Current Issue

25 April 2024, Volume 24 Issue 2 Previous Issue   

---

For Selected:

Download Citations EndNote Ris BibTeX

Toggle Thumbnails

Select

Impact of Charging and Incentive Strategies on Commuting Mode Choice WANGDianhai, LIYiwen, CAI Zhengyi 2024, 24(2): 1-12.  DOI: 10.16097/j.cnki.1009-6744.2024.02.001 Abstract ( )   PDF (2668KB) ( )   This paper investigates the regulatory impact of two traffic demand management strategies, tolls and rewards, on travel mode choices, using the main urban area of Hangzhou as a case study. The stated preference (SP) and revealed preference (RP) surveys were performed to understand the intention of private car commuters' mode choice under parking charge and travel reward scenarios. The disaggregate theory was used to establish Nested Logit (NL) models for commuting mode selection under separate and joint implementation of parking fees and travel rewards. The results indicate that both parking fees and travel incentives can reduce private car travel demand and promote public transportation. Only when the parking price reaches a certain level can private car trips be effectively reduced, and appropriate incentives can actively encourage travelers to switch to other modes of travel. If charging and incentive strategies are implemented simultaneously, it will manifest a joint effect of charging as the main approach and incentive as a supplement. In all three scenarios, income is a significant factor influencing travel mode choices. The higher the income, the more likely the continuation of private car usage. In the scenario with only a parking fee, the elasticity of parking fees increases with the rate; there are limited elasticity when the rate is low. The elasticity of travel rewards initially raises and then drops with the increase in the reward amount; Small rewards also show elasticity. References | Related Articles | Metrics

Select

Analysis of Residents' Travel Mode Choice in Medium-sized City Based on Machine Learning LI Wenquan, DENGAnxin, ZHENGYan, YIN Zijuan, WANG Baifan 2024, 24(2): 13-23.  DOI: 10.16097/j.cnki.1009-6744.2024.02.002 Abstract ( )   PDF (3143KB) ( )   This paper aims to investigate the characteristics of travel behaviors and the influencing factors on travel mode choice in a medium-sized city. Utilizing travel data from a medium-sized city in China, a random forest model embedded with a particle swarm optimization algorithm adding a variation procedure (PSO-RF) was proposed for travel mode choice prediction, due to the distinctions in prediction accuracy and modeling rationality of discrete choice model and machine learning model, as well as the characteristics and efficiency of hyperparameter optimization algorithms. The predictive accuracy, predictive mode proportion's absolute deviation, and expected simulation error were used to statistically compare the predictive performance among PSO-RF, machine learning models, and the multinomial Logit model. The SHAP (SHapley additive exPlanation) model was employed to thoroughly analyze the nonlinear relationships among individual socio-economic attributes, travel attributes, mode attributes, and residents' travel mode choices. The results indicate that PSO-RF has the highest average overall prediction accuracy (0.856), and the lowest average predictive mode proportion's absolute deviation (0.062) and average expected simulation error (0.306). Statistically significant differences in models' predictions are observed. Distance has the most prominent impact on the choice of different travel modes. The modes of walking and private cars show higher sensitivity to distance, with probability changes exceeding 35% at different distances. Individuals under 30 years old exhibit greater variations in the probability of choosing different travel modes compared to other age groups. Gender, car ownership, and bus IC card ownership notably affect the probability of choosing a bus and a private car. References | Related Articles | Metrics

Select

Effects of Carbon Emission Reduction from Freight Structure Adjustment at Provincial Level in China LIU Zhaoran, ZHU Lichao 2024, 24(2): 24-33.  DOI: 10.16097/j.cnki.1009-6744.2024.02.003 Abstract ( )   PDF (2327KB) ( )   The freight sector is one of the most challenging sectors in supporting China's goal of achieving "carbon peak and carbon neutrality", and the adjustment of freight structure is a key means to achieve CO2 emission reduction in freight transportation. However, there is an urgent need for in-depth quantitative research from both management and academia to assess the CO2 emission reduction effects of freight structure adjustment, especially considering the spatial connection of freight CO2 emissions between regions. To address these issues, this study employs a "top-down" approach to estimate freight CO2 emissions for 30 provinces in China from 1999 to 2019. It also develops a spatial econometric model that incorporates social-economic variables and freight characteristics to quantify the impact of freight structure and other factors on CO2 emissions from freight transportation. The findings reveal that the average annual growth rate of freight CO2 emissions in most Chinese provinces exceeds 10%, although the growth rate is decelerating. Coastal provinces, such as Shandong, Guangdong, Shanghai, Liaoning, and Jiangsu exhibit the highest emissions, with provinces within the same geographical region displaying similar changes. Given the freight CO2 overall effects of rail and water freight transportation on CO2 emissions from freight transportation are-0.193 and 2.378, respectively, prioritizing the transition from "road-to-rail" should be a focal point in most provinces, while the shift from "road-to-water" should be approached cautiously to effectively achieve CO2 emission reduction in freight transportation. References | Related Articles | Metrics

Select

Natural Peak Characteristics and Peak Forecast of Carbon Emissions in Transportation Industry YANGDong, LIYanhong, TIAN Chunlin 2024, 24(2): 34-44.  DOI: 10.16097/j.cnki.1009-6744.2024.02.004 Abstract ( )   PDF (3274KB) ( )   The peak of carbon emissions in the transportation industry is a natural long-term evolutionary process. In order to study the process of carbon peaking in China's transportation industry, this paper adopts the international analogy method, selects typical foreign countries, and compares the time of the peaks of the overall national carbon emissions, transportation industry carbon emissions, and converted turnover. The natural peak characteristics of transportation industry carbon emission are analyzed and the time of the natural peaks of carbon emissions in China's transportation industry is predicted according to transportation demand forecast. Then, the STIRPAT (Stochastic Impacts by Regression on Population, Affluence, and Technology) prediction model is constructed by introducing the core influencing factors such as carbon emissions per unit of converted turnover and the ratio of railroad to road freight transportation. Finally, through analogical analysis and model prediction, the time and volume of peak carbon emissions of China's transportation sector are obtained. The results of the international analogy show that there is no clear causal relationship between the peak carbon emissions of the transportation industry and the national peak carbon emissions, but it is closely related to the peak converted turnover, and the converted turnover reaches the peak or is close to the peak when the carbon emissions of the transportation industry reach the peak. It is predicted that China's converted turnover will reach a plateau period of 26 trillion ton-kilometers in about 2048. From the perspective of the international analogy, the time of the natural peak of China's carbon emissions from transportation is roughly roughly between 2040 and 2043. The STIRPAT model shows that the carbon emissions of China's transportation industry will increase by 1.201%, 0.259%, 0.454%, and-0.389%, respectively, for every 1% increase in urbanization rate, per capita GDP, carbon emissions per unit converted turnover, and railway- road freight ratio. Based on the combination prediction of international analogy and STIRPAT model, China's transportation industry will achieve peak carbon emissions in 2038-2040, with about 1.3 billion tons. References | Related Articles | Metrics

Select

Impact Analysis of Objective Weighting on Critical Distance for Express Goods Transportation Modes SUNZongshenga, SHUAI Bin, YANG Junjiea, XU Minhaoa 2024, 24(2): 45-52.  DOI: 10.16097/j.cnki.1009-6744.2024.02.005 Abstract ( )   PDF (1674KB) ( )   The transportation in China has shifted from independent development among various modes to integrated development. Express goods transportation is an emerging key field for the integrated development of integrated transportation. The study of the critical distance for express goods transportation modes is of great significance to determine the reasonable transportation structure. In response to the shortcomings of subjective weighting in existing research on critical distance issues, expression for critical distance and weight coefficients of service attribute is established based on the express goods sharing rate model. Then, the weight coefficients are objectively calculated based on the CRITIC (Criteria Importance Though Intercriteria Correlation) method. The model is analyzed and validated through example analysis, and the results are compared between different scenarios. The results indicate that: the objective assignment method can make the weight coefficients of each service attribute heterogeneous with the increase of transportation distance; economy has the largest fluctuation in 300 to 400 km and 900 to 1100 km; the transportation distance threshold between convenience and economy is 327 km, and its importance will be higher than economy with the increase of transportation distance. The influence of objective weighting on the critical distance between the modes of express goods transportation is mainly reflected in the expansion of the absolute dominant distance range of high-speed rail express transportation, which changes from 700~1500 km to 400~2000 km. References | Related Articles | Metrics

Select

Connection Optimization of Container Sea-rail Combined Transport Based on Vehicle-ship Direct Access Mode WANGShuang, SUN Xiang 2024, 24(2): 53-62.  DOI: 10.16097/j.cnki.1009-6744.2024.02.006 Abstract ( )   PDF (2269KB) ( )   To address the optimization problem of container-sea-rail intermodal connections under the direct access mode of vehicles and ships, this study considers the connection between trains and container ships in terms of time and volume, based on the fixed ship schedule. A mixed-integer nonlinear programming model is constructed for the timetable optimization of sea-rail intermodal transportation, aiming at maximizing the number of connected trains, minimizing the number of ships involved in the connection and minimizing the timetable adjustment. The solution is obtained based on a hierarchical optimization method. The model is verified by using Yantian Port in Shenzhen as an example. The results show that the optimization of train schedule considering time and volume connection can increase the number of connected trains by 6, and the number of directly transshipment containers by 75% with 466 TEU. The number of ships involved in the connection increased by 1, and the number of containers taken directly by the three ships increased by 2.6%, 4.2% and 2.5%, respectively. The total dwell time of the containers in the port was reduced by 26%. The study shows that the model can provide useful references for improving the connection level of sea-rail combined transport direct access mode. References | Related Articles | Metrics

Select

Method for Identifying Dangerous Driving Behaviors in Heavy-duty Trucks Based on Multi-modal Data WUJianqing, ZHANG Ziyi, WANG Yubo, ZHANGYu, TIANYuan 2024, 24(2): 63-75.  DOI: 10.16097/j.cnki.1009-6744.2024.02.007 Abstract ( )   PDF (3402KB) ( )   This paper proposes a multi-modal method to identify dangerous driving behaviors of heavy-duty trucks, which integrates driving operation data, eye-tracking data and electrocardiogram (ECG) data in the analysis. A naturalistic driving experiment is designed to collect driving data using three types of devices: vehicle inertial navigation systems, eye-tracking decoders, and physiological data recorders. A multi-modal driving dataset is established through data synchronizing and data cleaning processes. The dangerous driving behaviors are divided into two categories: dangerous manipulation behaviors and fatigue driving behaviors. By extracting data features, nine dangerous driving behavior indicators are defined to represent five types of dangerous driving behaviors, including speeding, unstable speed, rapid speed changing, rapid lane changing, and fatigue driving. For dangerous manipulation behaviors, characteristic thresholds are determined through literature review, indicator calculation, and interquartile range method. For fatigue driving behaviors, fatigue driving levels are identified through factor analysis and K-means clustering methods. A random forest (RF) classification model is then developed to identify dangerous driving behavior. When compared to traditional methods, including back propagation neural network (BPNN), K-nearest neighbors (KNN), support vector machine (SVM), the proposed model surpassed others in terms of accuracy and fitting performance. The model achieved a classification accuracy of over 90% for all types of dangerous driving behaviors. The results prove that the proposed methods are effective in identifying dangerous driving behaviors and it provides a theoretical basis for multimodal warning systems of dangerous driving behaviors. References | Related Articles | Metrics

Select

Formation and Evolution Mechanism of Connected and Autonomous Fleet Based on Fish Streaming Effect WEILiying, WU Runze 2024, 24(2): 76-85.  DOI: 10.16097/j.cnki.1009-6744.2024.02.008 Abstract ( )   PDF (3532KB) ( )   With the rapid development of connected and autonomous vehicles (CAV), the research on traffic characteristics and cooperative control of the intelligent mixed traffic that is composed of CAVs and human-driven vehicles, has become a research focus. In this paper, a multi-lane cellular automata model for the mixed traffic is established to simulate the formation and evolution process of a CAV fleet. Firstly, the fish streaming effect is introduced to describe the formation process of four kinds of CAV fleets based on their networked characteristics. Secondly, the Markov property is used to calculate the fleet scale transfer probability from the perspective of the fleet, and the evolution process of the CAV fleet state is described. Thirdly, the rule of Gipps safety distance is introduced to improve the NaSch model, and CAV vehicles and fleet are subjected to the speed guidance. Finally, this paper carries out simulation experiments on the established mixed traffic flow cellular automata model based on fish streaming according to the measured vehicle arrival rate. The results show that the CAV fleet can effectively improve the operating state of mixed traffic and alleviate traffic congestion; Under the condition of a 60% penetration rate, the congestion rate can be reduced by 43.9% when the CAV fleet scale is 3 compared with the non-fleet, the traffic flow speed can be increased about 43%, and the average speed tends to be stable with the increase of fleet scale. References | Related Articles | Metrics

Select

Flexible Traffic Signal Control Method Based on Non-conflict Merging Strategy HUANGWei, LI Shichang, ZENG Haipeng 2024, 24(2): 86-95.  DOI: 10.16097/j.cnki.1009-6744.2024.02.009 Abstract ( )   PDF (2679KB) ( )   The development of intelligent and connected vehicles makes it possible to finely control the vehicle trajectories and the new ideas and concepts have appeared for urban traffic signal control. This paper proposes a flexible signal control method based on non-conflict merging strategies. The merging phase is introduced to accommodate both through and left-turn vehicles. Based on the NEMA (National Electrical Manufacturers Association) dual-ring phase structure, the merging strategy is integrated to form a dual-ring merging signal phase structure with 12 action spaces. With the optimized phase structure, this paper proposes an improved reinforcement learning algorithm for the signal timing. The algorithm fully considers the actual rules of phase switching under the NEMAdual-ring structure and learns the optimal phase control strategy for the current state. The performance of the proposed control method is tested with the SUMO simulation. Different control strategies are compared, including the traditional NEMA dual-ring structure, and the NEMA dual-ring merging phase structure under both actuated control and improved reinforcement learning algorithms. The impact of the penetration rate of connected vehicles on the control performance is also analyzed. The results show that the proposed method can generate more realistic and flexible phase timing plans while maintaining the NEMA dual-ring merging phase structure. Under different traffic flow conditions, especially high traffic volume and uneven distribution of traffic volume scenarios, compared with the traditional NEMA dual-ring phase structures and actuated control methods, the proposed method can reduce vehicle delay by about 37% and hence improve the traffic operation at the signalized intersections. References | Related Articles | Metrics

Select

Traffic Signal Control with Deep Reinforcement Learning and Self-attention Mechanism ZHANGXijun, NIE Shengyuan, LI Zhe, ZHANG Hong 2024, 24(2): 96-104.  DOI: 10.16097/j.cnki.1009-6744.2024.02.010 Abstract ( )   PDF (2200KB) ( )   Traffic signal control (TSC) is still one of the most important research topics in the transportation field. The existing traffic signal control method based on deep reinforcement learning (DRL) needs to be designed manually, and it is often difficult to extract the complete traffic state information in the real operations. This paper proposes a DRL algorithm based on the self-attention network for the traffic signal control to analyze the potential traffic from limited traffic state information and reduce the difficulty of traffic state design. The vehicle position of each entry lane at the intersection is obtained, and the vehicle position distribution matrix is established through the non-uniform quantization and one-hot encoding method. The self-attention network is then used to analyze the spatial correlation and latent information of the vehicle location distribution matrix which is an input of the DRL algorithm. The traffic signal adaptive control strategy is trained at a single intersection and the adaptability and robustness of the proposed algorithm are verified in a multi-intersection road network. The simulation results show that in a single intersection environment, the proposed algorithm has better performance on the average vehicle delay and other indicators compared with three benchmark algorithms. The proposed algorithm also has good adaptability in the multi-intersection road network. References | Related Articles | Metrics

Select

Collaborative Study of Decision-making and Trajectory Planning for Autonomous Driving Based on Soft Acto-Critic Algorithm TANGBin, LIU Guangyao, JIANG Haobin, TIAN Ning, MI Wei, WANG Chunhong 2024, 24(2): 105-113.  DOI: 10.16097/j.cnki.1009-6744.2024.02.011 Abstract ( )   PDF (2319KB) ( )   To improve the learning speed, safety and rationality of autonomous driving decision-making, this paper proposed a collaborative method of autonomous driving decision-making and planning based on Soft Actor-Critic (SAC) algorithm. The autonomous driving decision planning collaborative agent was designed by introducing the SAC algorithm with the rule-based decision planning method. Combined with the Self-Attention Mechanism (SAM) and the Gated Recurrent Unit (GRU), a preprocessing network was constructed to improve the agent's ability to understand traffic scenarios and improve the agent's learning speed. Considering the specific implementation mode of the planning module, the study used the action space to improve the executability of the decision- making results. The reward function was designed by using the information feedback, adding the constraints of vehicle driving conditions to the agent, and transmitting the trajectory information to the decision-making module. The collaboration of decision-making and planning improved the safety and rationality of decision-making. The dynamic traffic scenarios were built in the CARLAautonomous driving simulation platform to train the agent, and the proposed decision-making and planning collaboration method was compared with the conventional decision-making planning method based on SAC algorithm in different scenarios. The experimental results showed that the learning speed of the agent designed in this paper had increased by 25.10%. The average vehicle speed generated by its decision outcomes was higher and closer to the expected road speed. The speed variation rate produced by its decision outcomes was smaller, and the path length and curvature variation rate resulting from its decision outcomes were also smaller compared to traditional methods. References | Related Articles | Metrics

Select

AVehicle Lane-changing Trajectory Prediction Model Based on Temporal Convolutional Networks and Attention Mechanism YANGDa, LIU Jiawei, ZHENG Bin, SUN Feng 2024, 24(2): 114-126.  DOI: 10.16097/j.cnki.1009-6744.2024.02.012 Abstract ( )   PDF (2493KB) ( )   An accurate vehicle trajectory prediction model can provide self-driving vehicles with precise information about the motion states of surrounding vehicles in mixed traffic flow environments, allowing it to assess the possibility of conflicts with neighboring vehicles in the short term. This paper proposes a vehicle lane-changing trajectory prediction model based on Temporal Convolutional Networks with Attention Mechanism (TCN-Attention) to improve the accuracy of vehicle lane-changing trajectory prediction. This model uses Temporal Convolutional Networks as the current input's feature extractor and utilizes a temporal and spatial attention mechanism to establish dynamic correlations between different time steps and spatial positions. Specifically, the combination of temporal and spatial attention mechanisms helps the model extract essential semantic features in both the temporal and spatial dimensions before and after lane-changing, enabling it to more accurately capture the dynamic spatiotemporal relationships between vehicles. This enables precise predictions of lane-changing trajectories on highways. Different from the traditional only using a trajectory features as input, our method achieves the multi-dimensional expansion and fusion of the input features, and further improves the accuracy of the trajectory prediction. In addition, this paper proposes a new method to define the start and end time of lane-changing in the dataset more accurately. Experiments show that the proposed model can predict the trajectory of the lane-changing with high accuracy, and the overall effect is better than other deep learning models. Compared with the Convolution Long Short- Term Memory(ConvLSTM), the Mean Absolute Error( EMAE ) of TCN-Attention is reduced by 69.8%, the Root Mean Square Error( ERMSE ) is reduced by 49.15% and the MeanAbsolute Percentage Error( EMAPE ) is reduced by 14.24%. References | Related Articles | Metrics

Select

Assessment of Drivers' Potential Hazard Perception at Unsignalized Intersections PENGJinshuan, CHENG Jiajia, ZHAO Liuchang, LUO Shuang, YUAN Hao, XU Lei 2024, 24(2): 127-135.  DOI: 10.16097/j.cnki.1009-6744.2024.02.013 Abstract ( )   PDF (2593KB) ( )   To quantify the potential hazard perception ability of drivers at unsignalized intersections, this paper conducted a real vehicle driving test to collect real-time parameters of drivers' eye-movement characteristics. The fixation and saccade law of skilled and unskilled drivers were analyzed when they drive straight through the unsignalized intersections. Based on the Markov chain model, the driver's fixation transfer characteristics were illuminated to reveal the internal relationship between drivers' visual characteristics and potential hazard perception characteristics. The characteristic parameters of drivers' potential hazard perception ability were extracted combining with the statistical analysis under each mapping index of skilled and unskilled drivers. Based on the gray near-optimal comprehensive evaluation method, this study evaluated driver's potential hazard perception ability at unsignalized intersections. The results show that when driving straight through the intersections, the horizontal search breadth, vertical search depth and saccade intensity of unskilled drivers are significantly lower than those of skilled drivers. The ability of unskilled drivers searching for information on both sides of the road is weaker than that of skilled drivers, and timely allocation mechanism of fixation probability is not flexible. Potential hazard perception score of skilled drivers at unsignalized intersection is 31.2% higher than that of unskilled drivers. Among skilled drivers, men's potential hazard perception ability is significantly higher than that of women, while gender has no significant effect on the perceived performance of unskilled drivers. The research results can enrich the theoretical system of defensive driving, and provide important reference for the optimization and improvement of traffic facilities at unsignalized intersections, as well as driver safety education and evaluation. References | Related Articles | Metrics

Select

Human-machine Collaborative Decision-making for Transportation Scheduling Optimization LIU Tao, YOU Hailin 2024, 24(2): 136-148.  DOI: 10.16097/j.cnki.1009-6744.2024.02.014 Abstract ( )   PDF (3361KB) ( )   The paper proposes a human-machine collaborative decision-making methodology based on the deficit function model to solve transportation scheduling problems with flexible constraints. The methodology mainly consists of two stages. In the first stage, by making use of mathematical programming models and the powerful computing capacity of computers, a feasible solution is quickly obtained. In the second stage, with the help of the deficit function model, human beings' own knowledge and experience are employed to further optimize the feasible solution obtained in the first stage, while taking into account flexible constraints. The two stages interact in real time through a graphical user interface composed of deficit function figures, thereby realizing human-machine collaborative decision-making. The effectiveness of the proposed human-machine collaborative decision-making methodology based on the deficit function model is demonstrated through two case studies, i.e., app-based customized bus scheduling problem and civil aviation aircraft scheduling problem. Computation results show that the proposed methodology can realize the automatic construction of vehicle/flight chains and the automatic insertion of deadheading trips. It is useful for solving complex transportation scheduling problems with flexible constraints. References | Related Articles | Metrics

Select

Prediction of Bicycle Trajectory Considering Stressful Avoidance Behaviors LI Yan, LIANG Shujuan, LIU Linjian, SHAO Jin, WANG Fan 2024, 24(2): 149-156.  DOI: 10.16097/j.cnki.1009-6744.2024.02.015 Abstract ( )   PDF (1964KB) ( )   When the space in non-motorized lanes is restricted, cyclists in an overtaken scenario will generate stressful avoidance behaviors to ensure their own safety. In order to clarify their stress reaction when being overtaken, and to design the non-motorized lane according to their behavioral characteristics, a bicycle trajectory prediction model for stress behavior classification is proposed. The model decomposes the dynamic characteristics of bicycles from the frequency domain perspective, classifies the stressful avoidance behaviors into uniform speed, acceleration, and deceleration behaviors based on the cadence range, and uses the whale algorithm to improve the long and short-term memory neural network model to predict the classified cycling trajectories. The proposed method was tested using the 2415 overtaken events obtained from Xi'an City. The results indicate that the proportions of the three avoidance behaviors are 11.3%, 38.3%, and 50.4%, respectively. The predicted trajectory of uniform speed avoidance has a small fluctuation throughout the whole process, with an average lateral displacement of 0.15 m. The predicted trajectory of accelerated avoidance shows larger lateral displacement, with an average of 0.83 m. The predicted trajectory of decelerated behavior has a lateral displacement of 0.47 m. The root mean square errors of these three behaviors are 0.0619, 0.0513, and 0.0587, with their goodness of fit as 0.9589, 0.9774, and 0.9687, respectively. Compared to the results of the model without the consideration of stress-based behaviors, the proposed model improves the prediction accuracy by 11.07%, 13.22%, and 12.21%, respectively. References | Related Articles | Metrics

Select

An Energy-efficient Train Driving Strategy Based on Regime Sequences Optimization ZHAODongsheng, ZHAO Peng, YAO Xiangming, YANG Zhongping, ZHANG Bonan 2024, 24(2): 157-165.  DOI: 10.16097/j.cnki.1009-6744.2024.02.016 Abstract ( )   PDF (2639KB) ( )   The traditional "three-phase" and "four-phase" fixed regimes sequences of train driving mode can hardly meet the demand for energy-efficient train operation under the complex line conditions of urban rail transits. An energy efficient operation strategy optimization method based on condition sequence optimization is proposed. A physical line section is discretized into several equidistant sub-sections, the mapping relationship between the sequence of driving regimes and sub-sections is established, and the multi-objective optimization model of train energy-saving maneuvering strategy is constructed with the objective of minimizing the total traction energy consumption and running time in the section. In order to improve the solution efficiency, the crossover operator and distance operator of the Non-dominated Sorting Genetic Algorithm-II are improved, and the set of energy-efficient train driving strategies is solved based on the discrete simulation method. Two typical sections of Fuzhou Metro Line 1 are used for the case study and the results show that the average reduction of traction energy consumption of trains after optimization is about 19% compared with the traditional manipulation mode. The proposed method can effectively construct operating condition sequences adapted to different line conditions by reasonably selecting the algorithm parameters, and generate a set of energy-efficient train operation strategies with accurate seconds of running time. References | Related Articles | Metrics

Select

Identification of Aggressive Lane-changing Behaviour Based on Unsupervised Cluster Analysis WANGWanqi, CHENG Guozhu, XU Liang 2024, 24(2): 166-178.  DOI: 10.16097/j.cnki.1009-6744.2024.02.017 Abstract ( )   PDF (2855KB) ( )   To effectively guide drivers to adopt safer lane-changing behaviours, this paper proposes a method to identify aggressive lane-changing behaviour based on a modified Self-Organising Mapping Neural Network (SOM-Kmeans) cluster analysis. Driving data and eye movement status are obtained by driving simulation equipment and eye movement equipment. Then, a change-point detection algorithm is applied to extract lane-changing behaviour event data from the multimodal dataset by combining the steering wheel angle and lateral gaze position. Afterwards, SOM Kmeans cluster analysis is used to extract key feature parameters of driver lane changing behaviour and identify aggressive lane changing behaviour. The effectiveness of the SOM-Kmeans clustering method is compared with the density-based clustering algorithm (DBSCAN) and the fuzzy C-mean clustering algorithm (FCM), respectively, for the identification of aggressive lane changing behaviour. The results show that SOM-Kmeans is able to classify aggressive lane-changing behaviour into two types: emergency lane-changing and squeezing lane-changing. The proposed method can establish the corresponding behavioural indicators and thresholds, and identify the lane changing behaviour as aggressive when the acceleration fluctuation in the process of lane changing is greater than 8.22 m·s-3 and the steering wheel angle is greater than 0.83 (°)·s-1. Based on aggressive lane changing behaviour, when the lane changing gap is less than 7.5 m and the duration of the lane changing is greater than 10.3 s, the lane changing is identified as crowded lane changing, otherwise it is emergency lane changing behaviour. Crowded lane changing behaviours are mostly found in mandatory lane changing with heavy congestions, and emergency lane changing behaviours are mostly found in free lane changing with low-to-moderate traffic densities. The accuracy of the proposed method identification is 92.5% when compared with the traditional cluster analysis. The proposed method can identify the types of aggressive lane changing behaviour in a more detailed way, and the results of the study can be used as a way to assess whether there is a deviation from the normal lane changing behaviours of a driver and to measure the driver's lane changing habits. The results of the two-layer clustering can also be used as a referential criteria of the radical type of lane changing behaviours. References | Related Articles | Metrics

Select

Driver's Working Memory and Stress Response Analysis Under Typical Accident Types MUYuwei, YUANWei, ZHANGHuiming, HAO Shuaijie 2024, 24(2): 179-187.  DOI: 10.16097/j.cnki.1009-6744.2024.02.018 Abstract ( )   PDF (2298KB) ( )   To examine the relationship between different working memory patterns in the accidents, drivers' stress response ability and visual behavior, this study carried out driving simulator experiments on 31 drivers under three typical accident types of side collision, rear-end collision and scraping pedestrians on urban roads. Based on the output video of the camera and eye tracker, dynamic clustering was used to divide the driver's gaze area into 7 categories. Meanwhile, the stress response ability and visual behavior of people with different accident types and different working memory capacities were analyzed based on one-way Analysis of variance (ANOVA), Mann-Whitney U test and Kruskal-Wallis test. The results show that there is a significant correlation between the accident type and the driver's stress response ability. The driver's reaction time in the emergency braking event of the front vehicle is the shortest, which is reduced by 25.46% compared with the side collision. Reaction time in the face of danger is affected by the driver's working memory capacity, that is, the driver with a high working memory capacity has a shorter reaction time than the driver with a low working memory capacity. In addition, the total duration of fixation, mean saccade time and pupil diameter were correlated with working memory capacity. The research results can provide theoretical support for targeted early warning of vehicle safety. References | Related Articles | Metrics

Select

Metro Maintenance Tasks Scheduling Considering Resource Constraints LUOQin, HUANG Shan, SONG Jianwei, ZENG Cuifeng, CHEN Jingjing, LI Wei 2024, 24(2): 188-198.  DOI: 10.16097/j.cnki.1009-6744.2024.02.019 Abstract ( )   PDF (2261KB) ( )   Metro maintenance construction is normally undertaken with heavy workload, limited resources, and strict timelines. This paper focuses on the metro maintenance task plan to establish an optimization model and algorithm for maintenance scheduling. The model takes task priority, person, and workspace capacity constraints as constraints, and aims to minimize the makespan and balances person workload. A hybrid algorithm is designed in combination of the linear programming and resource crossover (CPLEX-ROC). A case study verifies the feasibility and effectiveness of the model and algorithm. The results show that compared to manual scheduling, genetic algorithm (GA), and Teaching learning- based optimization (TLBO) methods, the proposed method reduces the makespan reduced by 32.90% , 15.11% , and 10.75% respectively. The person workload balance is improved by 15.44% compared to GA, and improved by 10.62% compared to the TLBO. The proposed method improves the overall operation efficiency of metro maintenance tasks and balances the personal work load in the constructions. References | Related Articles | Metrics

Select

Collaborative Optimization of Train Timetabling and Train Capacity Allocation for Large Passenger Flow Scenario GONGCongcong, YANG Lixing, SHI Jungang, QI Jianguo, ZHOU Housheng 2024, 24(2): 199-207.  DOI: 10.16097/j.cnki.1009-6744.2024.02.020 Abstract ( )   PDF (1962KB) ( )   To alleviate platform congestion at large passenger flow stations and their downstream stations on urban rail transit commuter lines, and reduce the crowding level and safety risks, this paper systematically optimizes the allocation of train capacity resources from temporal and spatial levels. By considering time-varying passenger demand and the train carriage reservation strategy, a collaborative optimization method for train timetabling and train capacity allocation problems is proposed. Specifically, decision variables related to train departure time, number of reserved train carriages, and passenger assignment plan are introduced and an integer linear programming model for the train timetabling and train capacity allocation problem is formulated, to minimize the operational cost of train carriage reservation and the maximum number of waiting passengers on platforms. Among them, the passenger assignment constraints formulated by the Big-M method obey the first-in-first-out principle in time and space levels. To validate the effectiveness of the constructed model, four sets of numerical experiments are implemented by the Gurobi solver directly. The results show that, compared to planned train timetable and two single optimization strategies, the collaborative optimization method can significantly reduce the maximum number of waiting passengers by approximately 60%, 52%, and 31%, and reduce the total passenger waiting time by about 29%, 17%, and 29%, respectively. That is, the collaborative optimization method can balance the temporal and spatial distribution of the urban rail train capacity and effectively mitigate the risk of passenger crowding at platforms. References | Related Articles | Metrics

Select

Classification and Threshold Research on Multivariant Relationship Between Shared Bicycles and Public Transit DENGYajuan, LIU Shuang, BAI Yu, LIU Wenfeng, CUI Liangbin 2024, 24(2): 208-216.  DOI: 10.16097/j.cnki.1009-6744.2024.02.021 Abstract ( )   PDF (2813KB) ( )   To accurately evaluate the complex spatiotemporal relationship between shared bicycles and public transit, this study identifies the multivariant relationship between shared bicycles and public transit based on the causes of substitution or complementation relationships, combined with the distribution characteristics of shared bicycle trips' origins and destinations. Furthermore, considering the differences between bus and rail transits, this study proposed a classification model for the multivariant relationship between shared bicycles and public transit based on a weekly supervised fully connected neural network and calculated the coverage of public transit, the duration of shared bicycles, and the walking distance for integrated boundary thresholds under different relationship classifications and modes of transit using shared bicycles trajectory data. The results indicate that the multivariant relationship between shared bicycles and public transit can be classified as complementation, integrated complementation, and two types of substitution modes. Specifically, the threshold values for the three parameters between shared bicycles and bus transit are 329.75 m, 5.07 min, and 182.93 m, while for rail transit, the threshold values are 816.96 m, 10.27 min, and 653.91 m. The main relationship between shared bicycles and bus transit is the first substitution mode, accounting for 54.98% of total trips, while the main relationship between shared bicycles and rail transit is complementation, accounting for 48.90% of total trips. The relationship between shared bicycles and bus transit is mainly characterized by multivariant substitution and complementation, while with rail transit, it is mainly characterized by multivariant complementation and the other substitution mode. The relationship between shared bicycles and buses is more mixed compared to rail transit. This study provides support for promoting the coordinated development of shared bicycles and public transit at their respective advantageous distances. References | Related Articles | Metrics

Select

Exploring Nonlinear Effects of Built Environment on Dockless Bike Sharing Usage CHENYifan, ZHANG Buhao, DANG Zhen, GUOTangyi, GU Ziyuan, ZHANGYuliang 2024, 24(2): 217-224.  DOI: 10.16097/j.cnki.1009-6744.2024.02.022 Abstract ( )   PDF (2466KB) ( )   To investigate the dependency of dockless bike usage characteristics on the built environment, this paper used dockless bike order data and electronic fence information from Xiamen city in 2020 to analyze the nonlinear explanatory power of the built environment at both the aggregated (grid-level) and disaggregated (individual trips) levels. A machine learning model, namely, extreme gradient boosting model (XGBoost) was adopted. First, the relative importance of six dimensions of built environment variables (density, design, destination accessibility, land use diversity, public transport accessibility, and demand management) were identified on bike trip generation, attraction, and the user's departure time choice. Then, according to partial dependence plots, the impact trends and the threshold effects of built environment variables were evaluated. The results revealed that at the aggregate level, electronic fence density was the most critical factor, affecting travel generation and attraction by 26.88% and 51.88% respectively. A threshold effect was approximately 150 per · km-2. At the disaggregate level, the probability of dockless bike users borrowing bikes during the morning peak was associated with the built environment features of both the origins and destinations. Among these, the proportion of workplaces in the destination grid was the most significant factor (18.17%), followed by the proximity to Central Business District (CBD) of the origin grid (7.34%) and bus stop density in the origin grid (5.91%). References | Related Articles | Metrics

Select

Spatio-temporal Ride Path Selection in Morning Peak for Express-local Trains WANGJing, SHAOYuchen, HU Hua, LIU Zhigang 2024, 24(2): 225-233.  DOI: 10.16097/j.cnki.1009-6744.2024.02.023 Abstract ( )   PDF (2067KB) ( )   Examining the spatial and temporal travel patterns of commuters during the morning peak period can help to better organize the express-local mode of rail transit. Starting from a common express-local mode, this paper designed the ride paths based on commuters' boarding train at origin and direct/ transfer options. Then, the travel time, early arrival/late delays and inside crowding were taken as the factors affecting the travel option of morning peak commuters to estimate the travel costs under different ride paths. This paper developed a ride path allocation model based on the user equilibrium theory, and proved the existence and uniqueness of the user equilibrium segment solutions of the model. The equilibrium model was verified through a case study. The results indicate that short distance commuters are more likely to accept trains that are highly congested but the arrival time are close to the work start time. The long-distance commuters show a more even distribution of trip choices. In addition to the travel time factor, the delay penalties and ridership crowding also affect the transfer behavior of morning peak commuters, and transferring to express trains has the potential to reduce the travel time. High delay penalties and congestion costs may cause commuters at some stations to not choose the transfer path. References | Related Articles | Metrics

Select

Spatio-temporal Characterization of Urban On-street Illegal Parking and Modeling of Role Influencing Factors LIU Keliang, CHEN Jian, QIU Zhixuan, ZHANG Di, TANG Zhen, PENG Qian 2024, 24(2): 234-248.  DOI: 10.16097/j.cnki.1009-6744.2024.02.024 Abstract ( )   PDF (3857KB) ( )   To reveal the spatio-temporal patterns and influence mechanisms of on-street illegal parking in cities, a total of 10396 electronic police parking violation capture data for two months were combined, and 16 influencing factors were selected from the three dimensions of land use, parking supply and road design. Standard ellipse method, hierarchical clustering method and Pearson correlation test are used to analyze the spatial and temporal characteristics of on-street parking violations and screen the influencing factors, and a Bayesian quantile regression model is further constructed based on the characteristics of the dependent variable to analyze the relationship between the number of parking violations and the influencing factors in the morning and evening peaks. The results of the study show that: (1) From the perspective of the temporal characteristics of the number of parking violations, the number of parking violations on weekdays is significantly higher than that on weekends, and is mainly concentrated in the middle of the week. The results of the hourly clustering of the number of weekday parking violations show that the number of parking violations is divided into three time patterns, which are the morning and evening peak time periods (8:00-10:00 and 17:00-19:00), as well as morning and evening peak convergence time periods (10:00-11:00 and 15:00-17:00), and other time periods. (2) In terms of the spatial characteristics of the number of violations, the overall spatial distribution of the number of weekday violations is mainly concentrated in the middle of the week, and the number of weekend violations is mainly concentrated in the middle of the week. The overall spatial distribution of the number of parking violations on weekdays is mainly concentrated in areas with high commercial and residential densities. The spatial change trend of the number of parking violations in a day shows the trend of "expanding-stabilising-contracting". (3) The quantile regression results show that the three dimensions of land use, parking supply, and road design have a non-linear impact and a threshold effect on on-street parking violations. There is a difference in the mechanism between the morning and evening peak parking violation. The management of parking violation needs to take into account the spatial and temporal characteristics of the differentiated management. References | Related Articles | Metrics

Select

An Interpretable Machine Learning Framework-based Approach for Predicting Passenger Flow Distribution in Train Riding Sections SUNGuofeng, JING Yun, LI Hebi, TIAN Zhiqiang, TIAN Xiaopeng 2024, 24(2): 249-262.  DOI: 10.16097/j.cnki.1009-6744.2024.02.025 Abstract ( )   PDF (4383KB) ( )   In order to clarify the impact of railway passenger transportation services on the prediction of passenger flow distribution, we propose a method based on an interpretable machine learning framework to predict passenger flow distribution in high-speed railway sections. First, we propose a framework capable of predicting passenger flow distribution in sections by using gradient-boosted tree models. Meanwhile, we construct different gradient-boosted tree models, including GBDT, XGBoost, LightGBM, and CatBoost. Secondly, the importance of feature contributions and feature variables are calculated using the SHapley Additive exPlanations (SHAP) method. A non-linear relationship between different features and passenger flow distribution is revealed. The experiment from Beijing South to Shanghai Hongqiao shows that all four models accurately predict the distribution. The coefficients of determination for GBDT, XGBoost, LightGBM, and CatBoost in the test set are 0.9664, 0.9601, 0.9680, and 0.9715 respectively. After optimizing the features, the order of importance in the contribution is as follows: benchmark train, ticket price, travel time, date, day of the week, and train code departure time. The coefficient of determination for the CatBoost-7 model in the validation set after feature optimization is 0.9458. Both the date and the benchmark train show a non-linear positive correlation with the passenger flow distribution prediction, while the travel time shows a non- linear negative correlation. In addition, low travel time, high ticket price and the benchmark train departing exactly at the scheduled departure time positively influence the passenger flow distribution prediction. This study provides valuable insights into the design of high-speed rail passenger transportation services. References | Related Articles | Metrics

Select

ARandom Parameters Frequency Model for Highway CrashAnalysis Considering Spatial Spillover Effects QI Weiwei, QIN Shuolei, ZHU Shufang, LIU Yan 2024, 24(2): 263-271.  DOI: 10.16097/j.cnki.1009-6744.2024.02.026 Abstract ( )   PDF (1449KB) ( )   To enhance the accuracy of parameter estimation for highway crash frequency models, this paper utilizes crash records, road attributes, traffic flow conditions, and weather condition data as samples to compares the fitting performances of several conventional crash frequency models. The Poisson-lognormal distribution model, which exhibits the best performance, is chosen as the foundational model for optimization. Then, the paper considers the spatial spillover effects of adjacent road segments and investigates additional spatial effects influencing crash frequency on highways. A model with spatial spillover covariates is developed to analyze the impact of spatial spillover effects on crash frequency of road segments in consideration of conditional autoregressive priors. Additionally, a random parameter model is developed to capture the influence of data heterogeneity on crash frequency of road segments. The results demonstrate the effectiveness of spatial spillover effects, and the goodness-of-fit of the proposed models have been improved compared to the control model. Based on the parameter estimation results of the optimal model, risk factors are identified, including "ln(MADT)", "ln(road length)", "category 1 vehicles", "category 4 vehicles", and "precipitation" as ordinary variables, as well as spatial spillover covariates such as "category 1 vehiclesS" and "curvature lengthS", which exhibit significant correlations with crash frequency. References | Related Articles | Metrics

Select

Interactive Effect on Traffic Accident Severity Considering Built Environment WANGJianyu, CHEN Xiantian, JIAO Pengpeng, QIN Chuliang, WANG Zehao 2024, 24(2): 272-280.  DOI: 10.16097/j.cnki.1009-6744.2024.02.027 Abstract ( )   PDF (1910KB) ( )   To explore the mechanism of various factors influencing traffic accidents under the impact of the built environment, this paper proposes a method that integrates the ADASYN (Adaptive Synthetic Sampling) balancing algorithm with the CatBoost model to study road traffic accidents in Shenyang from 2015 to 2020, and to analyze the interactive effects of accident causation. Firstly, by employing geographic information matching, the study supplemented 14 built environment factors around the accident locations with to construct a multi-source dataset. Secondly, by comparing four classic machine learning models—namely, CatBoost, Random Forest, XGBoost, and LightGBM—the study selected the model with the strongest generalization ability. Subsequently, the SHAP (Shapley Additive Explanation) attribution method was used to interpret the optimal model to reveal the effect of individual risk factors and their importance ranking. Finally, based on single-factor analysis, the study explored the interactive effects between the built environment and accident characteristics. The research indicates that the same features have different impacts on the mechanism of accidents in both single-factor and dual-factor interaction analyses. In single-factor analysis, two factors, season and mode of transportation, have a significant positive impact on fatal accidents; whereas five factors, including trunk road density, expressway density, industrial land proportion, site morphology, and physical road separation, have a significant negative impact on fatal accidents. In dual-factor interaction analysis, high trunk road density interacting with autumn and winter seasons, and low industrial land proportion interacting with spring, have a positive impact on fatal accidents; while a high industrial land proportion interacting with pedestrian traffic has a negative impact. The findings of this study offer precise insights into the factors influencing the severity of traffic accidents, providing a theoretical foundation for optimizing and developing urban transportation systems. References | Related Articles | Metrics

Select

Location and Capacity Planning of Electric Bus Charging Station in Cold Regions HUXiaowei, SONG Shuai, QIU Zhenyang, WANG Jian 2024, 24(2): 281-292.  DOI: 10.16097/j.cnki.1009-6744.2024.02.028 Abstract ( )   PDF (1973KB) ( )   Low-temperature environments in cold regions can lead to a decrease in the travel ranges of electric buses (EB), which affects the service area and planning of charging facilities. In order to improve the low-temperature adaptability of EB charging stations in cold regions, a location algorithm and capacity planning model for EB charging stations are proposed. Firstly, in the charging station location problem, the gradual coverage service radius of the charging station was constructed and an improved affinity propagation (AP) clustering algorithm was used to determine the charging station location. Based on the algorithm clustering centers, a Voronoi diagram of the charging stations was constructed to divide the charging clusters. In the capacity planning problem, a low-temperature capacity attenuation model for power batteries was constructed to determine the charging demand for EBs in cold regions. Based on the truncated queue theory model with limited capacity, the constraints associated with effective service intensity, denial of service rate, and charging satisfaction of charging stations were formulated. By introducing cost balancing coefficients, a charging station capacity planning model in cold regions was established to minimize the whole social cost within the planning period. A genetic algorithm (GA) was designed for the solution. Finally, the Harbin urban area is taken as an example to conduct the numerical study. The results obtain nine locations with charging clusters for charging stations as well as the number of charging facilities and the costs for each charging station. Based on the sensitivity analysis of environmental temperature and the cost balancing coefficients, the results show that the low-temperature environment in cold regions has a significant impact on the number of chargers and various costs of charging stations. Reasonably balancing the interests of charging stations and EBs can improve the satisfaction of charging service and reduce overall social costs. References | Related Articles | Metrics

Select

Carbon Emission Boundary Analysis for High-speed Railway Signal System SUPengfei, WANG Guo, MINYongzhi, TAO Jindou, LI Xudong, LIU Xinyue 2024, 24(2): 293-303.  DOI: 10.16097/j.cnki.1009-6744.2024.02.029 Abstract ( )   PDF (2443KB) ( )   Accurate quantification of carbon emissions of high-speed railway signal system is important for carbon emission estimation and the low-carbon development of high-speed railway. This paper considers the auxiliary activities of the signal system for the carbon emission calculation, and investigates the carbon emission boundary of the signal system. The proposed carbon emission calculation method for signal system is based on energy consumption. The carbon flow index of each device in the signal system is analyzed and calculated using the carbon flow theory. Taking the high-speed railway station of Lanzhou West in Gansu Province as a case study, this paper verifies the feasibility and effectiveness of the proposed method. The results show that the carbon emissions of the signal control and monitoring system are greatly related to the carbon emission factors of electricity, while the average carbon emission generated by the maintenance of the coded equipment is 0.12 kgCO2 ·d-1. The average branch carbon flow density among the devices in the signal system is 0.0076 kgCO2 ·kWh-1, and the node carbon potential of each device is equal to the branch carbon flow density. The carbon emission of each subsystem is associated with the type of the energy consumption. The indirect carbon emission caused by power consumption is the main component of carbon emission in the signal system. Direct carbon emissions from signal-system related activities are proportional to the primary energy consumption. References | Related Articles | Metrics

Select

Fleet Deployment and Voyage Planning Optimization for Liquefied Natural Gas Shipping Under Power of Mixed Boil Off Gas ZHAORuijia, SUN Yuxuan, NIU Dongxiang, XIE Xinlian 2024, 24(2): 304-312.  DOI: 10.16097/j.cnki.1009-6744.2024.02.030 Abstract ( )   PDF (1896KB) ( )   Liquefied natural gas ships can use the natural and forced boil off gas from their laden cargo as the propulsion power during navigation. This paper proposes an optimization method of fleet deployment and voyage planning for liquefied natural gas shipping powered by mixed boil off gas. A correlation function between liquefied natural gas ship speed, consumption volumes of boil off gas, and volumes of voyage delivery is constructed to balance the relationship between forced boil off gas of liquefied natural gas and voyage costs. Considering factors of owning and chartering ships, voyage transportation time, and boil off losses, this paper develops a management strategy of mixed boil off gas, and a joint optimization model for fleet deployment, voyage planning and speed of liquefied natural gas shipping is formulated with the goal of minimizing the annual operating cost of liquefied natural gas shipping companies. A nested solution algorithm based on model decomposition is proposed by combining genetic algorithm with optimization software. Using the actual operational data of liquefied natural gas shipping companies as a case study, this paper verified that the management strategy of mixed boil off gas can effectively reduce the operating cost by 7.13% and ship idle rate by 4.8%. The sensitivity analysis results indicate that as the spot price of liquefied natural gas increases, the management strategy of mixed boil off gas to adjust volumes of forced boil off gas has a more significant effect on reducing operating costs for shipping companies. References | Related Articles | Metrics

Select

Evaluation of Driver's Eye-catching Effect in Highway Tunnel Entrance Zone HANLei, DU Zhigang, MAAojun, JIAO Fangtong 2024, 24(2): 313-322.  DOI: 10.16097/j.cnki.1009-6744.2024.02.031 Abstract ( )   PDF (1947KB) ( )   To comprehensively and systematically evaluate the impact of visual attraction in the entrance zone of highway tunnels on the visual performance of drivers' eye-catching effects, this study carried out the natural driving experiments with 30 subjects. The experiment collected eye movement data of drivers under different visual attraction conditions in the entrance zone of highway tunnels, selected eye-catching effect sensitive evaluation indicators based on factor analysis method, and constructed linear mixed effect models and data envelopment analysis models. The study then analyzed the impact characteristics and mechanism of visual attraction in the entrance zone of highway tunnels on the visual performance of drivers' eye-catching effects. The results show that the sensitive indicators of drivers' eye-catching effects are gaze duration, pupil diameter, scanning duration and scanning amplitude. Different visual attraction conditions in the entrance zone of highway tunnels have a significant impact on the visual performance and comprehensive efficiency of drivers' eye-catching effects, and are significantly affected by individual characteristics factors such as drivers' age and driving experience, while gender factors have no significant impact on them. The existence of visual attraction will affect the normal visual performance of drivers to varying degrees, reducing its rationality and effectiveness. The visual attention level of drivers under the visual attraction condition of prompt slogan is the worst, the degree of visual cognitive load is the highest, and the negative impact of eye-catching effect is the greatest, followed by the billboard condition. References | Related Articles | Metrics

Select

Analyzing Heterogenous Effects of Contributing Factors Affecting Injury Severity of Helicopter Accidents ZHOUYue, JIANG Weian, FU Chuanyun, WEI Lin, YOU Yi 2024, 24(2): 323-332.  DOI: 10.16097/j.cnki.1009-6744.2024.02.032 Abstract ( )   PDF (2357KB) ( )   The study explores the contributing factors affecting the injury severity of helicopter accidents based on the modeling exercises that capture unobserved heterogeneity. The data include 1241 helicopter accident records collected from the National Transportation Safety Board (NTSB) database, categorizing the accident information into temporal, flight, pilot, helicopter, and meteorological factors. Afterward, the study uses Chi-square tests to examine the statistical differences of factors' probability distributions across injury severities (property-damage-only, minor injury, and serious injury or fatal). Several mixed Logit models are developed to model the injury severity, which set the random parameters with diverse priori densities. Through a meticulous comparison, the optimal model is selected based on data fitting metrics. Further, the estimates of this model are used to analyze the effects of the contributing factors on injury severities and their heterogenous effects. The results indicate that unobserved heterogeneity can be found in modeling the injury severity of helicopter accidents. The mixed Logit model with independent normal and multivariate normal random parameters outperformances others in data-fitting. Among the factors, encountering IMC, loss of control in flight, fuel issues, low altitude operation or external loading (including contact with powerlines or trees), maneuvering or operation, approaching, elder pilots, pilots with helicopter instrument qualification, and accidents involving fire or explosion significantly increase the probability of serious injuries or fatalities. For the factor's heterogenous effects, the coefficients of weekend flight, elder pilots, and the use of 3-point restraints can be modeled as random parameters. References | Related Articles | Metrics