Archive

25 April 2023, Volume 23 Issue 2 Previous Issue    Next Issue

---

For Selected:

Download Citations EndNote Ris BibTeX

Toggle Thumbnails

Select

Research Review of Influence of Social Network Information on Travel Behavior CHEN Jian, ZHANG Chi, FU Zhi-yan, LIU Ke-liang 2023, 23(2): 1-10.  DOI: 10.16097/j.cnki.1009-6744.2023.02.001 Abstract ( )   PDF (1616KB) ( )   PDF(English version) (478KB) ( 118 )   To quantitatively review the research results of the influence of social network information on travel behavior, this paper retrieved and screened 133 English and 32 Chinese literatures from 2010 to 2022 based on the database of Web of Science and China National Knowledge Infrastructure. Through the combination of knowledge graph and qualitative literature analysis, the paper quantified and counted three indexes of annual publication volume, research hotspot countries, and keyword graph. The research results were presented in four aspects, including research methodology, social network information behavior, the influence of social network information on travel decisionmaking, and the influence of social network information on travel activities. The results show that: (1) in terms of data sources, the basic data of existing research haven't achieved the integration of feature dimension and decision dimension, and it is necessary to further integrate multi-source data to improve the robustness of research conclusions. (2) In terms of research methods, the existing research lack mutual support among analysis methods, and a variety of research methods can be integrated to analyze the influence of social network information on travel behavior across disciplines. (3) In terms of research content, the existing research results cannot fully reflect the development trend of future travel, and the heterogeneity of travelers can be given more attention. It is necessary to analyze the connection mode between social network information and travel behavior considering traveler heterogeneity in combination with new scenarios such as autonomous driving and shared travel. References | Related Articles | Metrics

Select

Carbon Emission Efficiency Evaluation and Driving Factors Analysis of Logistics Enterprises ZHENG Yan, BEN Yu-shu, WANG Kang-de, JIANG Xiao-hong 2023, 23(2): 11-21.  DOI: 10.16097/j.cnki.1009-6744.2023.02.002 Abstract ( )   PDF (1570KB) ( )   In recent years, while the logistics industry is promoting economic growth, its energy consumption is also increasing year by year. The environmental pollution caused by logistics enterprises in the transportation process cannot be ignored. To facilitate the energy-saving and low-carbon development of logistics enterprises, the "top-down" method is used to calculate the mobile carbon emissions of land transportation and air transportation modes. This paper selects energy consumption, capital stock and labor as input variables, and business income and carbon emissions as output variables to develop a Super-efficiency Slack-based Measurement (Super-SBM) model, which is used to analyze the efficiency of carbon emission. In addition, the Logarithmic Mean Divisia Index (LMDI) is used to decompose the driving factors of carbon emissions into four categories. Their impact on carbon emissions of logistics enterprises is analyzed. This paper takes Shunfeng Express Enterprise as an example, and evaluates its efficiency of carbon emission according to its operating data from 2016 to 2021. The results show that the comprehensive technical efficiency value continues to grow and remain above 1.0 after 2019. The pure technical efficiency value fluctuates in the range of 1.1. The scale efficiency values are all below 1.0 but continue to grow. The results show that Shunfeng can carry out technological innovation on the original scale, and strive to obtain the highest output with the least input. Thus their resources can be reasonably allocated. Moreover, the level of economic development and population size both are the influencing factors that related to carbon emissions. The energy efficiency largely inhibits the increase of carbon emissions. The influence of energy consumption structure in promoting carbon emissions is relatively limited. The experimental results verified the effectiveness of the proposed measurement method. The Super- SBM and LMDI models can effectively evaluate the efficiency of carbon emission of logistics enterprises and analyze the influence of driving factors on carbon emission. References | Related Articles | Metrics

Select

Monitoring and Assessing Carbon Footprint of Individual Trip Chain in Environment of Mobility as a Service LI Wen-xiang, CHENG Jia-nan, LIU Xiang-long, MU Kai, CAI Jin-jin, LIU Wei-wei 2023, 23(2): 22-31.  DOI: 10.16097/j.cnki.1009-6744.2023.02.003 Abstract ( )   PDF (3014KB) ( )   Under China's strategic goal of "carbon peaking and carbon neutrality", scientific and accurate monitoring and evaluation of the carbon footprint of individual travel is the basis for promoting the low-carbon transformation of urban transportation, but it also faces great challenges. Therefore, based on the data openness and sharing of the Mobility as a Service (MaaS) platform, a carbon footprint monitoring and evaluation method for urban individual trip chains in the MaaS environment is proposed. Firstly, we designed an urban transportation carbon source monitoring index system based on the MaaS platform and realized the extraction and integration of multi-modal transportation characteristics of users' individual trip chains. Then, we established the carbon emission calculation models for trips by private motor vehicles and rail transit, respectively, and the carbon emissions of different transportation modes are calculated and then added up to obtain the carbon footprint of individual complete trip chains. Finally, the carbon emission reduction of the individual trip chain was evaluated by using the baseline scenario of car trips. The case analysis of 8424 travel sections collected in Beijing shows that the average person-kilometer carbon emissions of the trip chain dominated by motor vehicles, buses, rail, and non-motor vehicles are 0.238 kg · pkm- 1 , 0.031 kg · pkm- 1 ,0.039 kg · pkm-1 , and 0.0017 kg · pkm-1 , respectively, and the average person-kilometer carbon reduction to the baseline is 0.029 kg · pkm-1 , 0.22 kg · pkm-1 , 0.23 kg · pkm-1 , and 0.28 kg · pkm-1 , respectively. The carbon emission reduction of person-kilometers in the trip chain is positively correlated with the proportion of green travel in the trip chain. Electrifying vehicles on MaaS platforms could increase carbon reduction benefits by 52.5%. References | Related Articles | Metrics

Select

Influencing Factors and Decomposition of Social Logistics Cost Evaluation Indicators in China and United States FAN Dong-fang, JIN Zhi-hong 2023, 23(2): 32-39.  DOI: 10.16097/j.cnki.1009-6744.2023.02.004 Abstract ( )   PDF (1330KB) ( )   This paper proposes a two-stage research method to analyze the difference between the internal influencing factors of the social logistics cost evaluation indicator "the proportion of social logistics cost to Gross Domestic Product (GDP)" between China and the United States. The“logistics science and technology level (calculated by the R&D, Research and Development)”is introduced as one of the internal influencing factors of the evaluation indicator, and its relevance to social logistics costs is discussed. The "proportion of social logistics costs to GDP" is decomposed into the product of "logistics rate per unit turnover", "average transportation distance" and "freight volume per unit GDP”. The logistics costs of China and the United States are evaluated and analyzed from these three dimensions. In the first stage, the Grey System Correlation theory is used to calculate the correlation between the three indicators (GDP, the proportion of three industries, and the R&D level) and the evaluation index. The results show that the correlation between the three reference indicators between China and the United States and the evaluation index is positive, but the difference between GDP and R&D level is obvious. In the second stage, based on the data- driven method, the three-dimensional evaluation indicators obtained by decomposition are compared and analyzed respectively. In fact, the results show that the“per unit turnover logistics cost”in China is lower than that in the United States, but the GDP output of the same freight volume in the United States is much higher than that in China. China's industrial structure reform should promote the joint development of logistics and service industries. Then the specific suggestions are provided for reducing costs and increasing efficiency of China's logistics industry based on the analysis results. References | Related Articles | Metrics

Select

Inferring Spatial-temporal Travel Patterns of Vehicles Combining Topology of Trips and Sequence Analysis JIN Sheng, SU Hong-yang, ZHANG Jing 2023, 23(2): 40-53.  DOI: 10.16097/j.cnki.1009-6744.2023.02.005 Abstract ( )   PDF (4496KB) ( )   The analysis of spatial-temporal characteristics of vehicle travel is the basis of urban traffic planning, design, and traffic demand management. License Plate Recognition (LPR) data can detect every vehicle in a road network, breaking the limitations of aggregated or sampled data used in traditional travel behavior research. In this paper, LPR data was employed to infer spatial-temporal travel patterns of individual vehicles. Firstly, trip chains were extracted from LPR data and then cut off to obtain vehicle trips with time, location, frequency, and topology features. Travelactivity sequences were constructed based on the locations where vehicles stay during the day. Point of Interest (POI) data was used to identify the land use characteristics associated with trip origin- destination that were regarded as features of staying location. A total of 30 spatial-temporal vehicle travel patterns were mined using the k-modes algorithm. It is found that all the travel patterns were divided into four categories: regular commuting pattern, special commuting pattern, short-term activity pattern, and non-native travel pattern. The population, features, and typical behaviors of each pattern are discussed. The results show that 95% of vehicle travel activities can be represented by a simple topology structure consisting of less than 3 edges, 30% of which can be used for constructing activity sequences. The k-modes clustering algorithm is capable to distinguish various spatial-temporal travel patterns of vehicles based on activity sequence. Regular commuting pattern dominates vehicle travels on weekdays, while shortterm activity pattern is the majority on weekends. With the analysis of microscopic behaviors, the combination of topological structure and activity sequence benefits the mining of individual vehicle travel patterns. This paper provides theoretical support for detailed vehicle travel behavior analysis as well as traffic management and control strategy for decision-makers. References | Related Articles | Metrics

Select

Platoon Control Strategy for Connected and Automated Vehicles Under Intermittent Communication Failures LIU Run-kun, YU Hai-yang, CHENG Meng-yue, REN Yi-long 2023, 23(2): 54-66.  DOI: 10.16097/j.cnki.1009-6744.2023.02.006 Abstract ( )   PDF (2838KB) ( )   There will be intermittent communication failure when the Connected and Automated Vehicles (CAVs) platoons drive in special areas. In order to improve vehicle control stability in the case of intermittent communication failure, this research proposes a new platoon control strategy. First, we design an upper Cooperative Adaptive Cruise Control (CACC) controller structure, named Multi-information Predicted CACC (MIP-CACC), which includes a predictor to predict the state information of the vehicle ahead. Then, a Dynamic Low-Rank Tensor Prediction (D-LRTP) model is constructed that can be applied to this controller. By analyzing the string stability of the controller, the stability is affected by the combination of information weight, number of communication connections, communication delay, cut-off frequency, time headway, and other parameters. There is a minimum time headway that satisfies string stability in different communication and control scenarios. Finally, the control strategy proposed in this study is compared with the traditional control strategy. The simulation results show that the control strategy proposed in this study has less disturbance to the driving state of the platoon under intermittent communication failures. Driving speed and acceleration disturbances can be reduced by 79.43% and 72.53% respectively. This strategy has reference significance for the design of CAVs control scheme with communication robustness. References | Related Articles | Metrics

Select

Configuration of Spatial-temporal Resources at Intersections Considering Travel Efficiencies of Mixed Flows QIAN Guo-min, REN Lu, JI Qing-yuan, FENG Yuan-jing, ZHANG Li-hui, WANG Dian-hai 2023, 23(2): 67-73.  DOI: 10.16097/j.cnki.1009-6744.2023.02.007 Abstract ( )   PDF (1721KB) ( )   To optimize the configuration of spatial-temporal resources at intersections with mixed traffic flow of passenger cars and heavy vehicles, this paper proposes a set of mixed integer linear programming (MILP) model with aim to maximize the intersection capacity. In the model, different impacts of the locations of permitted lanes on flows of different modes and vehicle turning maneuvers are considered to increase the model applicability. The constrains on the relative setting position of permitted lanes for left-turn, through and right-turn are released. That's to say, the permitted left-turn lanes don't have to be on the left of the permitted through lanes, and the permitted through lanes don't have to be on the left of the right-turn permitted lanes. This may induce some new flow conflicts (potential conflicts) and whether these conflicts exist is affected by the lane channelization scheme. Thus, some constraints are added in the model, among which one is to identify which potential conflicts are arisen with specific lane scheme, the other is to determine these conflicts' impact on the configuration of temporal resources at intersections. The results from numerical tests show that this impact will change the configuration of spatial-temporal resources at intersections. Compared with the conventional lane-based model, the modified model can significantly improve the intersection maximum capacity while ensure safe flow propagation. The improvement on the intersection maximum capacity increases with the heavy vehicle proportions and the impact of lane location on travel efficiency. But with the increase of the difference of travel efficiencies between heavy vehicles and passenger cars, the improvement first increases and then decreases. References | Related Articles | Metrics

Select

Inverse Identification Method of Urban Traffic Congestion Source Based on Typical Coupled Optimization Algorithm ZHAO Xue-ting, HU Li-wei 2023, 23(2): 74-83.  DOI: 10.16097/j.cnki.1009-6744.2023.02.008 Abstract ( )   PDF (2236KB) ( )   This paper applies simulation-optimization methods to identify the location of urban traffic congestion sources and data derivation. First, a hypothetical urban traffic congestion scenario is established and the model is improved by introducing the groundwater contamination mass transport model and considering the urban traffic congestion characteristics. The continuous field of urban traffic outflow rate is proposed to describe the inhomogeneity of urban traffic congestion based on the Cholesky decomposition method. Then, the Kriging and back propagation (BP) neural network are used to develop an alternative model for the numerical simulation model of urban traffic congestion. The accuracy of the alternative model is tested by the mean relative error, deterministic coefficient and root mean square error. The optimized model is solved through the Sparrow Search Algorithm (SSA) and Genetic Algorithm (GA), and the inverse identification results are tested by the average relative error. The results show that (i) using the Cholesky decomposition method, the urban traffic outflow rate is unevenly distributed and consistent with the characteristics of urban traffic congestion heterogeneity. The obtained mean value is 322.15. The result is at a medium diffusion level. (ii) The Kriging alternative model is more accurate than Back Propagation (BP) neural network alternative model, and the average relative error is 0.98%. (iii) Both the SSA and GA can be applied to identify the location of urban traffic congestion sources and traffic diffusion volume fast and accurately. Compared to the GA, the SSA improves the overall relative error of traffic congestion source location by 1.68% and improves the overall relative error of traffic volume by 2.52%. The proposed method can effectively identify urban traffic congestion sources and traffic diffusion traffic volumes with high accuracy, which can provide important references for urban traffic congestion source control and making traffic diffusion control schemes. References | Related Articles | Metrics

Select

Collaborative Optimization Model of Truck Speed and Signal Timing Based on Minimum Fuel Consumption ZHANG Peng, GU Yun-xiang, SUN Chao, LI Wen-quan 2023, 23(2): 84-91.  DOI: 10.16097/j.cnki.1009-6744.2023.02.009 Abstract ( )   PDF (1779KB) ( )   This paper proposes an integer nonlinear programming model for collaborative optimization of truck speed and signal timing to reduce the fuel consumption caused by frequent stopping and start of trucks at signalized intersections. The objective function of the model was the truck travel fuel consumption based on the Virginia Tech Comprehensive Power-based Fuel Consumption Model. Truck travel fuel consumption included road fuel consumption, intersection stopping fuel consumption and speed recovery fuel consumption. The road section was divided into acceleration area, uniform speed area and deceleration area. Considering the acceleration process of the multi-gear gradient of the truck, the upper and lower limits of acceleration and shift time were defined for each gear of the truck. The queue dissipation process at the downstream intersection was analyzed using the traffic wave theory, and the spacetime trajectory constraint of the truck was considered. The optimization variables include acceleration at each gear of the truck, acceleration time, uniform speed, uniform driving time, deceleration, deceleration time, intersection green light extension time and red light early return time. The example analysis included 10 typical truck arrival situations generated at uniform intervals. The results show that compared with no speed guidance, the fuel consumption from the proposed model is reduced by 29.9 % at the maximum, 3.9 % at minimum, and 13.7 % on average. The number of truck stops is reduced by 71.4 %, and the total stopping time is reduced by 174 seconds (89.2 %). The model effectively reduces the fuel consumption and number of stops for trucks crossing the intersections. References | Related Articles | Metrics

Select

Truck Stop Purpose Identification Method Based on Trajectory Data YAO Ya-hui, ZHANG Rong 2023, 23(2): 92-99.  DOI: 10.16097/j.cnki.1009-6744.2023.02.010 Abstract ( )   PDF (1772KB) ( )   The truck trajectory data can provide a wide range of information on freight demand, but the method for identifying stop purposes from the trajectory data needs further research. Using the trajectory data of intracity trucks, this paper designs a truck stop extraction algorithm based on reverse inference of movement records and proposed a stop-purpose calibration method incorporating data from the truck travel survey data. A feature set containing five feature subsets: stop features, trip features, trip chain features, nearby stop features, and nearby point of interest features is constructed and used as input variables for the identification model. A truck stop purpose identification model based on the binary tree Support Vector Machine (SVM) is developed using the grid search algorithm to determine the optimal combination of parameters. The model can distinguish between loading stops, unloading stops, and non-loading stops. Taking heavy goods vehicles traveling within Shanghai as an example, this paper applies the truck stop purpose identification method based on trajectory data. The results indicate that: trucks stop on average 4.6 times a day. Besides, more than 95% of stops have a stopping time greater than 7 minutes and a movement time greater than 17 minutes between two consecutive stops. The error in the number of stops of the truck stop extraction algorithm is about 6.5%. And setting the time-matching criteria and space-matching criteria is a feasible method to calibrate the stop purpose. Furthermore, the accuracy of the truck stop purpose identification model based on the binary tree SVM reaches 85.83%. This study could provide technical and data support for mega-city managers to analyze urban freight demand and formulate urban freight management policies. References | Related Articles | Metrics

Select

Impact of Built Environment on Flow of Transfer Passengers Between Subway and Bus Considering Spatial Heterogeneity LI Xiang, YAN Qi-peng, LUO Chen 2023, 23(2): 100-110.  DOI: 10.16097/j.cnki.1009-6744.2023.02.011 Abstract ( )   PDF (6558KB) ( )   The travel mode of subway and bus interchange has become the main travel mode in densely populated cities. Exploring the influencing factors of the interchange between the two modes can help to improve the public transportation sharing rate. In this paper, we use AFC data and AVL data to identify the interchange flows of metro-tobus and bus-to-metro modes. With the metro station as the core, the built environment index system for metro stations is built in four dimensions: the degree of development around the metro station, the traffic system, the urban design, and the structural characteristics of the metro network. A multi-scale geographically weighted regression (MGWR) model is used to investigate the mechanism and scale effects of the urban built environment on the interchanges. An empirical study is conducted in Chengdu. The results show that the MGWR model can characterize the regional differences between the interchange flows and the built environment and its scale of influence, with better estimation results than the global OLS and GWR models. There is spatial heterogeneity in the effects of built environment elements on the bus and metro interchange passenger flows, with the most significant spatial heterogeneity in the number of bus lines and weaker spatial heterogeneity in the density of non-motorized lanes, land use mix, and the number of metro lines. The effect of different built environment elements varies on interchange flows. The number of bus stops and the number of routes have a greater impact on interchange flows, while non-motorway density has a significant inhibiting effect. References | Related Articles | Metrics

Select

Optimization of Bus Bridging Considering Multi-station Coordination Under Metro Disruption SUN Hui-jun, FENG Zong-xu, ZHENG Han-kun 2023, 23(2): 111-120.  DOI: 10.16097/j.cnki.1009-6744.2023.02.012 Abstract ( )   PDF (2428KB) ( )   In the event of unexpected disruptions to the metro system that require the operation to be sectioned, the evacuation of stranded passengers to the nearest turnover station can result in congestion. To enhance the efficiency of passenger evacuation, this paper proposes the implementation of an emergency bridging bus service to transfer stranded passengers to the nearest turnover or transfer stations based on their travel direction. The proposed emergency bus bridging scheme involves cooperation between the turnover and transfer stations to facilitate the evacuation task. A mixed integer programming model is established, which considers constraints such as dispatching of bridging bus lines, connection of bus lines, time connection of connecting lines, passenger flow evacuation relationship, and passenger flow control at evacuation destination stations. The objectives of the model are to minimize passenger waiting time and maximize bus load rate. To solve the model, a hierarchical sequence algorithm nested improved genetic algorithm is developed. The model and algorithm are verified using Beijing Metro Line 4 as an example. The results demonstrate that the proposed bus bridging scheme reduces the bus load rate by 2.5%, while the average waiting time of passengers is decreased by 30.2%, significantly enhancing the efficiency of evacuation. References | Related Articles | Metrics

Select

Effects of Built Environment on Metro Ridership Considering Stage of Growth 2023, 23(2): 121-127.  DOI: 10.16097/j.cnki.1009-6744.2023.02.013 Abstract ( )   PDF (1709KB) ( )   To more accurately grasp the generation law of metro ridership, the relationship between metro ridership and the surrounding built environment is explored from the perspective of the growth stage. Taking Shanghai Metro as the studied case, the built environment is described by 14 factors such as population and employment density, land use, road network density, the number of entrances and exits, betweenness, etc., based on multi-source data including Shanghai smartcard data, population and economic census data, Point of Interest (POI), and road network. The Ordinary Least Square (OLS) model and the eXtreme Gradient Boosting (XGBoost) model are used to quantify the effects of the built environment on metro ridership. The results show that the XGBoost model based on a machine learning algorithm has better model performance than the OLS model. As for the contribution of independent variables, in the early stage, the number of entrances and exits (21.9%), the density of the population (15.9%), and the density of the road network (9.8% ) are the most important built environment factors affecting the metro ridership. In the short term, the built environment such as commercial land (16.5% ), floor area ratio (11.1% ), and job density (8.5% ) have become the key to improving subway passenger flow. In the long term, metro ridership depends on the level of integration between land use and transportation, such as the number of entrances and exits (18.9% ), commercial land development (16.6% ), and the number of transfer lines (7.7% ). The results confirm the progress characteristic relationship between metro ridership and the built environment around the station, which provides an important reference for formulating the integrated development strategy of transit-oriented development according to the time and contexts. References | Related Articles | Metrics

Select

Investigating Influencing Factors on Metro-bus Transfer Demand Incorporating Spatial Heterogeneity Based on Multi-source Data ZHENG Yue, GAO Liang-peng, CHEN Xue-wu 2023, 23(2): 128-138.  DOI: 10.16097/j.cnki.1009-6744.2023.02.014 Abstract ( )   PDF (3113KB) ( )   As an important part of multi-mode public transport, the transfer between metro and bus is a key link for urban passenger transport integration. Based on the multi-source data of Nanjing City, this paper analyzes the transfer demand between the metro and bus. To better understand the forming mechanism of the transfer demand, a multi-scale geographically weighted regression (MGWR) model was constructed to reveal the impact and spatial heterogeneity of four kinds of factors, including shared bicycle usage, bus supply characteristics, transfer accessibility, and subway network characteristics. The results show that the MGWR model has stronger explanatory power than the linear regression model and the geographically weighted regression (GWR) model, and the influencing factors of metro-bus transfer volume have significant spatial heterogeneity. The increase in bus operation shifts and the number of accessible bus stops can significantly promote the metro-bus transfer. The number of residential POIs promotes the transfer demand, while the number of enterprise POIs inhibits the transfer demand in suburban areas. Shared bicycle usage is revealed to be negatively correlated with the metro-bus transfer demand, especially in suburban areas. References | Related Articles | Metrics

Select

Metro Passenger Flow Prediction Based on Dynamic Spatio-temporal Neural Network Model SHI Jun-qing, LI Ruia, CHENG Ming-huia, RUAN Jun-huia, XIE Xinga 2023, 23(2): 139-147.  DOI: 10.16097/j.cnki.1009-6744.2023.02.015 Abstract ( )   PDF (2247KB) ( )   This paper proposes a Dynamic Spatio-Temporal Neural Network (DSTNN) model based on attention mechanism for urban rail transit station passenger flow forecast. The DSTNN adopts a multi-branch parallel architecture to effectively extract the complex spatio-temporal features of metro passenger flow. In the spatial dimension, the global and local attention mechanisms are combined to capture dynamic spatio-temporal correlation between stations and static topology. In the temporal dimension, the bi-directional long short-term memory and attention mechanisms are used to learn the time-varying patterns of passenger flow data. In the experiments on Hangzhou Metro dataset, the results show that the DSTNN has higher prediction accuracy and training efficiency compared to classical prediction models and deep learning models. The average mean absolute error (MAE) over four different prediction durations is respectively 6.63% and 2.57% lower than that of the Diffusion Convolutional Recurrent Neural Network (DCRNN) and Physical-Virtual Collaboration Graph Network (PVCGN). In addition, the visualization analysis demonstrates the dynamic learning ability of this model for spatio-temporal correlations, and the ablation experiments verified the effectiveness of each branch. References | Related Articles | Metrics

Select

Bus Arrival Time Prediction Model Based on Bidirectional Long Short-term Memory Network ZHANG Bing, ZHOU Dan-dan, SUN Jian, NI Xun-you 2023, 23(2): 148-160.  DOI: 10.16097/j.cnki.1009-6744.2023.02.016 Abstract ( )   PDF (3951KB) ( )   To improve the accuracy of bus arrival time prediction and increase the bus usage in the cities, this paper proposes a bus arrival time prediction model based on a bidirectional Long Short-term Memory (BiLSTM) neural network and the hyperparameter search. The improved seagull algorithm optimization adding Attention mechanism to bidirectional LSTM (ISOA-BiLSTM-Attention) prediction model was developed by introducing nonlinear convergence factor, sine cosine operator, and adaptive parameters to improve the seagull algorithm to achieve hyperparametric optimization of the bidirectional LSTM model. The Attention mechanism was added to improve the information processing ability of bidirectional LSTM. Then, the trajectory data of bus route 220 in Nanchang, Jiangxi Province of China, were used to predict the bus arrival time for different directions and time to validate the model prediction accuracy. The results show that, the proposed model has better performance than the traditional bidirectional LSTM model. The improved seagull algorithm can achieve a better optimization effect on the bidirectional LSTM-Attention model. Compared with the existing model and seagull algorithm (SOA) optimized bidirectional LSTM-Attention model, the mean absolute percentage error was reduced by 5.96%, the root mean square error was reduced by 9.87%, and the mean absolute error was reduced by 7.99% in the ISOA-BiLSTM- Attention for bus arrival time prediction. Moreover, the ISOA-BiLSTM-Attention has the largest model decision coefficient R2 value, which indicates the good generalization ability and stability of the proposed model, and can provide good fitness of accuracy for bus arrival time. References | Related Articles | Metrics

Select

Influence of Individual Attribute and Built Environment on Convenience of Public Transportation for Elderly LI Kang-kang, YANG Dong-feng 2023, 23(2): 161-167.  DOI: 10.16097/j.cnki.1009-6744.2023.02.017 Abstract ( )   PDF (1623KB) ( )   Existing studies focus mostly on the spatial layout of transportation facilities such as public transportation stations and hubs, and on natural environmental factors such as terrain and climate, and few studies focus on the impact of the built environment on the convenience of public transportation for the elderly. This paper analyzes the impact of individual socio-economic attributes and the built environment on the elderly's public transportation convenience by constructing a multi-level linear model. It is found that age growth has obvious restrictions on the convenience of public transportation for the elderly, and education level and living years can alleviate the lack of adaptability of the elderly to the traffic environment caused by physiological decline. The male elderly is more aware of the convenience of public transportation than females. There is no significant relationship between empty nests and not. The intersection density, road network density, NDVI, and public service facility density at the connection path level are the key built environment elements that affect the convenience of public transportation for the elderly. Among them, the density of intersections will aggravate the restriction of age and living years on the convenience, while the road network density has a positive regulating effect. Public facilities have a greater impact on the male elderly, and the elderly with higher education levels show a high-quality demand for NDVI. References | Related Articles | Metrics

Select

Elderly's Acceptance of Autonomous Vehicles in Context of Digital Divide LIU Zhi-wei, LIU Jian-rong 2023, 23(2): 168-175.  DOI: 10.16097/j.cnki.1009-6744.2023.02.018 Abstract ( )   PDF (1573KB) ( )   Autonomous vehicles can increase mobility and improve the quality of life of the elderly, and it is important to study the acceptance of autonomous vehicles by the elderly. Firstly, a latent class analysis was established to divide the sample into two latent classes: the elderly with a weak digital ability and the elderly with a strong digital ability. Then, with all the samples, three comprehensive models with the technology acceptance model and the theory of planned behavior were established to study the acceptance of autonomous vehicles by the elderly with different digital abilities. The results show that the elderly with a weak digital ability and the elderly with a strong digital ability account for 27.93% and 72.07% , respectively. Significant heterogeneity exists in the acceptance of autonomous vehicles among the elderly population. For all elderly people, perceived behavioral control is the most important factor influencing the acceptance of autonomous vehicles. For the elderly with weak digital ability, the subjective norm is the most important factor affecting the acceptance of autonomous vehicles, while for the elderly with strong digital ability, attitude is the most important factor affecting the acceptance of autonomous vehicles. References | Related Articles | Metrics

Select

Airport Taxi Supply and Demand Equilibrium Game Model Considering Ride-hailing Competition HUANG Ai-ling, LIU Meng-han, LI Ying 2023, 23(2): 176-186.  DOI: 10.16097/j.cnki.1009-6744.2023.02.019 Abstract ( )   PDF (2504KB) ( )   PDF(English version) (1788KB) ( 87 )   To balance the supply and demand of taxi at airports, this paper proposes a strategy game model for taxi drivers and passengers decision-making based on the static non cooperative game theory under complete knowledge. The competition influence coefficient (CIC) is introduced in the profit function for taxi driver, and changes in driver's decision-making behavior are analyzed considering the impact of the ride-hailing services. The benefit function of passenger group is proposed to reflect the impact of different decision-making results of passenger group on drivers, and the major factors include queue length, boarding speed, traffic conditions, traffic fare, and comprehensive impact of multiple modes of transportation. Using Beijing Capital International Airport as an example for the empirical analysis, the results show that: when the two groups reach Nash equilibrium (NE) through mutual feedback, the overall supply of airport taxi is slightly higher than the demand under the average competition level of Beijing's ride-hailing services. It also shows a state of oversupply in the morning and evening peak hours, while in the early morning the demand is far greater than the supply with the closure of subway. Taxi drivers' awareness of the effects of online taxi competition can be suitably increased considering the mismatch between the daily supply and demand of airport taxis under the existing level of ride-hailing competition when it reaches NE. References | Related Articles | Metrics

Select

Collaborative Optimization of Urban Rail Timetable and Flow Control Under Mixed Passenger and Freight Transportation PAN Han-chuan, LU Jun-bo, HU Hua, LIU Zhi-gang, SHA Yue 2023, 23(2): 187-196.  DOI: 10.16097/j.cnki.1009-6744.2023.02.020 Abstract ( )   PDF (3311KB) ( )   With the development of the economy and society, ground traffic road congestion is serious, resulting in increased carbon emissions. Due to the advantages of energy saving, environmental protection, and emission reduction of urban rail transit, full use of the redundant transportation capacity of rail transit is currently a feasible solution to alleviate the pressure of ground cargo transportation and reduce carbon emissions. This paper studies the timetable and flow control problem under the mixed passenger and freight transportation mode. Firstly, a model for mixed passenger and freight transportation in an urban rail line is constructed, which takes the departure time of the train, the layout of the train carriages, and the number of passenger and freight demand allocated to the train compartment as the decision variables to minimize the waiting time of passengers and goods and the energy consumption of the train compartment, considering flow balance constraints, train capacity limitation, and timetable. To verify the validity of the model, Shanghai Metro Line 17 is taken as an example for empirical research, and the model is solved by the optimization solver Gurobi. The results show that the collaborative optimization method proposed in this paper has a good optimization effect and computational efficiency. Compared with the one-by-one solution method, the number of passenger and cargo delays can be significantly reduced. Collaborative optimization can reduce the number of passenger delays by 21.92%, the number of cargo delays by 9.73%, the average waiting time of passengers and cargo by 35.88% , 25.56% , and carbon emissions by 1.7% . This method can improve the load rate, reduce the waste of capacity during the peak period, and improve the safety and efficiency of rail transit operations. References | Related Articles | Metrics

Select

Collaborative Optimization of Demand-oriented Train Timetabling and Stop Planning for Intercity Railways TIAN Xiao-peng, NIU Hui-min, HAN Ying 2023, 23(2): 197-207.  DOI: 10.16097/j.cnki.1009-6744.2023.02.021 Abstract ( )   PDF (2995KB) ( )   To collaboratively optimize train timetables and stop plans for intercity railways, this paper used the hourdependent origin-destination passenger demand as the input. Considering variable train running times on segments, the paper developed a bi- objective linear integer programming model including train safe operation constraints and passenger demand loading constraints, and the construction of the objective functions depending on two aspects of train operation efficiency and passenger travel quality. Combining the characteristics of the proposed model, the bi-objective optimization model was transformed into a single-objective model by using the ε-constraint method, and then a branchand-cut algorithm was designed based on two sets of valid inequalities. Several different-sized numerical experiments on Guangzhou- Zhuhai intercity railway were conducted to assess the effectiveness of the proposed approach. The results show that the proposed approach can efficiently solve the real-life problems, and the obtained train timetables can highly match the dynamic passenger demand distributions. Specifically, compared with the non-peak periods, the peak periods account for 70% of the train lines and 69% of the train stops. For medium-scale and large-scale problems that cannot be solved directly using GUROBI, the proposed method can obtain satisfactory feasible solutions within an acceptable computational time, showing good solution performance. References | Related Articles | Metrics

Select

A Timetable Optimization Method Considering Train Operation Net Energy Consumption and Passenger Load Difference ZHANG Hui-ru, DOU Fei, WEI Yun, LIU Jie, NING Yao 2023, 23(2): 208-216.  DOI: 10.16097/j.cnki.1009-6744.2023.02.022 Abstract ( )   PDF (2510KB) ( )   Urban rail transit is of great significance to alleviate the huge passenger flow pressure brought about by the urbanization process. However, the daily operation always consumes a considerable amount of energy. An effective and feasible energy-saving method is to optimize the train operation strategy while keeping the existing infrastructure unchanged. In this paper, considering the spatial distribution of passenger flow, a timetable optimization model for the energy-saving operation of trains is established to make the net energy consumption of trains and passengers' traveling time achieve bi-objective optimization. A simulation-based non-dominated sorting genetic algorithm is designed to solve the model. With the optimal driving strategy of maximum acceleratio-cruising-coasting-maximum deceleration braking, a set of Pareto optimal solutions of time vs. energy consumption are obtained by optimizing the running time in the interstation sections and the dwell time at stations. A series of experiments are carried out using the actual data of the Changping Line in Beijing Metro. The results show that Pareto optimal solutions with uniform distribution and complete convergence under different passenger load conditions can be obtained by the proposed optimization model. Compared with the original timetable, the net energy consumption optimization rate of fixed train passenger load and the one considering the spatial difference of train passenger load can reach 5.5% and 18.79% respectively. References | Related Articles | Metrics

Select

Comprehensive Optimization of Operation Planning and Pricing of Backhaul for Heavy-haul Railways Transportation TONG Rui-yong, MAO Bao-hua, DU Peng, WEI Run-bin, HUANG Jun-sheng 2023, 23(2): 217-224.  DOI: 10.16097/j.cnki.1009-6744.2023.02.023 Abstract ( )   PDF (1691KB) ( )   Heavy-haul railway is an efficient method for single cargo one-direction transportation but there is a waste of capacity in backhaul. This paper proposes a bi-level multi-objective model to improve the railway sharing rate, increase revenues of railway sector, and reduce the environmental impact. The upper layer of the model aims to maximize railway sector's revenues and the carbon dioxide emission reduction rate. The lower layer minimizes shippers' generalized cost. Taking the heavy-haul railway from one company as an example, the study uses non-dominated sorting genetic algorithm with elite strategy (NSGA-Ⅱ) to solve the model and obtain the optimized plans for operation planning and pricing. Besides, the influences of government subsidies and carbon tax are also considered for the optimization. The results show that the algorithm obtains multiple sets of Pareto frontiers, and provides multiple optimization plans for decision makers. The revenues of railway sector are significantly negatively correlated with the carbon emission reduction rate on the routes of backhaul for heavy-haul railways. When road freight rate is 0.500 yuan per ton-kilometer, heavy-haul railway freight rate corresponding to maximum revenues is 0.285 yuan per ton-kilometer. Compared with scenario of roadway transportation, carbon emission reduction rate is increased by 40.62%. Moreover, when the value of government subsidies for railway sector is increased from 0 to 0.10 yuan per ton-kilometer, the maximum revenues of railway sector increase by 51.65%, and freight rates corresponding to the maximum revenues drops from 0.285 yuan per ton-kilometer to 0.225 yuan per ton-kilometer. In addition, carbon emission reduction rate increases from 40.62% to 49.11%. At last, when carbon tax increases from 0.0004 yuan per gram carbon emission to 0.0012 yuan per gram carbon emission, the carbon emission reduction rate increases by 0.26%, and the revenues of railway sector increases by 9.06% . Reasonable pricing plan, government subsidies and carbon tax schemes can effectively improve revenues and carbon emission reduction rates for heavy-haul railway backhaul transportation. References | Related Articles | Metrics

Select

Modelling and Simulation of Railway Station Check-in Process Considering Social Distance TANG Tie-qiao, ZHONG Jing-ran, YUAN Xiao-ting 2023, 23(2): 225-232.  DOI: 10.16097/j.cnki.1009-6744.2023.02.024 Abstract ( )   PDF (2294KB) ( )   With the COVID-19 pandemic in China gradually turned better, the railway passenger transport services are gradually back to normal, and the railway stations accommodate a large number of passengers every day. The social distancing between passengers will affect their behaviors under the "Category B management" policy in China. This study investigated the check-in area of Beijing South Railway Station and developed a cellular automaton model based on potential energy field. The paper also proposed the decision function of passenger queue spacing under the bounded rationality mechanism to study the influence of social distance between passengers on the ticket checking process based on the check-in process in Kunming Railway Station and Chongqing West Railway Station . The simulation results showed that: the passenger's average arrival time and the total check-in time are affected by different strength of "bounded rationality". Under the experimental conditions of this paper, the close contact times of passengers are positively correlated with the strength of bounded rationality. In the queuing process, when the difference between the initial selection of passengers increases by 0.4 meters, the average share of gate increases by 3.3%, the average arrival time increases by 14.9%, the average check-in time increases by 20.5%, and the ratio of check-in time to arrival time of passengers increases by 5.7%. In addition, the increase of the proportion of completely rational passengers will increase the passenger average arrival time. When the proportion of completely rational passengers is small, the decrease of the proportion has little effect on the check-in time, but the number of close contacts between passengers increases accordingly. References | Related Articles | Metrics

Select

Semantic Segmentation of Railway Scene Based on Reticulated Multi-scale and Bidirectional Channel Attention LU Tong, YU Zu-jun, GUO Bao-qing, RUAN Tao 2023, 23(2): 233-241.  DOI: 10.16097/j.cnki.1009-6744.2023.02.025 Abstract ( )   PDF (2245KB) ( )   Semantic segmentation is the basis of intelligent perception. This paper proposes a semantic segmentation network for railway scenes based on reticulated multi-scale fusion and bidirectional channel attention to address the difficulties that railway scene categories are complex and effective features are challenging to extract. To enhance the discriminant ability of the model for various railway facilities, a reticulated multi-scale fusion module is proposed. The module is embedded in the backbone network to obtain the parallel connection of features of different scales and carries out reticulated information interaction in the fusion layer to realize the feature fusion of different branches. By aggregating inputs from other branches, the output of the model can simultaneously retain multi-resolution features. In order to improve the extraction performance of effective features in complex railway scenes, a bidirectional channel attention module is proposed. After the up-down sampling operation of the backbone network, the forward channel attention module makes the output feature map weighted by the input features of different scales, so as to adaptively improve the expression of effective features. The inverted channel attention module is inserted before the final output of the model, to generate effective high-level semantic information while retaining the underlying spatial information. Experimental results on the RailSem19 railway dataset show that the proposed method can significantly improve the segmentation performance of easily confused categories as well as railway facility categories such as track area, catenary mast, train and protective fence, and mIoU reached 65.12%, which is a certain improvement compared with other methods. References | Related Articles | Metrics

Select

Dangerous Driving Behavior Spectrum in Merging Area Based on Maneuver Indicators LIU Tang-zhi, BI Hui-yun, YANG Zhuo-si, XIA Hao-jie, SHANG Ting 2023, 23(2): 242-251.  DOI: 10.16097/j.cnki.1009-6744.2023.02.026 Abstract ( )   PDF (2143KB) ( )   The traffic environment in the highway interchange merge area is complicated, with a high interweaving ratio and high frequency of lane changing, and high risk coefficient. To quantitatively describe the main driving behavior in the merging area and their degree, this paper proposes a maneuver indicator to characterize dangerous driving behaviors in interchange merging areas and develops a driving behavior spectrum based on the characteristic parameters of vehicle trajectory data. Four kinds of dangerous driving behaviors are quantified by using the risk measure method, including acute direction, stomp pedal, dangerous following, and dangerous lane changing. Quartile deviation and criteria importance though intercrieria correlation (CRITIC) is used to determine the threshold values of dangerous driving behavior characteristics and weights, and calculate the characteristic values of drivers' dangerous driving behavior spectrum. Combined with the spatial distribution of dangerous points and related research, the effectiveness of the maneuver indicator is verified. The results of the study showed that the correlation coefficients of the four risky driving behaviors were all less than 0.2 and there was no collinearity. Among the four dangerous driving behaviors, the highest weighting of 0.405 is given to acute direction, which has the greatest impact on driving safety in merging area. The spatial distribution of maneuvering volume index hazard points is consistent with existing studies of traffic conflicts in merging areas. The above results prove that the dangerous driving behavior spectrum established using the maneuver indicators can identify the dangerous drivers in the merging area in real time combined with the vehicle trajectory data, and provide a reference basis for active safety prevention and control in the highway interchange merge area. References | Related Articles | Metrics

Select

Multi-objective Optimization of Variable Speed Limit Control Strategy on Expressway TANG Jin-jun, FU Qiang, WANG Cheng-cheng, YUAN Chen, ZHANG Xuan, XU Chang-jing 2023, 23(2): 252-261.  DOI: 10.16097/j.cnki.1009-6744.2023.02.027 Abstract ( )   PDF (2058KB) ( )   Variable speed limit control is an important part of active management and control technology of expressways. It dynamically updates the speed limit value according to the road operating conditions and effectively controls the traffic flow state. Due to the impact of traffic incidents, it is still a challenge to design a variable speed limit control strategy considering the efficiency and safety of traffic operations simultaneously. With the spatial characteristics of highway traffic flow under the influence of traffic incidents, we improve the basic Cell Transmission Model (CTM) model by extending the variable length of cells and integrating metastable state theory. Considering the decrease in traffic capacity in bottleneck areas under the influence of incidents and the change of speed under traffic control, a variable CTM model is constructed. To balance traffic efficiency and safety in practical applications, a multiobjective optimization method of the variable speed limit control strategy with the objective of shortest total travel time and minimum total collision risk is proposed, and the NSGA-II algorithm is applied to solve the optimization problem. The Ganggou-Xinglong section of the Jinan Ring Expressway was selected as the case study. The simulation scenario was constructed and adjusted based on the actual traffic flow data. A comparative experiment under different conditions, with different cell numbers and speed limit control cycles, was designed. The results show that the proposed method can effectively reduce the total travel time by more than 32% and the total collision probability by about 70% under the influence of the incident, which can effectively improve the efficiency and safety level of road traffic. At the same time, the advantages and disadvantages of the number of cells and the length of the control cycle of the speed limit were compared and discussed. References | Related Articles | Metrics

Select

Highway Network Post-disaster Recovery Decision Optimization Considering Resilience and Equity ZHAO Xue-ting, JIA Peng, YANG Yan-bo, KUANG Hai-bo 2023, 23(2): 262-274.  DOI: 10.16097/j.cnki.1009-6744.2023.02.028 Abstract ( )   PDF (2816KB) ( )   This paper focuses on the decision-making of highway network repair after disasters. Considering large-scale natural disasters normally affect a wide area, this paper divides the large disaster area into several small regions and then optimizes the repair time sequence of damaged road network for the sub-areas. The emergency rescue resources between multiple regions were allocated from the global perspective. Based on the topology structure of the transportation network and the reliability of the road network, the resilience value indicator is proposed to measure the overall performance of the network. Considering the resilience of the internal road network in the sub- region and the equity of the sub-region during the rescue process, this study develops an optimization model for the post-disaster highway network repair. The feasibility of the model and algorithm was verified by numerical examples. The results show that: in 8 rescue scenarios, the distribution scheme considering both resilience and equity can ensure the fairness of the rescue in the sub-regions with lower resilience value when compared with the schemes that only consider the total resilience value of the transportation network. Compared with the random repair scheme, the recovery effect of the optimized repair scheme can be improved by 57.92% on the overall network performance. To restore the overall performance of the network quickly, it is necessary to prioritize the repair of road sections with relatively high network density, particularly those located in the center of the network. Improving the speed of maintenance and repair is more beneficial than increasing rescue resources. The results also indicate that the road network restoration by regions addresses the equity of all sub-areas and improves the recovery efficiency of the road network. References | Related Articles | Metrics

Select

Multi-objective Optimization of Cold Chain Logistics Distribution Path Considering Time Tolerance 2023, 23(2): 275-284.  DOI: 10.16097/j.cnki.1009-6744.2023.02.029 Abstract ( )   PDF (2138KB) ( )   Aiming at the new characteristics of customers' high requirements on service time under new modes such as "central kitchen & cold chain distribution of food materials", this paper analyzes the cold chain logistics vehicle routing considering time tolerance. According to the time sensitivity of customers, the concept of time tolerance and its quantitative method were proposed, and a multi-objective optimization model of cold chain logistics distribution path was developed with the objectives of maximizing customers' time tolerance and minimizing the cost of cold chain logistics enterprises. The improved multi-objective plant growth simulated algorithm is designed to obtain the Pareto non-inferior solutions with the improvements of generating the initial base point by saving algorithm, adopting mixed random-fixed step size and elite strategy. With the distribution information of Pareto front, the optimal solution selection method was used to select the solution in Pareto's non inferior solution set that can be accepted by both customers and transportation enterprises. An example of a cold chain logistics company was analyzed, and the optimal distribution scheme was obtained to meet the requirements of customers and transportation enterprises. To verify the effectiveness of the model and algorithm, the number of 30, 50 and 100 customers in Solomon's standard calculation examples were selected for further analysis. By comparing with non-dominated sorting genetic algorithm-II algorithm, the superiority of the improved multi-objective simulated plant growth algorithm was verified. The proposed method can provide decision-making basis for cold chain logistics enterprises to make reasonable distribution. References | Related Articles | Metrics

Select

Reliability Logistics Network Design Based on Two Stage Robust Optimization SHI Chu-wei, MA Chang-xi, MA Cun-rui 2023, 23(2): 285-299.  DOI: 10.16097/j.cnki.1009-6744.2023.02.030 Abstract ( )   PDF (2809KB) ( )   This paper investigates the reliability logistics network design problem with uncertainty due to node and road damages. A reliability logistics network design method based on two-stage robust optimization is proposed. Taking the location of supply and transit nodes, determination of node connectivity, and distribution of cargo flow as decision variables, a two-stage reliability logistics network design model was constructed. The model has two independent objectives, i.e., the total network cost objective function and the total network operation time objective function. The former includes two stages of cost, in which the first stage calculates the network construction cost and the network operation cost under the normal state and the second stage calculates the network operation cost under a disruptive scenario set. The total network operation time objective function is used to calculate the network operation time under the normal state. A hybrid evolutionary algorithm with double-layer encoding structure chromosomes is designed. The NPGA is used as the main algorithm framework, and a large neighborhood search mechanism is designed to optimize the connectivity relationship genes of individuals. Cluster-based crossover and mutation strategies are combined to improve the search ability of the algorithm in the solution space. The effectiveness of the model and algorithm are verified by case studies with several groups of reliability logistics network design problems of different scales. The results show that the two-stage reliability logistics network design model can significantly reduce the operation cost of the network in case of damage and effectively improve network reliability through a small increase in the initial network construction cost. In the case comparison of 5 supply nodes, 10 transit nodes, and 15 demand nodes, the two sets of Pareto solutions for cost preference and time preference obtained by the model can save up to 20.6% and 28.2% of network operation cost in the same network damage scenario set, respectively, compared with the two sets of corresponding preference solutions of the traditional multi-objective logistics network model. The hybrid evolutionary algorithm converges to a better target value at the initial stage of iteration, showing a better search and optimization performance, which can effectively solve the two-stage reliability logistics network design model. References | Related Articles | Metrics

Select

Impact of Bike Sharing on Traffic Congestion in China's Major Cities HUANG Gan-xiang, ZHANG Wei, XU Di 2023, 23(2): 300-306.  DOI: 10.16097/j.cnki.1009-6744.2023.02.031 Abstract ( )   PDF (1354KB) ( )   To systematically investigate the governance effect of bike sharing on urban traffic congestion, this paper takes the staggered entry of ofo and Mobike into China's major cities as a quasi-natural experiment. Based on the 2016 to 2018 quarterly congestion delay indicator panel data of 45 major cities in China provided by Amap, this paper uses a difference-difference model to identify the impact of the bike sharing services on traffic congestion. The results show that: bike sharing service has a significant mitigation effect on urban traffic congestion, the bike sharing service reduces the congestion delay indicator by 2.9% on average, saving a total of about 309 million hours of commuting time and 15.1 billion yuan of economic losses annually, and reducing the total annual dioxide emission of urban vehicles in peak hours by about 3.55 million tons. The traffic congestion mitigation effect of bike sharing services is more effective in cities with heavier traffic volumes and larger populations, and is not restricted by air pollution. This study reveals the important role and potential value of bike sharing service in alleviating urban traffic congestion, and provides an innovative method and important policy enlightenment for urban traffic congestion management. References | Related Articles | Metrics

Select

Impact of Countdown Traffic Signal on Pedestrian Crossing Physiological Load PENG Jin-shuan, YANG Xiang-hao, CHEN Xin 2023, 23(2): 307-314.  DOI: 10.16097/j.cnki.1009-6744.2023.02.032 Abstract ( )   PDF (2610KB) ( )   To investigate the impact of signal countdown display on pedestrian physiological load, this paper conducted the experiments at two crosswalks with different signal countdown displays. The real-time parameters were collected such as physiology and speed to analyze the effect of the crossing initial time-phase and age on the physiological characteristics of pedestrians. The coupling pattern between physiological characteristics and crossing speed was revealed. Considering multi-source indicators such as heart rate, heart rate variability and skin conductance response, the paper proposes a pedestrian physiological load evaluation model. The results show that the heart rate growth rate (HRI), root mean square of R-R interval difference (RMSSD), mean value of skin conductance response (MEDA) gradually increase with the passage of the crossing initial time-phase. Nevertheless, the mean value of R-R interval (AVNN) gradually decreases. The physiological load of pedestrians in the whole process display is significantly higher than that in the part process display (10-second countdown). The physiological load of pedestrians all increases as crossing initial time-phase passes. The physiological load of elderly pedestrians is obviously higher than that of young and middle-aged pedestrians in each crossing initial time-phase. The results of the study provide a reference for traffic management departments to improve and optimize non-motorized traffic safety facilities. References | Related Articles | Metrics

Select

Fatigue Driving Detection Based on Spatial-temporal Electroencephalogram Features and Parallel Neural Networks ZHANG Bing-tao, CHANG Wen-wen, LI Xiu-lan 2023, 23(2): 315-325.  DOI: 10.16097/j.cnki.1009-6744.2023.02.033 Abstract ( )   PDF (2130KB) ( )   Since fatigue driving is one of the main inducements of traffic accidents, it is of great application value to explore objective and accurate method detection for fatigue driving. Considering the information complementary between different types of features, as well as the advantages complementary between different machine learning algorithms in the process of information mining, this paper proposes a fatigue-driving detection framework based on spatial-temporal electroencephalogram (EEG) features and parallel neural networks. To reduce the volume conductor effect, map the time series EEG data to the spatial brain functional network (BFN) based on the phase-locked value (PLV), and successively extract temporal domain EEG features and spatial metric features related to the driving process from the time-series EEG data and BFN. A feature contribution algorithm is designed by analyzing the relationship between features and target classes, to give different contribution coefficients for the temporal domain EEG features and the spatial domain BFN metric features. And the two types of weighted features are used as the inputs of the long short term memory (LSTM) network and the two-dimensional convolutional neural network (2D-CNN), to use the advantages of LSTM network in temporal data processing and CNN in the spatial data processing, and thus realizing the complementary information of spatial-temporal EEG features and the complementation of two types of neural network algorithms in data mining ability. A series of comparative experiments on public datasets show that the fatigue detection performance of the parallel neural network framework is superior to other methods, and the highest detection accuracy is 96.47%. This result means that this method can provide an effective solution for fatigue driving warning and assist safe driving. References | Related Articles | Metrics

Select

Scaled Tractive Power Distribution and Emission Model for Heavy-duty Trucks Based on Vehicle Weight HUANG Yi-ran, SONG Guo-hua, PENG Fei 2023, 23(2): 326-334.  DOI: 10.16097/j.cnki.1009-6744.2023.02.034 Abstract ( )   PDF (2687KB) ( )   PDF(English version) (752KB) ( 92 )   To quantify the relationship between vehicle weight of Heavy-Duty Trucks(HDT) and Scaled Tractive Power (STP) distribution, and thus improve the accuracy and efficiency of emission estimation, this study develops a model of STP distribution and emission for HDT based on vehicle weight. First, the actual STP distributions are developed based on trajectory data of HDTs with different vehicle weights in Beijing. Then, the STP distributions are fitted by the Gaussian functions, and the relationships between vehicle weight and the parameters of the Gaussian functions are quantified by the polynomial functions to develop the model. Finally, the NOx emission factors are calculated to verify the emission estimation and prediction accuracy of the model, and the impact of vehicle weight on the emission estimation for HDT is elaborated in comparison with the existing emission model MOVES. The results are as follows: (1) The emission estimation accuracy of the model is satisfied. The emission estimation errors of restricted and unrestricted access roads are 4.7% and 7.0% , respectively. (2) The emission can be predicted by the only variable vehicle weight, which reduces the cost of collecting data on the trajectory of HDTs with different vehicle weights and simplifies the traditional emission calculation process, and the emission prediction error of the HDT weighing 6.7 t is 5.3%. (3) Compared with MOVES based on default fixed driving cycles and fixed vehicle weights, the emission error drops by 16.7% according to the developed model. References | Related Articles | Metrics