Purpose The dynamic nature of production environments with uncertainty about future tasks places a high demand on the flexibility for assigning tasks to a fleet of Automated Guided Vehicles. The goal of our research is to reduce the number of required vehicles compared to the traditional approach while achieving the same throughput for economic and sustainability reasons by enabling a more intelligent battery management. Design/methodology/approach Our proposed approach considers charging as an option during robot task assignment. We extend a previously presented sequential single-item auction algorithm for online scheduling using utility functions to compare the options for performing tasks or charging. This enables a more intelligent decision, depending on the current system state, focusing on improved availability in phases with high demand and efficiency. A material flow simulation highlights potential issues when disregarding battery management and validates the proposed methods. Findings Despite the advancements in modern battery technologies, battery management remains an important aspect for continuous operation without breaks for charging in between shifts. Additionally, frequently driving to charging stations for opportunity charging can generate significant additional movement and power consumption. It is demonstrated that our approach improves system performance while reducing the total number of charging operations, which allows a reduction in required vehicle count. Originality/value Our approach represents a deeper integration of battery management aspects into the robot-task assignment problem. The proposed online scheduler is very flexible, as any number of system states can be considered in the form of additional utilities, enabling potential for further optimization.

Traffic analysis and control are playing an increasingly important role in the management of modern urban infrastructure. This paper provides an extensive literature review on techniques based on computer vision and machine learning for real-time video surveillance systems to imporove urban traffic control with the emphasis of vehicle identification, counting, tracking and violation detection. We review recent methods based on computer vision and machine learning. This paper attempts to show the improvement in machine learning based techniques for the analysis of video surveillance for the traffic monitoring, discussing issues such as occlusions handling, temporal meteorological variations, and computational efficiency. It includes trajectory analysis, multi-object tracking, and automatic violation detection systems. Results suggest superior results in detection settings with deep learning methods, particularly YOLO-based architectures and Transformer models. Nevertheless, complex traffic situations and bad weather are still an obstacle to handle the real-time processing. This review describes the current state-of-the-art methods and outlines future work.

The soundscape of a given location can be transformed over time due to factors such as urbanization, technological advances and social changes, among others. The main tourist attraction in Natal-RN, Brazil, Ponta Negra beach, has undergone extensive intervention to contain the effects of erosion caused by the advance of the sea. The beach has had its sand strip expanded, now reaching about 100 meters at low tide. This work has had a major environmental impact, also affecting the soundscape. The aim of this article is to assess the impact of the Ponta Negra beach renourishment project on users' perceptions of the soundscape. The methodology was based on Sound Pressure Level measurements, vehicle counts and questionnaires with local users before and after the construction work. The number of measurement points after the work had to be increased due to the widening of the sand strip. The results show that, after the construction work, the sound of the sea became more distant for passers-by on the promenade, and the urban sounds of the nearby traffic lane became more prominent. In addition, there were more scenes of different human activities, such as soccer and volleyball games, which consequently changed the soundscape.

Using open data to derive parsimonious data-driven models for uncovering the influence of local traffic and meteorology on air quality: The case of Madrid.

The Vehicle Counting Web is a web-based system designed to automatically record the number of vehicles in traffic survey activities. This system was developed as a solution to the limitations of manual methods previously employed by the Tegal District Transportation Agency, such as tally counters and stopwatches, which are prone to errors and time-consuming. The platform enables officers to digitally input location data, conduct real-time vehicle counting, and store as well as manage data within a centralized system. In addition to improving fieldwork efficiency, the system also provides data visualization in the form of daily traffic volume graphs, which can be utilized for traffic analysis and planning. The development of this website aligns with the broader agenda of digitalizing public services and supports the research roadmap on sustainable transportation at UNNES. The expected outcomes of this research include publication in an accredited national journal, the attainment of copyright, and direct utilization by relevant agencies for modern and adaptive traffic monitoring.

Traffic congestion and vehicle queue formation at signalized intersections represent critical challenges in modern urban transportation systems, requiring accurate real-time detection methods for effective traffic management. This study presents a deep learning-based approach for real-time vehicle queue state classification that integrates You Only Look Once (YOLO) object detection with Simple Online Real-time Tracking (SORT) algorithms using standard traffic camera footage. The proposed system performs multi-class vehicle classification, real-time vehicle tracking with unique ID assignment, and speed estimation through camera calibration techniques, achieving 16.42 FPS average processing speed across diverse video scenarios. A comprehensive queue state detection methodology is developed that categorizes traffic conditions into three categories: Heavy traffic, stable flow, and free flow based on the analysis of average speeds of the detected vehicles, excluding motorcycles and bicycles due to their distinct traffic behavior patterns. Experimental validation across several test datasets encompassing both high and low resolutions demonstrates robust vehicle detection performance across all vehicle classes. Speed estimation accuracy ranges from 89% to 99%, validated against vehicle counting and tracking in designated traffic lanes, providing essential data for queue analysis. The system achieves vehicle counting accuracy ranging from 78.57% to 100% across different scenarios. The system offers a cost-effective alternative to traditional sensor-based methods by utilizing existing traffic-surveillance infrastructure, making it suitable for widespread deployment in intelligent transportation systems. Results indicate the proposed approach successfully detects queue states in real-time conditions across diverse traffic scenarios, from heavy congestion to free flow conditions. This research advances computer vision-based traffic monitoring by demonstrating the practical effectiveness of integrated object detection and tracking algorithms, contributing to improved traffic flow optimization and congestion management.

Deep Learning is a popular Machine Learning algorithm that is widely used in many areas in current daily life. Its robust performance and ready-to-use frameworks and architectures enables many people to develop various Deep Learning-based software or systems to support human tasks and activities. Traffic monitoring is one area that utilizes Deep Learning for several purposes. By using cameras installed in some spots on the roads, many tasks such as vehicle counting, vehicle identification, traffic violation monitoring, vehicle speed monitoring, etc. can be realized. In this paper, we discuss a Deep Learning implementation to create a vehicle counting system without having to track the vehicles movements. To enhance the system performance and to reduce time in deploying Deep Learning architecture, hence pretrained model of YOLOv5 is used in this research due to its good performance and moderate computational time in object detection. This research aims to create a simple vehicle counting system to help human in classify and counting the vehicles that cross the street. The counting is based on four types of vehicles, i.e. car, motorcycle, bus, and truck, while previous research counts the car only. As the result, our proposed system capable to count the vehicles crossing the road based on video captured by camera with the highest accuracy of 97.72%

Traffic congestion and inefficient road space utilization are major challenges in modern urban transportation systems. The "Automatic Movable Road Divider" project aims to provide a dynamic and intelligent solution to these problems by designing a road divider system that can be automatically repositioned based on real-time traffic conditions. This system employs sensors, microcontrollers (e.g., Arduino), and motorized mechanisms to detect traffic density in multiple lanes and adjust the position of the divider accordingly. During peak hours, the divider shifts to allocate more lanes to the congested direction, thereby optimizing traffic flow and reducing delays. The system can be controlled automatically through an embedded program or manually via remote control, ensuring flexibility and safety The design integrates infrared or ultrasonic sensors to monitor vehicle count and lane occupancy in real-time. A central microcontroller processes the sensor data and controls a series of motors or actuators that move the divider units along embedded tracks or wheels. Safety features such as emergency stop mechanisms, LED indicators, and obstacle detection ensure reliable and secure operation. The system is powered by either a mains supply or solar energy, promoting sustainability. This adaptable infrastructure can be especially beneficial in cities with varying traffic patterns throughout the day, such as near schools, business districts, or event venues. The project has the potential to replace traditional static dividers in metropolitan areas, offering a cost-effective, adaptive, and technologically advanced approach to traffic management

Effective urban commuter traffic monitoring is essential for planning sustainable transport systems in metropolitan regions. Traditional methods, relying on labor-intensive on-site surveys, are often limited by small sample sizes and infrequent data collection. This study introduces a novel data fusion framework that integrates multiple data sources, including automated vehicle counts, manual occupancy counts, floating phone data, and public transport passenger counts. By merging these data sources, the framework enhances data accuracy and reliability, allowing for the calculation of detailed origin–destination matrices and the modal split across various cordon cross-sections at a cordon line. Applied in the Austrian Vienna region over two consecutive years, the framework demonstrated its ability to reduce survey costs, improve data reproducibility, and therefore enable timely evaluation of transport policies and infrastructure projects. The findings revealed a correlation between public transport accessibility and modal split. Furthermore, it was observed that longer travel distances were strongly associated with increased public transport usage, particularly along major rail lines. The presented approach offers valuable insights for urban transport planning, which can be used for data-driven subsidizing of public transport based on patronage and residency, optimizing timetables, and planning metro extensions to meet future mobility needs. The framework’s flexible design and reliance on widely available data ensure its transferability to other metropolitan regions, promoting the development of sustainable transportation strategies.

The vehicle routing problem with time windows is an NP-hard optimization problem vital to logistics and supply chain management. It involves optimizing vehicle routes to serve customers within time windows and capacity limits. This study proposes a hybrid genetic algorithm combining a nearest neighbor-based initialization with advanced mutation operators. The nearest neighbor method ensures high-quality initial solutions by prioritizing proximity and constraints, while multiple mutation operators enhance exploration and exploitation. Tested on the Solomon 100-customer dataset, NN-IHGA outperformed benchmarks, especially on random and mixed datasets, reducing travel costs and vehicle counts. Results highlight NN-IHGA's robustness and adaptability, offering a practical solution for real-world logistics optimization.

Vehicle counting is a crucial method used in traffic management. Computer vision can be employed for efficient detection and classification techniques for vehicle objects. This paper reports on a simultaneous process of vehicle classification and counting implemented on NVIDIA Jetson Nano. The use of YOLOv5 overcomes computational load issues in edge computing deployments, whereas its combination with the DeepSORT tracker algorithm enhances the accuracy of vehicle detection and counting in various directions. A total of 18200 images are used to train the detectors that are designed to target local vehicles. The average accuracy of the model for detecting cars, motorcycles, buses, and trucks is 72.1%, 21.56%, 70%, and 25.63%, respectively. Real-time tests obtained an overall average vehicle counting accuracy of 49.95%.

Atmospheric microplastics (AMPs) have become a growing concern in recent years, although research remains limited. This study investigated AMPs in Bandar Lampung City, Indonesia, by roadside particulate sampling using a High-Volume Air Sampler (HVAS) over eight hours in industrial zones, residential areas, busy roads, and city centers. AMPs were identified through visual analysis for their abundance and physical characteristics. Certain samples were further examined with Raman spectroscopy. Total Suspended Particulate (TSP) levels ranged from 16.96 to 427.8 μg/m³, with the highest concentrations in industrial areas. Microplastic concentrations ranged from 0.0021 to 0.0199 particles/m³, with fibrous microplastics most common. Blue and grey (faded black) microplastics were the most prevalent, with particles between 500-1000 µm making up 42% of the total. Raman analysis detected Polyethylene terephthalate (PET). In S4 (city center), the highest vehicle count was 3,388±270 vehicles/day, while S2 (residential area) recorded the lowest at 1,166±99 vehicles/day. No significant relationship was found between TSP levels, microplastic concentrations, or vehicle numbers. However, Northern area may be potential sources of AMPs along traffic flow.

In 1961, Los Angeles experienced the disastrous Bel Air fire, which swept through an affluent neighborhood situated in a hilly, WUI (wildland–urban interface) location. In January 2025, the city was devastated again by a nearly-simultaneous series of wildfires, the most severe of which took place close to the 1961 fire location. Disastrous WUI fires are, unfortunately, an anticipatable occurrence in many U.S. cities. A number of issues identified earlier remained the same. Some were largely solved, while other new ones have emerged. The paper examines the Palisades Fire of January, 2025 in this context. In the intervening decades, the population of the city grew substantially. But firefighting resources did not keep pace. Very likely, the single-most-important factor in causing the 2025 disasters is that the Los Angeles Fire Department operational vehicle count shrank to 1/5 of what it was in 1961 (per capita). This is likely why critical delays were experienced in the initial attack on the Palisades Fire, leading to a runaway conflagration. Two other crucial issues were the management of vegetation and the adequacy of water supplies. On both these issues, the Palisades Fire revealed serious problems. A problem which arose after 1961 involves the unintended consequences of environmental legislation. Communities will continue to be devastated by wildfires unless adequate vegetation management is accomplished. Yet, environmental regulations are focused on maintaining the status quo, often making vegetation management difficult or ineffective. House survival during a wildfire is strongly affected by whether good vegetation management practices and good building practices (“ignition-resistant” construction features) have been implemented. The latter have not been mandatory for housing built prior to 2008, and the vast majority of houses in the area predated such building code requirements. California has also suffered from a highly counterproductive stance on insurance regulation. This has resulted in some residents not having property insurance, due to the inhospitable operating conditions for insurance firms in the state. Because of the historical precedent, the details in this paper focus on the Palisades Fire; however, many of the lessons learned apply to managing fires in all WUI areas. Policy recommendations are offered, which could help to reduce the potential for future conflagrations.

Road infrastructure is an important element that must be well planned where in its implementation it is faced with limited resources. Therefore, traffic surveys conducted for a full year to obtain the Average Annual Daily Traffic (AADT) value used in road planning and design are difficult to obtain. To overcome these obstacles, the government has set the number of survey days for two days and seven days depending on its designation. This study aims to determine the error in estimating the Average Daily Traffic (ADT) based on the number of survey days. Data were obtained through field surveys using an automatic vehicle counter with a tool error value of 10%. The errors calculated include MAE, MAPE, RMSE, and Bias. The results of the study show that the number of survey days affects the magnitude of the error and the range of ADT Bias in predicting the AADT value. The number of survey days for two days, three days, and seven days showed good predictions. The Bias Range as a correction factor for the estimated AADT value is influenced by traffic conditions during the year. The correction factor with 95% CI based on the recommended research results for Road Type 4/2T or 2/1 is in the interval (1.40;-1.59) for a 2-day survey, and (3.92;1.17) for a 7-day survey.

IntroductionAccurate vehicle analysis from aerial imagery has become increasingly vital for emerging technologies and public service applications such as intelligent traffic management, urban planning, autonomous navigation, and military surveillance. However, analyzing UAV-captured video poses several inherent challenges, such as the small size of target vehicles, occlusions, cluttered urban backgrounds, motion blur, and fluctuating lighting conditions which hinder the accuracy and consistency of conventional perception systems. To address these complexities, our research proposes a fully end-to-end deep learning–driven perception pipeline specifically optimized for UAV-based traffic monitoring. The proposed framwork integrates multiple advanced modules: RetinexNet for preprocessing, segmentation using HRNet to preserve high-resolution semantic information, and vehicle detection using the YOLOv11 framework. Deep SORT is employed for efficient vehicle tracking, while CSRNet facilitates high-density vehicle counting. LSTM networks are integrated to predict vehicle trajectories based on temporal patterns, and a combination of DenseNet and SuperPoint is utilized for robust feature extraction. Finally, classification is performed using Vision Transformers (ViTs), leveraging attention mechanisms to ensure accurate recognition across diverse categories. The modular yet unified architecture is designed to handle spatiotemporal dynamics, making it suitable for real-time deployment in diverse UAV platforms.MethodThe framework suggests using today’s best neural networks that are made to solve different problems in aerial vehicle analysis. RetinexNet is used in preprocessing to make the lighting of each input frame consistent. Using HRNet for semantic segmentation allows for accurate splitting between vehicles and their surroundings. YOLOv11 provides high precision and quick vehicle detection and Deep SORT allows reliable tracking without losing track of individual cars. CSRNet are used for vehicle counting that is unaffected by obstacles or traffic jams. LSTM models capture how a car moves in time to forecast future positions. Combining DenseNet and SuperPoint embeddings that were improved with an AutoEncoder is done during feature extraction. In the end, using an attention function, Vision Transformer-based models classify vehicles seen from above. Every part of the system is developed and included to give the improved performance when the UAV is being used in real life.ResultsOur proposed framework significantly improves the accuracy, reliability, and efficiency of vehicle analysis from UAV imagery. Our pipeline was rigorously evaluated on two famous datasets, AU-AIR and Roundabout. On the AU-AIR dataset, the system achieved a detection accuracy of 97.8%, a tracking accuracy of 96.5%, and a classification accuracy of 98.4%. Similarly, on the Roundabout dataset, it reached 96.9% detection accuracy, 94.4% tracking accuracy, and 97.7% classification accuracy. These results surpass previous benchmarks, demonstrating the system’s robust performance across diverse aerial traffic scenarios. The integration of advanced models, YOLOv11 for detection, HRNet for segmentation, Deep SORT for tracking, CSRNet for counting, LSTM for trajectory prediction, and Vision Transformers for classification enables the framework to maintain high accuracy even under challenging conditions like occlusion, variable lighting, and scale variations.DiscussionThe outcomes show that the chosen deep learning system is powerful enough to deal with the challenges of aerial vehicle analysis and gives reliable and precise results in all the aforementioned tasks. Combining several advanced models ensures that the system works smoothly even when dealing with problems like people being covered up and varying sizes.

This paper presents a generalised framework to model the interactions between human-driven vehicles (HVs) and connected autonomous vehicles (CAVs) in multi-lane settings using Lagrangian coordinates. The framework integrates lane-specific fundamental diagrams and conservation laws considering lane-changing behaviour, available in both continuous and discrete formulations. A novel method estimates net lane-changing rates by balancing demand and capacity across lanes, enabling dynamic calculation of vehicle group spacing and simulating both longitudinal and lateral dynamics. Validation includes numerical experiments involving accidents and lane drops, and one real-world trajectory dataset. Results show the model effectively captures the emergence, propagation, and dissipation of congestion, reproduces capacity drop, and provides both macroscopic flow characteristics and microscopic vehicle group dynamics. Compared to Lagrangian single-lane and Eulerian multi-lane models, the proposed framework better captures multi-lane mixed traffic complexities and reduces numerical diffusion. Real-world validation confirms its accuracy in estimating vehicle counts, spacing, and lane-changing behavior.

ABSTRACTCrowd counting plays a crucial role in analyzing group behavior in smart cities. Traditional crowd‐counting models rely on large datasets gathered from diverse individuals for training while ignoring the privacy protection for each training client. Meanwhile, the scale variation has long been a difficult problem in crowd counting and has greatly reduced model accuracy. Therefore, it is essential to achieve privacy‐aware crowd counting and to solve the problem of scale variation in dense scenes. To this end, we propose a Privacy‐preserving Quantum‐enhanced Network (PQNet). The PQNet uses federated learning to share parameters rather than data, which ensures the privacy of each client. Subsequently, a multi‐scale quantum‐driven calibration module is designed to capture multi‐scale information via quantum states. It enhances counting accuracy in dense crowd environments where scale varies. Experiments on four crowd counting and two vehicle counting benchmarks demonstrate that PQNet outperforms state‐of‐the‐art methods subjectively and objectively. The code will be available at: https://github.com/sdutzhangchen/PQNet.

Abstract. This study develops an integrated framework combining computer vision and traffic simulation for optimizing traffic management in high-density urban commercial areas. The framework employs a two-phase data acquisition approach: UAV-captured video is first processed using an enhanced YOLOv10 algorithm to identify critical road segments and key intersections, followed by handheld video recordings at target intersections for extracting dynamic traffic parameters (including vehicle counts, speeds, pedestrian density, and mixed-traffic interactions). The proposed framework was applied to the Xianlie East Road-Lianquan Road intersection in Guangzhou, a conflict-prone hub adjacent to densely clustered garment wholesale markets. Key improvement measures evaluated include crosswalk relocation and non-motorized lane adjustments. Additionally, the signal cycle duration was extended from 80 to 90 seconds to alleviate phase transition conflicts. Simulation inputs integrate field-observed behavioral patterns (e.g., 83% pedestrian compliance with signals). Results demonstrate significant improvements: 16.7% reduction in eastbound queue lengths (240.23m), 80.4% decrease in vehicle delays (56.46s), 44.8% shorter travel times (51.61s), and 83% fewer pedestrian-vehicle conflicts. This approach provides a scalable technical pathway for adaptive traffic governance in complex urban environments.

The increasing adoption of smart mobility and connected vehicles necessitates significant improvements in underlying infrastructure, particularly in real-time data processing and decision-making. Vehicular Edge Computing (VEC) has emerged as a vital solution by enabling computation closer to data sources, thereby reducing latency and reliance on centralized cloud systems. However, efficient allocation of edge resources (processing power, bandwidth, and storage) remains a critical challenge due to the highly dynamic, decentralized nature of vehicular networks. Traditional optimization techniques often fall short under these conditions. This study explores a quantum-inspired optimization framework designed to enhance resource management in VEC environments by leveraging principles of quantum computing such as superposition and probabilistic state selection within classical hardware. Extensive simulations involving 10 vehicles and 3 edge servers were conducted to evaluate the framework's performance. The dynamic resource demand fluctuated between 7 and 18 units, and server utilization ranged from 0.2% to 1.4%, illustrating diverse operational conditions. The proposed quantum-inspired model showed superior efficiency, achieving up to 35% improvement in fitness gain compared to traditional algorithms, with convergence to optimal fitness in just 45 iterations. The solution space was explored effectively using quantum state amplitude representations, which improved solution diversity and robustness in decision-making. Furthermore, fairness in resource distribution was evaluated using Jain’s Fairness Index, yielding a high score of 0.914, demonstrating equitable allocation among vehicles. Additional results revealed that task completion times ranged from 1.5 to 3.5 seconds, with processing delays being the major contributor. The system exhibited sublinear scalability, performing well up to 50 vehicles but declining as the vehicle count increased to 200, indicating a need for further optimization strategies. Although the model operates in a classical environment without quantum hardware, it offers substantial performance benefits. This research highlights the potential of quantum-inspired optimization for real-time, fair, and scalable resource management in vehicular networks. Future work should incorporate real-world vehicular trace data, expand scalability tests, and explore integration with 5G and energy harvesting mechanisms. These advancements will further support intelligent, secure, and sustainable transportation systems driven by edge computing technologies.

The growing dependence on oil tanker trucks for transporting petroleum products along Iraq’s International Highway has raised major concerns regarding traffic fluidity, road safety, and infrastructure resilience. This study investigates the specific impact of these heavy vehicles on traffic congestion, speed reduction, and road surface deterioration. The research aimed to: (1) analyze how oil tanker trucks affect traffic flow in terms of congestion, delays, and speed variation; (2) evaluate the capacity and condition of road infrastructure under heavy vehicle pressure; and (3) recommend effective traffic and infrastructure management strategies. A mixed-methods approach was employed, combining GPS tracking, ITS-based traffic monitoring, radar speed detection, and manual vehicle counts, along with surveys and interviews conducted with truck drivers, road users, and transport authorities. The results show that oil tanker trucks reduce the average vehicle speed from 75 km/h to 40–50 km/h, increase travel delays by up to 25%, and significantly accelerate pavement damage, particularly near toll booths and refueling stations. Stakeholder feedback revealed a consensus on the need for immediate interventions. The study recommends implementing dedicated truck lanes, time-based truck movement restrictions, and investment in intelligent transportation systems (ITS) to enhance traffic efficiency and minimize infrastructure wear. These findings offer vital insights for improving transport policy, road safety, and long-term highway sustainability in Iraq.