This PDF file contains the front matter associated with SPIE Proceedings Volume 12989, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

The complexity and variability of traffic systems have made traffic prediction a highly sought-after research area. To enhance the accuracy of predicting traffic flow, we propose a Graph Convolution and Multi-Channel Neural Network (GCMNN) model that leverages spatio-temporal dynamic correlation in traffic data. Firstly, we use a graph convolutional network to extract the spatial structure of the traffic network and combine it with spatio-temporal correlation to derive dynamic correlation features of traffic conditions. Secondly, we employ an A-component comprising multiple onedimensional convolutional neural networks to learn how different combinations of past traffic conditions impact future traffic conditions. We then predict temporal embedding attention, extracting spatio-temporal dynamic correlation of the traffic road network by calculating weighted features. The outputs from the spatial and temporal modules are fused using a gated fusion mechanism, and final prediction results are obtained through linear layer mapping. We performed traffic prediction experiments using highway traffic data, the results showed that GCMNN was superior to other baseline. Furthermore, this model can be applied to short-term urban air quality prediction and other spatio-temporal predictions, providing valuable insights for government management and improving people's lives.

In recent years, with the rapid development of urban distribution and the upgrading of people's consumption pattern, the volume of urban logistics distribution and distribution vehicles have soared, which has brought about urban traffic congestion and environmental pollution problems. In the face of these problems, electric logistics vehicles due to its green characteristics such as low energy consumption and low pollution can largely improve the urban distribution of "external uneconomical" problem. For this reason, many logistics companies replace the traditional fuel vehicles by electric logistics vehicles. In this context, this paper establishes a vehicle path optimization model that comprehensively considers the charging cost, driving range, customer time window and other factors of the electric logistics vehicle, and uses the improved genetic algorithm to solve the problem through examples, then concludes that the partial charging strategy is better than the complete charging strategy.

Flooding is a typical sudden and widespread disaster, which requires a large amount of emergency materials and rapid transportation capacity to control the spread of the disaster. In order to make the distribution of emergency materials under flooding more scientific and reasonable, and to improve the rescue efficiency, this paper studies the prediction and distribution method of emergency materials under large flood disasters, extracts seven relevant factors affecting the disaster relocation population, and establishes the relationship model between the number of disaster relocation population and the number of emergency materials. Meanwhile, with the goal of minimizing the transportation time turnover of the distribution centers and maximizing the demand satisfaction rate of the disaster sites, we construct a large-scale flood emergency materials distribution model with multiple distribution centers, multiple disaster sites and multiple materials. And the model can be verified to achieve the minimum demand satisfaction rate of emergency materials at each disaster site.

Aiming at the strong correlation between the hazards of the autonomous driving system of intelligent networked vehicles and the operating environment and control action, this paper proposes a method for identifying hazards of the expected functional safety of autonomous driving based on a mealy state machine. Firstly, the vehicle action rule base is determined by the operating environment of the vehicle and the available control action; secondly, the mapping between the mealy state machine and the vehicle operating state is established to simulate the vehicle state operation logic in the real scene; finally, by identifying the relationship between the vehicle state and the Conflicts between the operating environment and control actions identify potential hazards. In order to verify the effectiveness of the proposed method, this paper identifies 257 potential hazards through the identification of the expected functional safety hazards of the ACC system of a L2 vehicle, which is more efficient and convenient compared with the traditional STPA method.

To address the significant impact of different driving paths on vehicle safety and comfort during the driving process of autonomous vehicles, a polynomial function based obstacle avoidance path planning method is proposed. Using the vehicle states at the starting and ending points of the path as boundary conditions, the obstacle avoidance path equation is derived based on polynomial curves. The LQR lateral control algorithm and dual PID longitudinal control algorithm were used to verify the planned path on a real vehicle. The results of the real vehicle test showed that the obstacle avoidance path planning method based on polynomial function has good executable performance, safety, and comfort.

Due to the increasingly severe global energy shortage and the growing awareness of environmental protection, electric vehicles (EVs) have gained attention in the field of fresh food logistics. EVs are favored for their energy efficiency, low carbon emissions, environmental friendliness, and green characteristics. Additionally, the Chinese government's strong support for the new energy industry has further promoted the adoption of EVs in logistics transportation. However, EVs face challenges such as limited driving range, inconvenient charging, and potential issues related to cargo damage and refrigeration for fresh food products. These factors contribute to higher operating costs compared to traditional fuel-powered vehicles. With the improvement of people's living standards in China, there has been an increasing demand for fresh food products and higher expectations for the satisfaction of fresh food cold chain logistics services. Therefore, for purpose of reduce distribution costs of EVs while considering customer satisfaction, this paper focuses on the optimization of routes for electric cold chain logistics vehicles with time windows, taking into account the objective functions of driving costs, time window costs, vehicle utilization costs, customer satisfaction costs during the delivery process, refrigeration costs for fresh food cold chain products, and cargo damage costs during the delivery process. A mathematical model is established, and a modified genetic algorithm is used to analyze the model. Finally, through sensitivity analysis, it is found that increasing the number of delivery vehicles or improving the battery capacity and range of EVs can reduce the total cost of cold chain logistics distribution and improve customer satisfaction.

Electric bicycles, motorcycles, bicycles and other two-wheelers have incomparable convenience in short trips, but as vulnerable road users, the probability of accidents in complex traffic environments is higher than that of other vehicles. Based on the analysis of CIDAS accident data, this paper calculates the types of traffic participants in traffic accidents in which the passenger car is the main vehicle, and analyzes the two common traffic accident scenarios that are easy to cause injury to the two-wheeled vehicle. The basic principle, test scenario, test method and evaluation method of two ADAS functions, rear traffic crossing warning and door opening warning, are introduced according to the new C-NCAP protocol of 2024 edition. And the validation test results of three vehicle models were selected for analysis, and the final conclusion was drawn.

The design of railway stations impacts transportation efficiency and production. Therefore, meeting the needs of the new era for station design is a concern for the railway industry. This paper proposes a new design method for intercity/urban railway stations to address the requirements of the CTCS2+ATO (China Train Control System Level 2 + Automatic Train Operation) train control system for public transportation operations, aiming to shorten yard length and improve transportation efficiency. First, we compare design principles of arrival/departure lines at home and abroad. Second, we analyze factors affecting the effective length of passenger dedicated line stations. Third, we analyze the technical conditions for the implementation of the CTCS2+ATO system, optimize various influencing parameters according to the design specifications while ensuring the safe operation of trains. Fourth, we establish a CTCS2+ATO system arrival and departure line effective length model under selected vehicle parameters and line parameters conditions, and provide examples, illustrations and explanations of the effective length. Finally, we verify the proposed station design scheme through the intercity/urban railway CTCS2+ATO system laboratory test platform. The experimental results show that under the selected parameters, the train is controlled by ATO, and after entering the station, it stops steadily and accurately without passing the outbound signal. This is consistent with the theory and can improve the transportation efficiency of railway passenger dedicated lines while shortening the platform distances. The research results can be used to guide the design of intercity/urban railway stations.

Short-term passenger flow forecast is an important task in urban rail transit operation. Emerging deep learning technologies are seen as an effective way to solve this problem. In this study, we propose a deep learning model called GCN-Conv2d, which combines graph convolutional networks and two-dimensional convolutional neural networks. Firstly, a GCN model is introduced to deal with three passenger flow patterns (recent, daily, and symmetrical). The GCN model can extract the spatio-temporal correlation and topological information within the entire network, and then the two-dimensional convolutional neural network can be applied to deeply integrate the passenger flow information, and the latter can extract the spatio-temporal features between different passenger flow patterns and between stations at different distances. Finally, a fully connected layer is used to output the results. The GCN-Conv2d model predicted the smart card data of Zhuzhou Intelligent rail express system at a time interval of 10 minutes. The results show that the error of the model in RMSE and MAE is smaller than that of the random forest model and the CNN model, which shows good performance. This study can provide important support for public transport operators to optimize the operation of urban rail transit and promote the intelligent operation of urban rail transit network.

Although offshore wind power engineering has significant economic and social benefits, it is relatively close to navigable waters and will become a permanent navigation obstacle in the waters after completion, which will have a significant impact on the navigation safety of nearby ships. It may even cause collisions between ships and wind power engineering maintenance ships or wind turbines, resulting in significant casualties and economic losses. In this paper, on the basis of analysing the factors affecting the navigation safety of the shipping lanes near offshore wind farms, we establish the index system for evaluating the navigation safety level of the shipping lanes around offshore wind farms, and use the coefficient of variation algorithm, the cloud model theory, and the physical element topology theory to construct a model for evaluating the navigation safety level of the shipping lanes around offshore wind farms. This study can provide a reference for future offshore wind farm siting research on the evaluation of navigational safety.

In response to the recent sluggish international trade demand and a year-on-year decrease in China's total import and export volume, a vulnerability assessment framework for container terminals under the new development pattern of "Double Circulation" has been proposed. By tracing the appearance of container terminal vulnerability and based on the three major characteristic elements of vulnerability principle, analyze and determine the influencing factors of container terminal vulnerability. Construct a vulnerability evaluation index system for container terminals from four aspects: personnel, machinery, environment, and management. Taking the Yangshan Phase IV intelligent container terminal as the research object, a comprehensive vulnerability assessment of the container terminal was conducted through DEMATEL-ISM and ANP-Fuzzy. The results indicate that the overall vulnerability level of the Yangshan Phase IV intelligent container terminal is relatively low, with a vulnerability level of 39%. Finally, based on the experimental results, suggestions for anti vulnerability measures of the Yangshan Phase IV intelligent container terminal are provided from three aspects: emergency risk management, collaborative management, and personnel management.

In order to address the problem of container allocation at the container terminal of a U-shaped terminal layout, taking into account the actual conditions of multi-container area storage at the terminal, a model for allocating multiple container areas for outbound containers is established with the objective of minimizing the rehandling time and IGV transport time during container loading and unloading operations. A variable neighborhood search algorithm, considering storage rules, is designed to solve the problem of container allocation in the multi-container area. The effectiveness of the model and algorithm is verified through numerical experiments, and a sensitivity analysis of the balance of operations between container areas is conducted. The experimental results demonstrate that considering the balance of operations between container areas can effectively reduce the rehandling time during container loading and unloading operations, then improving the efficiency of the container terminal.

Correct driving decisions are a prerequisite for safe driving. In this paper, we construct a lane-changing decision model considering driving style in order to provide drivers with more accurate and timely lane-changing decision interventions. Based on NGSIM data, two-step trajectory reconstruction technique is firstly used to deal with data outliers and noise, and then vehicle lane-changing trajectory extraction and phase classification are carried out according to the definition of lane-changing decision, and k-means clustering method is also used to realize driving style calibration and determine lane-changing decision variables based on following phase data. The final XGBoost lane-changing decision model was constructed, and the need to incorporate driving style features and the superiority of XGBoost over RF and SVM models were verified through comparative experiments.

In order to minimize the impact of typhoons on ship navigation at sea while ensuring the safety and cost-effectiveness of ship avoidance during typhoon encounters, a research has proposed a ship avoiding typhoon path based on an improved artificial potential field method. Firstly, introducing a gravitational distance threshold and repulsion between the ship and target points in the original function, and incorporating random noise in the resultant force, to improve the artificial potential field method. This aims to address the issues encountered in traditional artificial potential field methods, such as obstacle collision, unreachable targets, and getting stuck in local minima, thereby enhancing the efficiency of path planning. Next, using Python software to set up the simulation environment, and under the same parameter condition and obstacle information, the improved artificial potential field method is compared with the original method to verify the superiority of the improved algorithm. Finally, to further demonstrate the practicality of the improved artificial potential field method, the 2023 of Typhoon Damrey was chosen as the research object, and the path simulation of ship avoidance was carried out, a reasonable path with both safety and economy in the dynamic moving process of Typhoon is obtained. The results indicate that the application of the improved artificial potential field method enables ships to safely and efficiently avoid typhoons.

The multiple electric vehicles (EVs) charging scheduling problem is a challenging issue in the context of sustainable urban mobility. To address this complex problem, a mixed reward multi-agent deep deterministic policy gradient (MRMADDPG)- based method is proposed in this paper. The proposed MR-MADDPG provides a framework for multiple EVs to collaboratively and adaptively make charging decisions in a shared charging infrastructure. By learning and adapting from interactions with the charging environment, the MR-MADDPG empowers every EV within the fleet to make real-time charging decisions based on its local observation and eventually gets relatively low charging costs. This research contributes to the advancement of sustainable urban transportation by harnessing the capabilities of MRMADDPG, promoting efficient energy use and reducing charging cost of EV.

The operating time fluctuations of different types of bus routes vary greatly, causing difficulties in bus schedule and high carbon emission costs. Therefore, based on the mixed bus system of electric buses and fuel buses, the carbon emission is considered including fuel life cycle and the influence of speed, acceleration, and load. Taking the minimization of bus operation cost, passenger travel cost, reliability cost, and carbon emission cost as the objective function, the paper constructs a joint strategy optimization model for the speed regulation and limited-stop bus service considering the carbon emission of the fuel life cycle. The limited-stop bus station, the departure time, and the running speed of all buses are determined. In a real-world case study, the proposed model is solved by genetic algorithm. The results show that the joint strategy optimization model and the joint strategy optimization model considering the carbon emissions of the fuel life cycle can greatly reduce the total cost of the bus system. The former has a more prominent effect on reducing passenger travel cost and reliability cost, while the latter can significantly reduce the carbon emissions generated from the bus.

The road transportation of explosives has the characteristics of instability, professionalism and serious accident consequences. During the transportation process, due to the problems of vehicles’ failure and illegal transportation method, it is very easy to cause explosion accidents. The unsafe behavior of drivers is one of the important reasons for the danger of explosives transportation vehicles. In order to reduce casualties and economic losses, it is necessary to monitor the driving behavior of drivers during the transportation process. In view of the lack of driver behavior monitoring system for explosive transport vehicles at this stage, this study uses consumer-grade cameras to determine the deployment scheme, determines specific identification behavior and risk judgment criteria according to traffic regulations and the particularity of explosives, and establishes a special explosive transport vehicle driver behavior monitoring system. This explosive transport vehicle driver behavior monitoring system integrates the recognition and early warning functions of fatigue driving, smoking, answering and making phone calls, leaving the steering wheel with both hands into the same model, and realizes the real-time detection and early warning of explosive transport vehicle driver behavior.

Complex weather conditions have a significant impact on the safety of ship navigation, and there is currently a lack of effective means to assess the risks to ship stability caused by complex weather. Research on route planning can only simply avoid strong winds and obstacles, and cannot timely and effectively identify navigation risk areas caused by complex weather conditions. In response to the problem of identifying high-risk areas in complex weather conditions, this paper analyzes the factors that affect navigation safety and establishes a navigation risk area model with wind and waves, pitch and heave as constraint conditions. By setting navigation risk as the objective function and using the genetic algorithm to solve it, the optimal route is obtained. The experimental results show that the optimization model proposed in this paper can effectively avoid high-risk sea conditions and prevent pitch or heave from occurring, which is of great significance for improving ship navigation safety.

Aiming at the problems of large cost, serious energy consumption and low distribution efficiency in cold chain logistics distribution. In this paper, the cargo damage cost, distribution cost and distribution time of cold chain vehicles and drones are considered comprehensively, and the total transportation cost is minimized as the optimization target model. The model is solved by artificial bee colony and genetic algorithm. A sensitivity analysis is performed for the speed of cold chain vehicles and drones. The results show that the cost of cargo damage and the total distribution time will decrease with the increase of the speed of cold chain vehicles and drones, but the speed of cold chain vehicles affects the two to a lesser extent compared to the speed of drones. The conclusions can provide distribution programs for logistics companies as well as help the government in decision-making.

With the rapid development of urban rail transit systems in our country, the frequency of unexpected events has significantly increased. This poses a series of challenges for the networked operation and management of urban rail transit, especially the unpredictable and significant changes in passenger travel behavior during unexpected events. Therefore, studying the relevant issues of passenger flow distribution in urban rail transit during unexpected disruptions is of great theoretical and practical significance.
In this paper, we analyze the passenger flow distribution based on the truthful AFC data of Beijing's rail transit system during a selected week in October 2017. We calculate the Origin-Destination (OD) passenger flow during working days and extract the OD passenger flow during the morning peak period in Beijing, which is used for the following two researches:
(1) The normal situation passenger flow distribution. With the help of abundant historical data from the Automatic Fare Collection (AFC) system, the research establishes a passenger flow distribution model and algorithm based on the transformed OD matrix and the structural characteristics of the urban rail transit network, optimized through the MNL approach. This study serves as an experimental basis for passenger flow redistribution during emergency disruptions.
(2) The interrupted passenger flow redistribution algorithm. Building on the basis of normal passenger flow distribution, the study analyzes and compares the passenger flow volume in disrupted segments and the transfer passenger flow before and after the interruption. By understanding the patterns of changes in passenger flow distribution before and after disruptions, it provides valuable guidance for post-disruption passenger evacuation.

To reduce the operational risk of aids to navigation, we firstly analyze the four aspects of the management of aids to navigation equipment, operating personnel, operating equipment and operating environment, and construct a safety evaluation system for the risk of aids to navigation operation. Then we utilize the ANP network analysis method to evaluate the risk of aids to navigation operations. Based on the evaluation results, countermeasures to improve the safety of aids to navigation operations are proposed to provide reference for realizing safe aids to navigation operations.

Externality is one of the important issues in the field of transportation economics research. With the development of intelligent connected vehicles, the efficiency of public transportation continues to improve. But congestion problems cannot be completely avoided, and congestion delays increase travel costs. How to accurately calculate the delay time of vehicles and effectively quantify the economic value of delay time is one of the important issues that need to be solved in the future development process of intelligent transportation systems. In order to effectively quantify the external costs of traffic congestion in the context of intelligent networking, this paper proposes a method for calculating the value of delay of intelligent connected passenger vehicles based on vehicle road collaboration data. By comparing the difference between the time when intelligent connected passenger vehicles pass through the starting and ending points of the road link at actual speed and the time when they travel at the speed limit of the road link, the delay time can be calculated based on the number of passengers carried by the passenger vehicle. The unit time price factor is calculated based on the annual per capita GDP of the region, and then the value of delay of intelligent connected passenger vehicles operation is solved. Finally, an example analysis was conducted on the road network of X city, and it was found that the average daily delay value generated by intelligent connected passenger vehicles passing through the road network was ¥28453, and the average congestion delay value generated per kilometer per hour was ¥6.23, thus verifying the rationality of the calculation method. This method provides a reference for formulating road congestion pricing strategies in the context of intelligent networking.

With the vigorous development of automotive information network technology and artificial intelligence technology, the degree of automotive intelligence continues to improve, and the forms of interaction between users and the cabin are constantly diversified. Among them, screen, voice, buttons, and steering wheel are the four most common forms of interaction. The UI design of the screen, the recognition ability of the voice system, the ease of understanding of button icons, and the tactile feel of buttons can distract drivers, causing safety hazards and increasing the probability of safety accidents. This article uses eye movement technology to capture eye movement data of drivers applying different interaction methods in different scenarios. By analyzing eye movement data, the driver's level of distraction and corresponding interaction function design defects are obtained, which helps enterprises optimize product functions, improve the operational convenience of functions, and reduce driving safety risks.

In the context of growing demands for energy conservation and emissions reduction, green logistics has become a focal point in the logistics industry. However, the adoption of new energy vehicles requires substantial capital investment, necessitating an assessment of the economic benefits and overall operational costs of different vehicle types. This study establishes a comprehensive cost calculation model for three types of new energy vehicles: battery electric vehicles (BEV), hybrid electric vehicles (HEV), and fuel cell electric vehicles (FCEV). The model not only considers purchase costs but also analyzes energy consumption costs and their cost trends over time. Through comparative analysis of economic benefits at different distances in current and future stages, this study reveals the long-term economic advantages of new energy vehicles compared to Internal Combustion Engine Vehicles (ICEV). The results indicate that driven by continuous technological advancements, BEV demonstrates superior long-term economic benefits compared to ICEV.

The loading efficiency of the freight yard is related to the efficiency of railway transportation and the organization of trains. The use of fixed catenary or the existing translational mobile catenary in the freight yard will occupy the loading space above the trains, thus restricting the use of large-scale loading machinery. Based on this, this article designed a sunken mobile catenary, which moves the contact suspension part to the ground or pit height when it is not in working position, releasing the loading space above the train and improving the efficiency and safety of loading operations. After on-site inspection, the equipment has a reliable structure, flexible movement, mobile device with good integrity and synchronization, which can effectively save manpower and machine power and improve the efficiency of railway freight transport.

The label data on the occurrence of bad behaviors of power company suppliers is seriously unstructured, and a large amount of effective information is difficult to obtain and reuse. This has a certain impact on the subsequent supplier evaluation, especially on the work of the bidding and purchasing department. Therefore, it is necessary to extract the core information in the bad behavior data so that the business department can take targeted management countermeasures in the subsequent bidding and procurement and cooperation process. In this paper, the framework design, model construction and experimental evaluation of the bad behavior label model of power suppliers based on supervised machine learning are carried out according to the characteristics of the bad behavior data of power suppliers.

With the digital transformation of traditional enterprises and the acceleration of modern supply chain construction, for power enterprises, the selection of high-quality suppliers is related to the safe and stable operation of the power grid. However, with the large number of suppliers of power enterprises and the large types of materials involved, there are various challenges in information and intelligent construction. This paper uses the portrait technology to extract the label information from the massive data for data analysis. It innovatively combines the portrait technology with the supplier management theory, designs the portrait system for power suppliers, and provides the path of label database construction and supplier evaluation system construction for the supplier evaluation of power enterprises.

The interval time of trains and workshops during the peak period of high-speed railway is the main factor determining the anti-interference ability and transportation ability of train diagram. Taking the number of delay trains, delay total time, redundant time utilization rate and punctuality rate as evaluation indicators, the high-speed railway dispatching command simulation system was used to study the influence of train interval time on the reliability of the train diagram. The results show that the train diagram is highly reliable under the absence of interference with the system, and use of 5-minute interval time can ensure the reliability of the train diagram under the interference with 5-minute and 10-minute. When the 5-minute experimental scheme is disturbed more than 10-minute, the longitudinal propagation degree of the late point is greatly increased. In order to ensure the reliability of the train diagram, it is advisable to adopt an 8-minute interval train diagram.

By analyzing the collected accident reports of ship grounding incidents, our aim is to summarize the primary human factors that lead to ship groundings. Through an analysis of the causal chain of grounding accidents, we will determine the Bayesian network structure. Finally, we will apply accident samples to validate the effectiveness of the Bayesian network model established in this paper. Simultaneously, we will conduct a causal analysis of the Bayesian network model to identify the most probable causal chain leading to grounding accidents, thereby providing safety management strategies for seafarers, shipping companies, and maritime safety regulatory authorities.

In this paper, we introduce an enhanced YOLOv5-based model tailored to address challenges in detecting small vehicle targets within high-altitude surveillance contexts. Within the backbone network, the SE attention mechanism is incorporated after each convolutional module. To amplify the detection capabilities, a high-resolution detection head is integrated. Within the intermediate (neck) network, we apply context aggregation and feature fusion techniques before the small target detection head, while employing the CBAM attention mechanism for other detection heads. To further optimize performance, the loss function has been revised by substituting the traditional Ciou calculation with Focal Eiou, in tandem with enhancing object localization using the OTA algorithm. Additionally, the Retinex algorithm is deployed for data augmentation. Performance assessments on the VisDrone2019 dataset revealed that our proposed approach outperformed the standard YOLOv5, 2019 challenge sota, and TPH-YOLO (with prior +CBAM) models by margins of 20%, 7%, and 3%, respectively. A series of tests affirmed the robustness of our algorithm, particularly in the realm of small vehicle target detection.

Through the analysis of each important index in the traditional dynamic compaction construction process, combined with the principle of digital technology, the digital dynamic compaction system was designed and developed. The system combines the actual construction process to realize the functions of design data import, ramming point guide, ramming point location acquisition, ramming times, ramming process quantity, and falling distance monitoring. Meanwhile, combined with big data analysis and wireless transmission, digital dynamic compaction system can display process data in real time, remote control, and strict control of construction quality as well as permanent storage of construction data in the whole process of dynamic compaction construction. The developed system in the study can effectively reduce personnel investment and artificial error, realize the digitization of dynamic consolidation construction process, simplify construction process, and improve construction quality while reducing management cost and personnel cost.

Aiming at the shortcomings of current mainstream detection algorithms in intelligent vehicle environment awareness, a design scheme of intelligent vehicle object detection system based on DC-YOLOv8s (Deformable Convolution-You Only Look Once version 8s) is proposed. The intelligent vehicle object detection system based on DC-YOLOv8s is composed of the preprocessing module, the feature extraction module, the feature fusion module and detection module. The preprocessing module performs geometric transformation and data enhancement on the input image. The feature extraction module is used to extract the feature information of the image to be detected. The function of the feature fusion module is to fuse low-level features with high-level features, and the detection module predicts the category and location of the detection target according to the output of the feature fusion module. By means of experiments, the feasibility of intelligent vehicle object detection system based on DC-YOLOv8s is verified. Through the analysis of experimental results, the detection system can give full play to the robustness of deformable convolution and the effectiveness of ICIoUL (Improved Complete Intersection of Union Loss) for improving positioning accuracy. Experimental results show that our system's object detection accuracy can reach 94.6%. Compared with the YOLOv8s model, the Precision is increased by 7.4%, the Recall is increased by 6.9%, and the mAP@0.5 is increased by 6.5%.

In order to support the research on the influence of connected autonomous vehicle on traffic flow, taking into account the influence sensitivity of multi front vehicles speed and acceleration and the influence weight within the communication range, connected longitudinal-control-model (C-LCM) model of connected autonomous vehicle is built based on the longitudinal-control-model (LCM). In the scenario of uniform headway and uniform traffic flow, the conditions for model stability were discussed by applying small disturbances. The relationship between the stability domain of the model and the number and position of the preceding vehicles was analyzed through simulation environment. After determining the values of C-LCM parameters, the impact of model parameters on model performance was explored in specific common scenarios such as acceleration, deceleration, and disturbance. The research results indicate that connected autonomous driving vehicles can obtain more information about the front vehicle, and give following vehicles more time to change their operating behavior to maintain stable traffic flow. At the same time, in disturbed scenarios, this car following model can reduce unexpected fluctuations in speed. The model has good stability and anti-interference ability.

In response to the issue of limited timeliness in the current monitoring of ship anchor dragging, a method of early warning of ship anchor dragging risk utilising fuzzy set theory was proposed. Based on sailing simulator, ship anchoring simulation experiments were conducted to collect trajectory data of ship anchoring under various environmental conditions, and the characteristics of the trajectory of normal anchored ships were analyzed. During a yawing motion, the yawing width, ground speed and heading change rate of were established as early warning indicators for ship anchor dragging risk. Triangular fuzzy numbers and the distribution of early warning indicators obtained from simulation experiments were combined to determine the membership functions of each indicator, and the entropy weight method was used to determine indicator weights. Then ship anchor dragging risk was comprehensively evaluated and the anchor dragging probability was quantified. Warning levels were set according to actual conditions, and the warnings will be alarmed if the anchor dragging probability reaches the corresponding warning level. The results indicate that during the yawing motion of anchored ships, larger yawing width, higher ground speed, and smaller ship heading change rate correspond to higher anchor dragging probability. Finally, sailing simulators were utilized for simulation experiments ,and the feasibility of the proposed method was validated. The method exhibits practical operability and demonstrates favorable timeliness in early warning capabilities.

Maritime safety has always been a topic of close interest to the shipping community. In order to accurately grasp the characteristics and key causative factors of global maritime accidents, improve navigation safety, and reduce the occurrence of maritime accidents, global maritime accident data from 2014 to 2022 were collected and processed from the Marine Casualties and Incidents (MCI) module of the Global Integrated Shipping Information System (GISIS). Firstly, statistical analysis is used to analyze the temporal characteristics of global maritime accidents, and then Gaussian mixture model (GMM) clustering is used to analyze the distribution characteristics of accidents. Finally, based on the C5.0 algorithm, the key causal factors of maritime accidents are excavated. The results show that most of the maritime accidents occurred in the eastern part of the North Atlantic Ocean and the western part of the Pacific Ocean; Crew and environmental factors are the main causes of maritime accidents. Based on the above results, targeted management suggestions are given to provide decision-making reference for the relevant maritime management organizations to implement on-site supervision more effectively and to ensure the safety of maritime navigation.

Close parallel runway operations play a critical role in maximizing airport capacity and facilitating efficient aircraft arrivals and departures. However, the close proximity of parallel runways presents unique challenges during landing operations, including the risk of runway incursions, wake turbulence, and communication issues. This research paper aims to conduct a comprehensive survey of existing studies on close parallel runway landing operations, analyzing the various approaches employed by airports worldwide. The objective is to identify the most effective method for managing very closely recommendations. The survey encompasses factors such as runway spacing, air traffic control procedures, knowledge, and the identification of best practices. This study aims to contribute to the enhancement of safety and operational efficiency in close parallel runway environments. The findings of this survey will be valuable for airport operators, air traffic control agencies, and aviation authorities in making informed decisions.

In order to further improve the disturbance suppression ability of smart city two-axis monitoring, this paper proposes a active disturbance suppression(ADS) algorithm for smart city two-axis monitoring. The disturbance torque of two-axis monitoring loop is estimated through the establishment of nominal model, disturbance design and disturbance estimator design. The interference torque of the two-axis monitoring loop is compensated by the compensator and tracker, and the proposed method is compared with the traditional PID and LADRC algorithms by simulation experiments. The simulation results show that the disturbance suppression capability of the proposed linear active disturbance suppression algorithm is improved by 74.996% compared with PID algorithm and 52.557% compared with LADRC algorithm, which can effectively improve the disturbance suppression capability of two-axis monitoring.

In order to improve the construction of the assembly center, this paper builds an assembly center location optimization model with the smallest operation cost based on the composition of the cost of China Railway Express operation, and designs a multi-population genetic algorithm to solve it. Secondly, a transportation network composed of 20 cargo source areas, 3 alternative assembly centers, 4 ports, and 3 destination cities was selected as an example to conduct analysis and verification, determine the optimal location strategy. Finally, this article draws conclusions and suggestions, which can provide reference for the construction of China Railway Express assembly center.

The admission qualification and attendance management of campus activities is a difficult problem in college student management. In order to solve this problem, this paper proposes a rapid deployment method of existing equipment and new equipment based on the Internet of Things technology. The collected data is transmitted to the resource server and data server through the Internet of Things and is combined with the "dynamic two-dimensional code" to complete the admission qualification and attendance management. As a result, it solves the problems of difficult deployment and poor scalability of traditional equipment. Therefore, compared with the traditional equipment recording or paper recording method, this method has the advantages of transformation, easy deployment and convenient maintenance. At the same time, real-time logs can be generated in the system after checking in the above way. Using the technology of "deep learning" to analyse these logs, we can accurately analyse and judge the rules of college students' daily activity attendance.

The operation of cableways inevitably raises concerns regarding the health of support structures, owing to various factors. However, traditional monitoring methods, such as manual observation and automatic monitoring systems, have proven inadequate in achieving continuous, real-time, and automated monitoring of support structure deformations over the long term. To ensure the secure operation of the cableway and to further investigate the factors contributing to the deformation of cableway supports, this study implemented a comprehensive automated deformation monitoring system by installing GNSS receivers, accelerometers, inclinometers, meteorological meters, and various other sensors on the cableway located in the Yinzuo Tianmeng Tourist Area of Yimeng Mountain.

Electric vehicles have been experiencing rapid growth in China in recent years, and this trend is expected to impact the grid load. Using Jiangsu Province in China as an example, this paper predicts that the market penetration rate of new energy vehicles will reach 50% in China, and the number of electric vehicle owners in Jiangsu Province will exceed 6 million by the year 2030. Further analysis reveals that the increasing adoption of new energy sources and electric vehicles leads to a rise in centralized EV charging, which amplifies the disparity between peak and off-peak electricity demand. Additionally, this paper introduces an objective function centered around peak-value electricity demand and a peak-valley time-sharing tariff strategy with the aim of minimizing the peak-to-off-peak load difference in the grid. It is found that, in order to reduce the peak-to-off-peak load difference in the grid, there is a need to increase the electricity price during peak hours to some extent.

Urban travel data typically encompass various temporal and spatial scales and features. Nonetheless, existing traffic travel feature analysis often focuses on a single spatial scale, which may overlook crucial information and hinder features at different scales. Consequently, this could reduce the accuracy of travel pattern recognition. To solve this problem, we introduce a multi-scale fusion transformation algorithm. In our approach, we first establish a multi-level taxi travel data grid system. Then, we apply the I-Clique clustering algorithm to perform grid clustering at each level of the system to identify taxi travel feature regions of different scales. Subsequently, a multi-scale fusion pyramid model is constructed utilizing the acquired travel feature regions at various scales. Finally, we determine the transformation relationships between scales in the pyramid model, enabling the identification of travel feature areas across both known and unknown scales. This approach facilitates the recognition of travel features within cities at any scale. Experimental results demonstrate that the multi-scale fusion transformation algorithm enhances our understanding of spatial aggregation patterns and the hierarchical structure of geographical entities, offering a novel approach for multi-scale urban traffic analysis.

In order to establish an around view monitoring system (AVMS) evaluation method suitable for China's transportation environment, standard requirements, and detection technology level, this article proposes typical testing scenarios and key indicators for AVMS based on the causes of commercial vehicle accidents in China, the operating principles of AVMS, and domestic and foreign AVMS standards such as GB/T 39265-2020. And based on China's testing technology level, propose an AVMS testing plan that meets China's technical level and regulatory requirements. This article not only improves the evaluation method of Chinese AVMS, but also provides reference for the revision of Chinese AVMS standards and system development.

In real life, high-pressure sodium lamps are still the main products of road lighting in China, but LED street lights replace traditional street lights faster and faster. Through experiments, we have designed and developed an energy-saving and environmentally friendly street lamp suitable for haze weather, and its light has high penetration in bad weather environments such as smog^[1], which can effectively solve the problems of weak penetration, large system power consumption and backward control mode of existing street lamps in haze weather. The street lamp combines the Internet of Things, artificial intelligence and other technologies to make the urban road lighting project more intelligent, environmentally friendly and energy-saving, and reduces the complexity and engineering cost of the entire lighting project through standardized technical interfaces.New energy smart street lamp can realize refined artificial intelligence light distribution and adjust color temperature according to the flow of people, traffic and weather changes, which controls the LED brightness by applying PWM signals, completes the dimming of the LED without changing the color, and uses PWM switch dimming to adjust the on-time ratio of the cool white LED array and the warm white LED array, and intelligently adjusts the color temperature of the road surface suitable for observing smog weather^[2]. Compared with the traditional high-voltage sodium lamp, supercapacitors are used instead of traditional batteries, thereby improving power density, prolonging life, and expanding the operating temperature range, while using DC-DC conversion circuits to realize the maximum power point tracking (MPPT) algorithm, so that the photovoltaic array always outputs the maximum power, charges the supercapacitor, and improves the charging efficiency. At the same time, the street lamp realizes efficient management of street lamps and reduces maintenance costs according to remote fault feedback and other functions.