Traffic Conditions Research Articles

Changes in traffic carbon emissions with traffic conditions and their control strategies

Road traffic carbon dioxide emissions have a significant impact on the atmospheric environment, and changes in traffic conditions can affect these emissions; while changes in traffic state and traffic infarction can reflect traffic conditions, there are also differences in carbon dioxide emissions generated by vehicles operating on different fuels. Therefore, this paper sequentially determines quantitative indicators of traffic states and traffic infarction as well as macro control strategies to improve traffic carbon dioxide emissions. Furthermore, taking the confluence area as the object of analysis, the impact of traffic state and traffic infarction on carbon dioxide emissions is sequentially analysed with actual traffic flow data and statistical methods, and corresponding models are constructed. Meanwhile, based on the analysis results, mainline speed limit control strategies that consider the traffic state and traffic infarction, as well as the control strategy based on the combination of multiple fuel–vehicle combinations, are presented. The analysis results show that changes in both traffic states and traffic infarction can affect traffic carbon dioxide emissions from the mainline and the on-ramps, which can fundamentally affect traffic conditions. Moreover, total carbon dioxide emissions from a combination of gasoline and diesel minibuses decrease with an increase in the ratio of gasoline minibuses.

Grading of Traffic Interruptions in Highways to Tibet Based on the Entropy Weight-TOPSIS Method and Fuzzy C-Means Clustering Algorithm

The interruption of transportation on the way to Tibet has brought great losses to the Tibetan region. The work proposed a model that integrated the entropy weight-TOPSIS method with the fuzzy C-means clustering algorithm to discuss the causes and characteristics of traffic interruptions on the four main highways to Tibet. This approach aimed to quantify and grade traffic interruption states. First, the entropy weight-TOPSIS method was used to mitigate dimensions among various indices and quantitatively evaluate the status values of traffic interruptions. Then, the fuzzy C-means clustering algorithm was employed to grade these values. The proposed model graded traffic interruption states into four levels by evaluating the duration, mileage, and severity of traffic interruptions. Moreover, the four-level classification scheme can reflect the severity of traffic blocking events more precisely while maintaining a lower PE (Partition Entropy) value. In the four-level classification, the Sichuan–Tibet Highway and Xinjiang–Tibet Highway experienced more level-3 and level-4 serious interruptions, while most high-level interruptions on the Qinghai–Tibet Highway were classified as level-2 ordinary interruptions. The Yunnan–Tibet Highway, with limited data and primarily level-1 classification, was not analyzed in detail. These findings provide a reference for highway management departments to formulate targeted maintenance and emergency measures, especially the Sichuan–Tibet highway, which needs more attention and resource investment to improve its disaster resistance and reduce the impact of traffic interruptions.

Developmental changes in how pedestrians cross streets in single- vs. dual-lane traffic conditions.

Pedestrian injury is a leading contributor to childhood deaths. This study compared the crossing behaviors of children with adults when crossing in virtual single-lane vs. two-lane traffic conditions. Using a fully immersive virtual reality system, children 7-10 years and adults crossed streets with traffic in one lane (near lane) or in two lanes (near and far lane) with 0 offset so cars from opposing directions arrived at the child's crossing line at the same time. On each trial, pedestrians made a decision on when to cross and then completed the crossing, with measures of attention and crossing behaviors automatically recorded by the system as they did so. In comparison to crossing a single lane of traffic, crossing two-lane roads increased injury risk for all pedestrians, though children experienced significantly greater risk than adults. Children predominantly crossed by stopping before entering the far lane, whereas adults showed greater synchronization of self-movement to traffic flow and more often crossed both lanes without stopping. Children experience more high-risk outcomes than adults when crossing single-lane roads. Crossing two-lane roads elevates risk for pedestrians of both ages, though this risk is significantly greater for child than adult pedestrians. The predominant strategy used to cross two lanes of traffic shows significant developmental changes. Implications for injury prevention are discussed.

Research on robust decision making for intelligent connected vehicle at highway on-ramp

Reinforcement learning has demonstrated its potential in the decision-making field of autonomous driving. By engaging in trial-and-error learning and interacting with the environment, we can adapt our behavioral strategies based on reward and enhance driving decisions. Nevertheless, real-world vehicle decision-making involves intricate and diverse information, and the limited state information hinders agents from making optimal decisions, which may result in catastrophic consequences like collisions. Therefore, this paper proposes a robust deep reinforcement learning method based on the Soft Actor-Critic (SAC) algorithm for highway intelligent connected vehicle ramp merging decision-making. The model represents the vehicle features of the target lane and its adjacent lanes as an environmental state space. Additionally, a hybrid action space is designed, which combines discrete lateral actions and continuous longitudinal actions. Finally, an on-ramp simulation platform is built using actual roads and SUMO to verify the feasibility of the model. Multiple sets of comparative analysis experimental results demonstrate that the proposed method outperforms others in terms of the average reward return value, robustness value, collision rate, and success rate. Moreover, the model exhibits a stable lane change success rate under various road traffic density conditions, which indicates its robustness. The code can be obtained at https://github.com/ColinFanghz/sac-on-ramp.git .

Reinforcement Learning for Transit Signal Priority with Priority Factor

Public transportation has been identified as a viable solution to mitigate traffic congestion. Transit signal priority (TSP) control, which is widely used at signalized intersections, has been recognized as a practical strategy to improve the efficiency and reliability of bus operations. However, traditional TSP control may fall short of efficiency and is facing several challenges of negative externalities for non-transit users and the need to handle conflicting priority requests. Recent studies have proposed the use of reinforcement learning (RL) methods to identify efficient traffic signal control (TSC). Some of these studies on RL-based TSC have incorporated the concept of max-pressure (MP), which is a maximal weight-matching algorithm to minimize queue sizes. Nevertheless, the existing RL-based TSC methods focus on private vehicles and cannot adequately distinguish between buses and private vehicles. In prior research, RL-based control has been implemented within the context of bus rapid transit (BRT) systems. This study proposes a novel RL-based TSC strategy that leverages the MP concept and extends it to incorporate TSP control. This is the first implementation of RL-based TSP control within the mixed-traffic road network. A significant innovation of this research is the introduction of the priority factor (PF), which is designed to prioritize bus movements at signalized intersections. The proposed RL-based TSP with PF control seeks to balance the competing objectives of enhancing bus operations while mitigating adverse impacts on non-transit users. To evaluate the performance of the proposed TSP method with the PF mechanism, simulations were conducted on an arterial and a grid network under dynamic traffic conditions. The simulation results demonstrated that the proposed TSP with PF not only reduces bus travel times and resolves conflicts between priority requests but also does not make a significant negative impact on passenger car operations. Furthermore, the PF can be dynamically assigned according to the number of passengers on each bus, suggesting the potential for the proposed approach to be applied in various traffic management scenarios.

Enhancing Planning for Autonomous Driving via an Iterative Optimization Framework Incorporating Safety-Critical Trajectory Generation

Ensuring the safety of autonomous vehicles (AVs) in complex and high-risk traffic scenarios remains a critical unresolved challenge. Current AV planning methods exhibit limitations in generating robust driving trajectories that effectively avoid collisions, highlighting the urgent need for improved planning strategies to address these issues. This paper introduces a novel iterative optimization framework that incorporates safety-critical trajectory generation to enhance AV planning. The use of the HighD dataset, which is collected using the wide-area remote sensing capabilities of unmanned aerial vehicles (UAVs), is fundamental to the framework. Remote sensing enables large-scale real-time observation of traffic conditions, providing precise data on vehicle dynamics, road structures, and surrounding environments. To generate safety-critical trajectories, the decoder within the conditional variational auto-encoder (CVAE) is innovatively designed through a data-mechanism integration method, ensuring that these trajectories strictly adhere to vehicle kinematic constraints. Furthermore, two parallel CVAEs (Dual-CVAE) are trained collaboratively by a shared objective function to implicitly model the multi-vehicle interactions. Inspired by the concept of “learning to collide”, adversarial optimization is integrated into the Dual-CVAE (Adv. Dual-CVAE), facilitating efficient generation from normal to safety-critical trajectories. Building upon this, these generated trajectories are then incorporated into an iterative optimization framework, significantly enhancing the AV’s planning ability to avoid collisions. This framework decomposes the optimization process into stages, initially addressing normal trajectories and progressively tackling more safety-critical and collision trajectories. Finally, comparative case studies of enhancing AV planning are conducted and the simulation results demonstrate that the proposed method can efficiently enhance AV planning by generating safety-critical trajectories.

The problem of obtaining vehicles speed input for road traffic noise microscopic models: a comparison between field measurements and a stochastic approach

Mitigating environmental noise harmful effects is a relevant goal in the EU "Zero Pollution Action Plan", which targets a 30% reduction in chronic annoyance from transportation noise by 2030. Since road traffic is a major noise source in urban and non-urban settings, much effort has been invested in analyzing and modeling it. To assess noise levels, numerous Road Traffic Noise Models (RTNMs) exist, each requiring multiple inputs like vehicle flows, percentage of heavy vehicles, and average speed. Specifically, when dealing with microscopic models, the single vehicle speeds are needed. To obtain them, on-site measurements offer accurate data but are time consuming and expensive. This paper explores two approaches: an alternative measurement approach, based on video processing and object detection tools, and a stochastic approach, which randomly assigns a speed to each vehicle from a distribution curve, depending on traffic conditions and vehicle category. These methodologies are used to provide input data to a road traffic noise microscopic model, whose results are compared with field measured data. This comparison will allow to get useful insights on the performances of the proposed techniques in providing reliable inputs and their potential applications in different scenarios where single vehicle measured speeds are not available.

Relation between measured noise levels and traffic flow

Road-traffic flow parameters are a critical input to urban noise pollution evaluation frameworks such as CNOSSOS-EU. The accuracy and resolution of the traffic flow improves the quality of the noise exposure assessment. Highly resolved traffic flow within an urban network is difficult to reliably measure or predict. The quality and availability of such traffic information may be improved using data from low-cost noise sensors that are mounted beside representative segments of the road network. A model is presented, which relates the noise levels measured at a single noise sensor with the local traffic flow conditions. The noise data is obtained from a noise sensor mounted alongside a road, which is part of a test-bed in Stockholm, Sweden. The local traffic flow is obtained from commercial vehicle-counting sensors mounted alongside the noise sensor. Effect of seasonality is observed when comparing models trained on different subsets of the training data. The effect of seasonality on the noise levels is investigated.

Road target recognition based on radar range-Doppler spectrum with GS – ResNet

ABSTRACT Road target recognition ability of automotive millimetre-wave (mmWave) radar is crucial in areas such as autonomous driving, advanced driver assistance systems (ADAS), and automated emergency braking systems (AEBS). Current mmWave radar systems are primarily utilized for speed, distance, and angle measurements, with limited capability or unsatisfactory performance in target classification. In response to this limitation, this paper proposes a novel method for road target recognition, using radar range-Doppler spectrum as input data and introducing a global attention mechanism and SCConv module. First, the range-Doppler spectrum is processed by the clutter removal algorithm block to remove background clutter and vibration interference. Second, considering the small features and susceptibility to noise interference in the range-Doppler spectrum, we have devised a deep residual network based on a global attention mechanism, significantly enhancing the model’s accuracy in range-Doppler spectrum classification. Finally, we introduce the SCConv module and improve the downsampling module in ResNet to enhance the model’s classification accuracy. Experimental results demonstrate that the model achieves an average accuracy of 99.79% in classifying six types of road targets, significantly outperforming other methods. This research is of significant importance in advancing the understanding of road traffic conditions by autonomous driving systems and enhancing system safety.

Anticipated buffer time – An evasive surrogate safety indicator for risk assessment of unsignalized intersections under heterogeneous traffic and aggressive driving conditions

Risk assessment of unsignalized intersections is particularly challenging when confronted with a combination of factors such as heavy traffic, diverse vehicle types, lane indiscipline, aggressive driving, and evasive manoeuvres. Understanding how people drive in these situations is crucial for accurately assessing the risks at unsignalized intersections. This study introduces a novel surrogate safety indicator, i.e. Anticipated Buffer Time (ABT), designed to account for these various factors. Additionally, three new indicators derived from ABT are introduced, namely ABT Negation Ratio, ABT Extremity Ratio, and ABT Progression Ratio. A risk assessment measure, denoted as UnSigRisk Score, is formulated using these three indicators for unsignalized intersections. Three intersections in Ahmedabad, India, were selected for the study due to their manifestation of these challenging conditions. Spearman Rank Correlation Coefficient was estimated to find out how well can UnSigRisk Score measure is able to quantify evasive behaviour. The results indicate that this score proficiently measures evasive behaviour, exhibiting coefficients exceeding 0.6 in all cases—significantly outperforming the current evasive indicators, Yaw Rate Ratio and Jerk. The proposed risk assessment score could serve as a practical tool for transportation authorities, enabling them to identify the most vulnerable intersections and allocate resources for targeted safety interventions wisely. The study unequivocally demonstrates that the use of ABT paves the way for a thorough examination of safety at unsignalized intersections, regardless of driving behaviour and traffic conditions.

An Algorithm for Predicting Vehicle Behavior in High-Speed Scenes Using Visual and Dynamic Graphical Neural Network Inference

Accidents caused by vehicles changing lanes occur frequently on highways. Moreover, frequent lane changes can severely impact traffic flow during peak commuting hours and on busy roads. A novel framework based on a multi-relational graph convolutional network (MR-GCN) is herein proposed to address these challenges. First, a dynamic multilevel relational graph was designed to describe interactions between vehicles and road objects at different spatio-temporal granularities, with real-time updates to edge weights to enhance understanding of complex traffic scenarios. Second, an improved spatio-temporal interaction graph generation method was introduced, focusing on spatio-temporal variations and capturing complex interaction patterns to enhance prediction accuracy and adaptability. Finally, by integrating a dynamic multi-relational graph convolutional network (DMR-GCN) with dynamic scene sensing and interaction learning mechanisms, the framework enables real-time updates of complex vehicle relationships, thereby improving behavior prediction’s accuracy and real-time performance. Experimental validation on multiple benchmark datasets, including KITTI, Apollo, and Indian, showed that our algorithmic framework achieves significant performance improvements in vehicle behavior prediction tasks, with Map, Recall, and F1 scores reaching 90%, 88%, and 89%, respectively, outperforming existing algorithms. Additionally, the model achieved a Map of 91%, a Recall of 89%, and an F1 score of 90% under congested road conditions in a self-collected high-speed traffic scenario dataset, further demonstrating its robustness and adaptability in high-speed traffic conditions. These results show that the proposed model is highly practical and stable in real-world applications such as traffic control systems and self-driving vehicles, providing strong support for efficient vehicle behavior prediction.

Assessing economic efficiency of driving behavior within traffic flow dynamics

The widespread adoption of electric vehicles has a positive impact on energy conservation and emission reduction, significantly contributing to sustainable development. This paper proposes an economic evaluation method for drivers’ driving behaviors based on actual operational data of pure electric vehicles. This method aims to eliminate the influence of traffic conditions on the evaluation results, ensuring that the outcomes solely reflect the economic efficiency level of drivers’ behaviors. Initially, short trips are identified and traffic conditions are classified and identified based on real vehicle data. A random forest identification model is established to recognize the traffic conditions of short trips in real-time. Subsequently, characteristic variables are selected and calculated. Pearson correlation analysis wand Spearman correlation analysis are utilized to identify variables strongly correlated with energy consumption and traffic conditions, which are then used as evaluation indicators. Finally, eco-scores are calculated using the energy consumption per unit mileage for short trip segments under different traffic conditions. A neural network model is established and trained between the evaluation indicators and eco-scores, enabling real-time evaluation of the economic efficiency of drivers’ driving behaviors. The evaluation method proposed in this paper can objectively assess the economic efficiency of driving behavior, facilitating drivers to improve their driving practices based on their evaluation results.

Self-adaptive equation embedded neural networks for traffic flow state estimation with sparse data

The data-driven approach in intelligent traffic systems has achieved successive breakthroughs, thanks to the ever-increasing volume of traffic data. Nonetheless, in practical scenarios, the collected data often contain some issues, e.g., missing values, significantly impacting the accuracy and efficiency of the algorithms. To enhance the precision of traffic estimation utilizing the sparse data, we have developed a physics-informed neural network (PINN) based algorithm in the line with the traffic flow theory and deep learning principles. In contrast to the conventional PINNs, our approach uniquely incorporates a self-adaptive macro model for mixed flow into the network's architecture, serving as an embedded source of physics information. With this algorithm, we can capture the dynamic behavior of an entire traffic flow including its spatiotemporal evolution with sparse traffic data such as initial and boundary value information. To realize the model's adaptability, we have revised the macro model by inverting its parameters and incorporating a data-driven nonlinear element, which simplifies the intricate macro model structure. The network's effectiveness has been validated through the experiments conducted on a mixed traffic flow system experiencing local agglomeration and real-world data, demonstrating its capability for precise traffic simulation, efficient traffic flow prediction, and interpretability. Our study offers a novel insight for data-driven traffic flow state estimation.

Hybrid vehicular access protocol and message prioritization for real-time safety messaging

Real-time safety services rely on the exchange of messages to enhance the operations of connected and automated vehicles (CAVs). These safety messages convey vital information about traffic conditions, enabling drivers to take necessary measures to prevent accidents. The timely and reliable delivery of these messages is essential, necessitating efficient channel access. Vehicular Deterministic Access (VDA) is employed as a channel access scheme with distinct priorities and stringent timing guidelines, particularly for urgent safety warnings. In this paper, we propose a hybrid approach that combines VDA and Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) protocols, along with a message prioritization algorithm, to ensure efficient and reliable communication of safety messages in vehicular networks. Our approach leverages the strengths of both VDA and CSMA/CA to avoid message collisions; VDA is more efficient under high traffic loads, while CSMA/CA is better suited for low traffic loads. Additionally, the incorporation of the message prioritization algorithm ensures strict deadline guarantees for high-priority messages, such as Decentralized Environmental Notification Messages (DENMs). We evaluate our proposed solution using the Artery simulation framework. Our results show over a 93% delivery rate for DENM exchanges while maintaining low collision probability across various traffic loads. This research provides practical guidance for the development of efficient and reliable communication systems for CAVs. It also offers a detailed analysis of the trade-offs among different access protocols and message prioritization strategies in vehicular networks.

Digital traffic state analysis for urban regions considering complex multi-directional flow changes

In this research, a digital analytical model based on three-flow full-cycle state macroscopic fundamental diagram (MFD) with observable different perimeter control management and flow loading modes has been proposed on the basis of real network and peak hour traffic flow loading scenarios in Rostov-on-Don. It is known that the aim of digital management of urban intelligent traffic is to ultimately optimize the overall state of multi-area traffic operation in the city. The goal of this research was to provide a strong basis to establish digital management strategies for intelligent transport, determine the change trends of steady-state equilibrium of regional traffic flow and their effects on full-cycle states of traffic flow under various peripheral control management strategies and traffic loading patterns. Simulation results confirmed that with the evolution of traffic state, stricter peripheral control management always results in a wider attraction area ultimately giving rise to steady-state equilibrium of traffic flow.

Cybersecurity in a Scalable Smart City Framework Using Blockchain and Federated Learning for Internet of Things (IoT)

Smart cities increasingly rely on the Internet of Things (IoT) to enhance infrastructure and public services. However, many existing IoT frameworks face challenges related to security, privacy, scalability, efficiency, and low latency. This paper introduces the Blockchain and Federated Learning for IoT (BFLIoT) framework as a solution to these issues. In the proposed method, the framework first collects real-time data, such as traffic flow and environmental conditions, then normalizes, encrypts, and securely stores it on a blockchain to ensure tamper-proof data management. In the second phase, the Data Authorization Center (DAC) uses advanced cryptographic techniques to manage secure data access and control through key generation. Additionally, edge computing devices process data locally, reducing the load on central servers, while federated learning enables distributed model training, ensuring data privacy. This approach provides a scalable, secure, efficient, and low-latency solution for IoT applications in smart cities. A comprehensive security proof demonstrates BFLIoT’s resilience against advanced cyber threats, while performance simulations validate its effectiveness, showing significant improvements in throughput, reliability, energy efficiency, and reduced delay for smart city applications.

A Life Cycle Cost Comparison of Flexible Versus Rigidpavements in Nigeria: Case Study of Papalanto - Sagamu Road

The condition of roads in Nigeria, most of which are made from asphalt, is generally considered unsatisfactory. This has led to calls for a blanket switch from asphalt pavements to concrete pavements for new road construction. This study examines the characteristics and relative advantages of the two pavement types; and carried out the design and life cycle cost comparison of a case study road with heavy traffic and poor subgrade conditions. The results indicate that, the initial cost of construction is actually less for the concrete pavement by 12% when compared with the asphalt alternative, while the life cycle cost analysis further yielded an advantage of 36% for the concrete pavement. The paper therefore recommends that major roads, particularly those on the national and international transit corridor, being of heavy traffic, should be reconstructed as the need arises, as concrete pavements. The paper however further recommends that, axle load control should be enforced through the introduction of weigh bridges and that routine maintenance should mandatorily be carried out, as even the best constructed concrete pavements have failed due to lack of routine maintenance like clearing the road verges of weed and road sand and desilting of drainage structures

A quadratically constrained mixed-integer non-linear programming model for multiple sink distributions

Rising traffic congestion and fuel costs pose significant challenges for supply chains with numerous retailers. This paper addresses these challenges by optimizing transportation routes for processed tomatoes within a long-haul and intercity distribution network. We use the heterogeneous capacitated vehicle routing problem framework to create a new quadratically constrained mixed-integer non-linear programming model that aims to meet demand at multiple destinations while minimizing transportation costs. Our model incorporates real-time data and route optimization strategies that consider traffic conditions based on freight time and route diversions for expedited deliveries. It aims to devise an optimal transportation schedule that minimizes fuel, operational, and maintenance costs while ensuring efficient delivery of tomato paste. By applying this model to a real-world case study, we estimate a significant 27.59% reduction in transportation costs, dropping them from GH¢20,270 ($1,638.91) to GH¢14,676 ($1,186.61) on average. This highlights the effectiveness of our strategy in lowering costs while maintaining smooth and efficient deliveries.

In-tunnel pollutant concentration measurement and ventilation control indexes for highway tunnels in mountainous area: A case study of no.1 Qinling tunnel, China

The pollutant concentration and hygiene standards in mountain tunnels are key factors determining the ventilation effectiveness of tunnel operation. Previous studies have mainly focused on the spatiotemporal distribution patterns of in-tunnel pollutants, while research on ventilation control indexes for pollutants in long mountain tunnels are relatively limited. Therefore, this paper aims to discussing the hygiene control indexes for operational ventilation of mountain tunnels. Field tests were conducted in the No.1 Qinling Tunnel of Xi'an to Han zhong Highway in China, and the distribution laws of traffic flow, wind speed, and concentrations of particulate matter (PM), carbon monoxide (CO), and nitrogen oxide (NOX) were on-site measured. The spatiotemporal distribution characteristics of pollutants in the long mountain tunnel were revealed, and the ventilation hygiene control indexes for harmful gases were calculated and compared to those of urban highway tunnels. The results show that the traffic volume of operational highway tunnels in mountainous areas of China is increasing rapidly, and the particulate matter emitted from diesel vehicle exhaust will pose a threat to the normal environmental quality in the tunnel. The average value of CO base emission factor in 2000 for the upline of No.1 Qinling tunnel was 0.0019 m3/(veh·km) by back-calculation, and the annual decline rate of CO in the ventilation design for mountain highway tunnels should be 6 %–8 % and revised every 8 years. It is recommended to consider the geographical location and traffic conditions of mountain tunnels when calculating the air demand required for diluting NO2. NOx has basically replaced CO as the primary pollutant with the most significant impact on human health in mountain tunnels. It is recommended that NOX be considered as the primary hygiene control indicator in the ventilation design of mountain tunnels in China. The pollutant distribution and ventilation control indexes of mountain and urban highway tunnels are quite different, which should be treated differently in engineering applications. The research results can provide a significant reference for ventilation design and environmental evaluation of highway tunnels in mountainous areas.

UAV Vehicle Detection System Based on YOLOv8

Abstract To utilize the drone vehicle detection technology to observe the road traffic conditions in real-time, thereby improving the operational effectiveness and safety of the traffic system, this paper adopts a vehicle target detection algorithm based on YOLOv8. YOLOv8 is the latest iteration developed by Ultralytics, which has further improvements and enhancements compared to YOLOv5. In the backbone network, YOLOv8 employs a C2f structure with richer gradient flow, boosting feature selection capabilities, and enhancing target detection performance and accuracy; In the detection head, YOLOv8 utilizes a Decoupled Head design and uses an improved Anchor-Free method to reduce computational complexity while improving detection efficiency; In terms of loss function, YOLOv8 utilizes the The Task-Aligned Assigner positive sample allocation strategy, and adds the Distribution Focal Loss to improve the model’s generalization ability. This paper divides the vehicles data set into 5 types, trains it using a pre-training model, and optimizes the network parameters through multiple iterations to achieve better detection performance. The experimental results indicate that YOLOv8 reaches the mAP of over 97% on the experimental data set, which is an improvement of approximately 8% compared to YOLOv5, and achieves a good detection performance, and is highly practical.