the Fréchet Video Distance by 20% for traffic-video generation, the Fréchet Inception Distance by 73% for safety-critical event generation, and depth error by 11% for point-cloud generation.
The researchers also evaluated the model on a camera-disjoint test set containing approximately 2,000 traffic-scene images collected from 23 previously unseen roadside cameras. The test data covered highways and intersections under sunny, cloudy, and rainy conditions.
Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University, and Professor Xiaopeng (Shaw) Li from University of Wisconsin–Madison. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.
It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan PhaseⅡ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in "TRANSPORTATION" category. In 2026, its 2025 IF was announced as 12.7, maintaining the Top1 position (1/66, Q1) in the same category.
From Volume 6 (2026), Communications in Transportation Research will be published by Tsinghua University Press on the SciOpen platform with the official journal website at https://www.sciopen.com/journal/2097-5023. We kindly request that all new manuscript submissions be made through the journal’s submission system at https://mc03.manuscriptcentral.com/commtr. For any submission-related inquiries, please contact the Editorial Office at commtr_e@mail.tsinghua.edu.cn.
Feature-based knowledge distillation framework. During the distillation process, the teacher model is kept frozen, and only the student model, specifically, its LLM backbone is optimized.
To answer this question: Can vision-language models predict imminent vehicle collisions accurately, explicably and efficiently? Researchers at Imperial College London have developed a framework that combines task-specific fine-tuning, two-step chain-of-thought prompting and feature-based knowledge distillation. The customised model achieved 84.7% average accuracy while generating textual rationales for its predictions. After distillation, the compact model retained more than 95% of the teacher model’s prediction performance and reduced inference latency by over 50%.
Vehicle collision warning systems can support timely warnings and interventions, but conventional deep learning models often provide limited explanation for their predictions. Vision-language models offer a potential solution by jointly analysing driving videos and generating natural-language rationales. However, general-purpose models lack specialised collision-related knowledge and often require substantial computational resources.
To address these challenges, a general-purpose vision-language model was fine-tuned using real-world pre-collision and normal driving video clips. A two-step chain-of-thought prompt guided the model to first analyse the driving scene and identify potential risk sources, and then predict whether a collision was likely to occur imminently. The resulting model could therefore provide both a prediction and an accompanying explanation.
Knowledge was then transferred from the customised model to a smaller student model by aligning their intermediate hidden representations. Experimental results showed that the customised model achieved average accuracy, precision and recall of 84.7%, 83.6% and 86.8%, respectively, outperforming conventional deep learning baselines. The distilled model retained more than 95% of the teacher model’s prediction performance while reducing inference latency by over 50%.
The study provides a new approach to balancing prediction performance, explainability and computational efficiency in vision-based collision warning systems. Future research could incorporate additional vehicle kinematic information and further investigate the internal reasoning mechanisms of vision-language models.
Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University, and Professor Xiaopeng (Shaw) Li from University of Wisconsin–Madison. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.
It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan PhaseⅡ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in "TRANSPORTATION" category. In 2026, its 2025 IF was announced as 12.7, maintaining the Top1 position (1/66, Q1) in the same category.
From Volume 6 (2026), Communications in Transportation Research will be published by Tsinghua University Press on the SciOpen platform with the official journal website at https://www.sciopen.com/journal/2097-5023. We kindly request that all new manuscript submissions be made through the journal’s submission system at https://mc03.manuscriptcentral.com/commtr. For any submission-related inquiries, please contact the Editorial Office at commtr_e@mail.tsinghua.edu.cn.
Which2comm is a collaborative 3D object detection framework based on object-level sparse features. LiDAR point cloud is input into Extractor, obtaining the sparse SemDBs. Multi-agent sparse features are encoded with RTE, and further processed through Fuser and Decoder.
Accurate perception is essential for the safe operation of connected and automated vehicles. However, a single vehicle is limited to its own viewpoint, making it vulnerable to occlusion, restricted sensing range, and detection failures in complex traffic scenes. Collaborative perception—where vehicles share sensing information—can address these shortcomings. Nevertheless, existing approaches often transmit full feature maps or dense sensor data, imposing significant communication overhead that hinders practical deployment.
To address this limitation, researchers from the Department of Automation at Tsinghua University and the Institute of Intelligent Transportation Systems at Zhejiang University developed Which2comm, an efficient collaborative perception framework for connected and automated vehicles. Instead of sharing broad region-level feature maps, the framework transmits compact object-level semantic information, termed Semantic Detection Boxes (SemDBs). These sparse features retain key spatial and semantic details necessary for 3D object detection while greatly reducing the data volume that must be exchanged between vehicles.
A central innovation of the framework is its safety-risk-aware collaboration strategy. Rather than treating all vehicles equally in communication, Which2comm adaptively selects the most informative collaborative agents based on their relevance to driving safety. The framework also incorporates temporal feature fusion and relative time encoding, enabling better handling of communication latency and asynchronous data from different vehicles.
The researchers evaluated Which2comm on multiple collaborative perception datasets, including V2XSet, the real-world V2V4Real dataset, and a newly constructed large-scale dataset named Multi-V2X. Multi-V2X comprises 146,000 LiDAR frames, 549,000 RGB images, and approximately 4.219 million 3D annotation boxes, supporting high-density connected-vehicle scenarios with up to 31 connected agents and a connected-vehicle penetration rate of 86.21%.
Experimental results demonstrate that Which2comm achieves strong detection performance at a very low communication cost. On V2XSet, it attains 0.929 AP@0.5 and 0.833 AP@0.7 while requiring only 0.022 MB of transmitted data. The framework also outperforms existing collaborative perception methods on Multi-V2X and V2V4Real, showing robustness under communication latency and scalability in dense connected-vehicle environments.
The study indicates that object-level sparse communication can render collaborative perception more efficient, scalable, and practical for future intelligent transportation systems. By reducing communication redundancy while maintaining high perceptual accuracy, Which2comm offers a promising step toward real-time cooperative perception in connected and automated transportation systems.
Looking forward, the researchers plan to extend the framework to more complex real-world scenarios, such as multi-modal sensor fusion, localization uncertainty, and packet loss, further enhancing the model’s applicability and readiness.
Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University, and Professor Xiaopeng (Shaw) Li from University of Wisconsin–Madison. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.
It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan PhaseⅡ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in "TRANSPORTATION" category. In 2026, its 2025 IF was announced as 12.7, maintaining the Top1 position (1/66, Q1) in the same category.
From Volume 6 (2026), Communications in Transportation Research will be published by Tsinghua University Press on the SciOpen platform with the official journal website at https://www.sciopen.com/journal/2097-5023. We kindly request that all new manuscript submissions be made through the journal’s submission system at https://mc03.manuscriptcentral.com/commtr. For any submission-related inquiries, please contact the Editorial Office at commtr_e@mail.tsinghua.edu.cn.
Traffic-induced vibrations are captured by roadside optical fibers and transformed into time–location DAS signal maps. A coarse perception stage first identifies preliminary spatiotemporal trajectory patterns, which are then refined by a trajectory recognition network to recover vehicle trajectories more accurately. By integrating distributed acoustic sensing with data-driven spatiotemporal analysis, the method enables robust, non-intrusive, and accurate traffic trajectory monitoring in complex road environments.
Researchers have developed a physics-informed neural network framework that turns standard fiber-optic cables commonly used for communication into large-scale traffic sensors. By embedding traffic flow equations into deep learning models, the team achieved accurate traffic speed and flow estimation across long-distance expressway using existing telecom infrastructure without installing new hardware.
Distributed acoustic sensing (DAS) turns standard fiber‑optic telecommunication cables, already buried along most highways, into dense vibration sensors that can monitor traffic in real time. But raw DAS signals are noisy and indirect: overlapping vehicle signatures often produce coarse, fragmented traffic maps that lack the precision needed for large‑scale traffic management.
To solve this, researchers at Hefei University of Technology, USTC, and Tsinghua University developed a physics‑informed neural network (PINN) framework that restores accurate traffic flow and speed maps from DAS data.
“We treat DAS‑based traffic state perception as an image restoration problem,” says Yang Ma, Associate Professor at Hefei University of Technology. “The key is embedding the fundamental laws of traffic flow, such as the Lighthill‑Whitham‑Richards conservation equation, directly into the neural network’s loss function. This ensures predictions are physically consistent, not just statistically plausible.”
The team built a realistic DAS simulation platform, validated with field data from China’s Chuxin Expressway, and tested 16 model combinations. The LWR‑constrained models outperformed purely data‑driven baselines by over 30% in error reduction, while more complex physics models actually degraded performance due to noise amplification. They also found that fusing DAS with camera data can help, but denser camera placement does not always improve accuracy, it can introduce spatial discontinuities that destabilize the physics loss.
Crucially, few‑shot transfer learning allowed the pre‑trained model to adapt to new highways (e.g., US‑101 and I‑80) with only a few hundred local samples, restoring high accuracy despite major differences in road geometry and traffic patterns.
“This means we can deploy a generalized model across diverse expressways with minimal site‑specific calibration,” Ma explains. “Instead of installing new hardware, we repurpose existing fibers as a distributed sensing network, which is scalable, cost‑effective, and can work in all‑weather conditions.”
The approach offers a paradigm shift from capital‑intensive sensor installation to infrastructure‑as‑sensor, paving the way for truly ubiquitous traffic monitoring on long‑distance highways.
Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University, and Professor Xiaopeng (Shaw) Li from University of Wisconsin–Madison. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.
It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan PhaseⅡ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in "TRANSPORTATION" category. In 2026, its 2025 IF was announced as 12.7, maintaining the Top1 position (1/66, Q1) in the same category.
From Volume 6 (2026), Communications in Transportation Research will be published by Tsinghua University Press on the SciOpen platform with the official journal website at https://www.sciopen.com/journal/2097-5023. We kindly request that all new manuscript submissions be made through the journal’s submission system at https://mc03.manuscriptcentral.com/commtr. For any submission-related inquiries, please contact the Editorial Office at commtr_e@mail.tsinghua.edu.cn.
