A new study by researchers at the University of Hawaiʻi reveals that traditional sea-level rise models underestimate the threat to coastal infrastructure on Oʻahu, with up to 85% of critical infrastructure, such as buildings and roads, at risk. The study finds that the number of buildings affected will nearly double, and residents affe...
A new radio-based positioning method allows electric vehicles to align precisely with wireless charging pads, enabling efficient charging. The method uses ultra-wideband radio signals to determine the vehicle's location and orientation, achieving median position and orientation estimation errors of 2.32 millimeters and 0.25 degrees.
Researchers developed a vision-language model that achieves 84.7% average accuracy while generating textual rationales for its predictions. The customised model retained more than 95% of the teacher model’s prediction performance and reduced inference latency by over 50%.
Researchers developed an efficient collaborative perception framework for connected and automated vehicles, reducing communication overhead and improving safety. The framework, called Which2comm, transmits compact object-level semantic information, enabling vehicles to share sensing data while maintaining high perceptual accuracy.
Researchers developed RADGCN, a relation-aware framework for estimating traffic states from sparse sensor data. The approach improved traffic-state estimates by 2.97%–11.82% compared to leading methods, providing more accurate estimates for large road networks with sparse sensor coverage.
A new framework uses LLMs to evaluate autonomous vehicle intelligence by combining measurable data with expert assessments, achieving closer agreement with human expert assessments. The framework assesses five aspects of behavior: safety, comfort, efficiency, social interaction and impact on surrounding traffic.
Researchers developed a physics-informed neural network framework to turn fiber-optic cables into traffic sensors. The approach achieved accurate traffic speed and flow estimation using existing telecom infrastructure without installing new hardware.
A research team developed a physics-informed platooning control framework combining traditional vehicle control methods with learning-based optimization techniques. The framework improved vehicle platooning stability and tracking accuracy under realistic operating conditions.
The Traffic World Model (TWM) generates controllable roadside images, traffic videos, safety-critical events, and LiDAR-style point clouds to overcome data scarcity. It reduces Fréchet distances by 39% for image generation, 20% for video generation, and 73% for safety-critical event generation.
The collaboration aims to accelerate the development of quantum technologies for vehicles, focusing on sensing, communication and energy storage. Researchers will explore applications for new technologies, protecting vehicle data and developing novel materials.
A new national research consortium, NAICT, is launching to help public transportation agencies use AI effectively. The consortium brings together three leading transportation research institutes and invites state and local agencies, industry, and researchers to join, starting with listening sessions and a panel workshop.
The University of Michigan is developing a crash-prevention system to detect vehicles, pedestrians, and cyclists on roads and sidewalks. The system, called SAFETI, uses LiDAR, cellphones, and traffic cameras to predict conflicts and respond through warnings, in-pavement lights, or signal changes.
Researchers have developed a metamaterial that combines stiffness, electromagnetic absorption, and noise reduction, inspired by the intricate scales of lepidopteran insects. The material, just 24mm thick, can support heavy loads, absorb microwave signals, and silence low-frequency noise, solving a long-standing trade-off in flight mate...
Researchers Cing-Dao Kan and Dhafer Marzougui are conducting crash tests to evaluate the performance of safety hardware devices in national parks. The project, funded by the Federal Highway Administration, aims to determine if the railings meet safety performance criteria and improve safety in these areas.
A real-time wildlife detection and driver warning system was successfully tested on a 14-mile stretch of Route 460 in Southwest Virginia, showing drivers reduced speeds by an average of 3 mph when approaching activated signs. The system, which utilizes thermal cameras and AI software, demonstrated an overall accuracy of 97 percent for ...
Dr. Steve Dellenback, VP of SwRI's Intelligent Systems Division, receives Lifetime Achievement Award for his work on intelligent transportation systems. He has developed and managed ITS software for the US public, deploying the ActiveITS platform across 15 state agencies.
Researchers developed Mediator-in-the-Loop-Driving (MILD) framework to enable collaborative decision-making between human drivers and automated systems. The AI-powered mediator continuously interprets driver state and external environment to provide personalized recommendations and explanations.
Researchers developed a hybrid AI model that combines physical driving laws with individual human behavior patterns to improve vehicle movement predictions. The model, CPIDLM, reduces distance prediction errors by 16.7% compared to existing methods, providing a more reliable foundation for autonomous vehicles.
Found-RL integrates vision-language models into reinforcement learning for autonomous driving, achieving competitive driving performance while maintaining real-time policy inference. The platform allows large foundation models to provide semantic feedback without slowing down high-frequency simulation training.
A new ship weather routing framework uses Deep Reinforcement Learning to optimize fuel consumption and safety under extreme weather conditions. The framework successfully planned detour routes that safely bypassed high-wave and strong-wind waters in extreme weather simulations.
Researchers develop COIN, a collaborative interaction-aware reinforcement learning framework for multi-agent self-driving systems, to improve safety and efficiency in complex traffic scenarios. The framework achieves promising results in various evaluations, demonstrating its potential for safer and more efficient autonomous driving.
A hierarchical risk-filtering safety assurance framework has been proposed to address AI-driven autonomous driving safety challenges. The framework uses a four-category assessment metric system and layered safety strategies to improve operation safety.
Researchers developed the Unstructured Scene Benchmark to test vision-language models in unstructured road scenes and degraded visual conditions. The study reveals uneven robustness, weak planning ability, and signs of insufficient visual grounding among current models.
Virginia Tech researchers will develop design tools to integrate certification and safety requirements into the design process from the start. This approach aims to reduce the time, cost, and risk of bringing new aviation technologies to market, with a focus on electrified propulsion technologies.
Researchers have developed a method to generate customizable virtual road networks with adjustable complexity, preserving real-world road features. The approach enables faster identification of functional defects and performance boundaries in autonomous vehicles.
Researchers found China VI trucks can significantly reduce NOx and particle number emissions, but some vehicles still produce high emissions under real driving conditions. After-treatment systems, such as SCR and EGR, play crucial roles in emission control, but their performance varies under different conditions.
Researchers developed a haptic steering wheel that forewarns drivers of autonomous maneuvers, improving driving performance and trust. The system, tested in a simulation, showed drivers drove better when negotiating with the automated system through haptic feedback.
A multidisciplinary team of researchers analyzed almost 300,000 km of Roman roads, revealing a complex network with provincial capitals acting as vital crossroads and Rome not being the dominant hub. The study also found a measurable relationship between Roman and modern roads, but with varying strengths across regions.
A new model forecasts subway passenger destinations and travel times one day ahead by analyzing multi-time-scale patterns in rider behavior. The model outperforms widely used forecasting methods, enabling subway operators to better plan train schedules and manage congestion.
Team Kinetiz from NTU Singapore, K2, and Alp Autonomy won the world's first multi-car autonomous race in Europe, taking the lead at Imola Circuit. The team demonstrated exceptional autonomous driving skills, showcasing the potential of AI in motorsport.
USC researchers will lead a $20M NSF effort to develop AI-powered optimization tools for power grids and supply chains. The project aims to improve decision-making in these complex systems and expand AI education for high school students.
Researchers found that composite metal foam (CMF) can absorb more energy during high-velocity impacts, reducing crash severity and improving driver and passenger safety. CMF outperformed conventional front rail designs, increasing safety limits by up to 40% and reducing maximum deceleration by 38%.
Engineers at Texas A&M University, NASA, and Purdue University create algorithms for managing spacecraft traffic around Gateway, a lunar spaceport. The system balances fuel efficiency and operational demands to reduce the risk of collisions during space missions.
A University of Houston study found that strategic investments in smaller and midsize cities can yield more new riders per dollar than larger metropolitan areas. Smaller systems may have more room to attract new riders due to limited practical transportation options before.
A McGill University-led study recommends cold-climate cities expand electric bus fleets and implement slow charging to optimize operations and reduce energy consumption. The approach aims to eliminate diesel reliance and minimize operating costs, making transit systems more cost-effective and eco-friendly.
A $1 million gift from Roger Krone establishes the University's first space systems endowment and laboratory, enhancing real-world learning experiences for students. The commitment will fund scholarships, research, and experiential learning opportunities in space systems engineering.
A new study from Carnegie Mellon University and Oxford Saïd Business School found that the entry of ride-hailing platforms drove measurable increases in economic output and intermittent employment. Regional GDP per capita increased, and the number of seasonal, temporary, or intermittent jobs rose after Uber and Lyft entered the region.
Research finds that one-pedal driving in electric vehicles yields narrower distributions of speed and acceleration when integrated with Adaptive Cruise Control, improving traffic flow stability. However, manual control produces pronounced speed oscillations and sharper decelerations, highlighting the need for automation.
Knowledge graphs provide explicit reasoning and traceable decision support, while large language models excel in semantic understanding and contextual generalization. The study reveals the growing synergy between these two AI paradigms, promising a more robust and trustworthy autonomous driving system.
Researchers developed a multi-feature fusion filtering framework to improve defect detection in heavy-haul lines, achieving accurate localization in over 90% of test cases. The method outperforms single-feature methods by reducing missed detections and large localization deviations.
This study reveals that off-the-shelf large language models do not exhibit universal risk profiles and can be affected by prompt language, leading to systematic distortions. The findings highlight the need for rigorous empirical calibration before deploying these models in computational social science and choice modeling.
Researchers developed an AI that navigates ships using human-like decision-making, handling complex situations such as narrow channels and multiple vessels. The AI achieved compliance with maritime traffic rules and demonstrated unexpected behavior like local navigation customs.
A study reveals a contrast in decision-making pathways between fuel vehicle and EV owners, highlighting pragmatic benefits for FV owners and habit-driven loyalty among EV owners. The research suggests segment-specific strategies to convert fuel-to-EV owners and retain existing EV users.
A new study introduces a standardized method for measuring greenhouse gas and air pollutant emissions generated across the entire retail journey. The protocol covers online and traditional in-store shopping and provides nationwide parameters for China from 1990 to 2023, identifying opportunities for meaningful reductions.
Researchers introduced CogDrive, a cognition-driven framework for autonomous vehicles to think ahead and make structured reasoning about multiple possible future behaviors. The framework achieves state-of-the-art performance in trajectory prediction and demonstrates smoother maneuvers in dense urban environments.
Global road injury rates declined by 38.3% for new cases and 32.3% for deaths between 1990 and 2023, but progress was uneven across income groups. Low- and middle-income countries bear the greatest burden, with nearly six times higher mortality rates than high-income countries.
A European energy system model determines that a uniform benchmark for vehicle charging infrastructure is not optimal, as the cost-optimal level varies by country. Customized targets for EU member states can unlock benefits through smart charging technologies like V1G and V2G.
A new collaborative study has developed an electrochemical device that can pull carbon dioxide directly from the atmosphere using electricity and water-based chemistry, addressing the planet's excess CO2 problem. The technology is designed to reduce new emissions and remove CO2 that has already accumulated in the atmosphere.
A University of Houston engineer developed a real-time safety system for quadrotor drones that can prevent accidents caused by unexpected events. The new technology uses a 'safety supervisor' module to monitor the drone's tilt and position in real time, ensuring it stays within safety limits.
Researchers developed a two-step simulation framework to assess vibration transmitted through tracks, stations, buildings, and soil near Changsha Huanghua Airport. They found that floor responses were stronger than structural columns at the same floor, with high-speed trains causing perceptible vibrations.
The U-Shift II project develops a vehicle that can transform into different types of vehicles, such as a shuttle or cargo van, using replaceable capsules. This concept enables flexible adaptation to various tasks, saving resources and promoting sustainable mobility.
The MOVEO general assembly showcased advancements in smart transport connections, with a focus on EU-funded research and innovation. The project aims to unite expertise from nine European countries to improve inspection, monitoring, logistics, mobility, and inclusiveness.
A machine learning model, trained on 19 key parameters, can predict wind shear events with a minimum of 15 seconds warning. The model's outputs showed deviations from real outcomes within 5% across all forecast horizons, suggesting improved aviation safety.
A new study found that protected bike lanes in NYC causally increase Citi Bike ridership, with a statistically significant effect of approximately 379 additional rides per station per month. However, painted bike lanes and sharrows showed no detectable causal effect, highlighting the importance of design details in cycling infrastructure.
SwRI's Intelligent Systems Division has achieved CMMI, ISO 9001, and CMMC certifications to demonstrate its commitment to delivering secure and reliable transportation solutions. The certifications validate SwRI's robust cybersecurity protections and quality management framework.
Researchers at King's College London developed a new algorithm that can automatically explain why some self-driving cars crash. The approach analyzes past events to identify the root cause of failures in complex and rare cases.
Researchers at Texas A&M University are designing how humans will build and survive on the moon, focusing on sustainable construction using lunar regolith. The institution's efforts aim to reduce costs associated with shipping materials to the moon, making it possible to produce rocket propellant locally.
The new center aims to make mobility systems more sustainable and efficient by leveraging AI, autonomy, digital infrastructure, and data-driven decision-making. Stanford faculty members across five schools are affiliated with the center, which will be funded through corporate partnerships.
A research team led by POSTECH developed an AI framework that can predict and account for microscopic defects in metal 3D printing, improving the reliability of metal components. The framework achieves a Mean Absolute Error (MAE) of just 9.51 MPa, outperforming conventional approaches.
A perspective paper on digital transportation explores its basic theories and proposed research directions. The author highlights the need for a comprehensive theoretical framework to guide the development of modern vehicles and digital transformation of transportation infrastructure.
SourceSciOpen·JournalJournal of Highway and Transportation Research and Development (English Edition)·DateMay 12, 2026