Integrated passenger flow forecasting method for metropolitan rail transit based on spatial circle demand

Through the method based on the needs of the spatial circle, the urban circle is divided and modeled to predict the distribution of rail transit passenger flow and travel methods, and the problem of analysis of passenger flow characteristics and spatial differences in multi-level rail transit systems in the existing technology is solved. It has achieved scientific planning guidance and multi-level integrated passenger flow forecasting.

CN117455078BActive Publication Date: 2024-11-15CHINA RAILWAY SIYUAN SURVEY & DESIGN GRP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311759302.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-20
Publication Date
2024-11-15
Estimated Expiration
2043-12-20

AI Technical Summary

Technical Problem

The existing technology is difficult to effectively predict the passenger flow characteristics and spatial differences of multi-level rail transit systems, resulting in a lack of scientific guidance on rail transit planning, especially in the multi-level demand analysis of urban agglomerations.

Method used

It proposes an integrated passenger flow prediction method for rail transit based on the needs of the spatial circle. By dividing the urban circles into circles, building a travel model, predicting the passenger flow distribution and travel modes of each circle, combining gravity model and Logit model, passenger flow generation, Distribution and method division prediction, and integration planning and distribution are carried out to obtain passenger flow distribution indicators of rail transit networks at all levels.

Benefits of technology

It realizes accurate passenger flow forecast for multi-level rail transit systems, meets the travel needs of different circles, provides scientific planning guidance, fills the gap in existing technology, and is suitable for the planning practice of multi-level rail transit integration.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117455078B_ABST
    Figure CN117455078B_ABST
Patent Text Reader

Abstract

The present invention provides a method for predicting passenger flow of integrated metropolitan rail transit based on spatial circle demand, which relates to the field of transportation technology and comprises the following steps: dividing the metropolitan circle into high-speed rail circle, intercity circle, urban area circle and urban rail circle, and analyzing the structure and characteristic requirements of travel demand by circle from the inside to the outside; constructing a traffic travel model for each circle to predict passenger flow, including prediction of travel passenger flow generation, passenger flow distribution and travel mode division; predicting passenger flow of external channel traffic of the circle, including prediction of spatial distribution of passenger flow of external travel of rail transit; the present invention establishes a model to predict related demand according to the spatial scope of the prediction object, superimposes it with the passenger flow outside the circle and the passenger flow through the circle, and distributes it in the integrated rail transit network to obtain passenger flow distribution indicators of rail transit lines at all levels, so as to guide the planning practice of multi-level rail transit integration in the next step, and has important application value.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of transportation technology, and in particular to a method for predicting integrated passenger flow of metropolitan rail transit based on spatial circle demand. Background Art

[0002] Rail transit demand analysis cannot be simply equated with passenger flow forecasting. Different levels have different focuses, and passenger flow scale may not be the only factor determining the layout of the rail transit system. The regional rail transit demand analysis of the metropolitan area is different from the passenger flow forecast of the single-system subway network in the central city that researchers have long been accustomed to. The main focus is not just the passenger flow level, but more attention is paid to the spatial differences in passenger flow characteristics. The rail transit demand analysis of the city and a larger area should be based on different traffic strategy plan tests, aiming to grasp the consistency between the passenger flow law analysis and the regional urban system layout. The analysis of passenger flow composition characteristics is more important than the passenger flow scale forecast. The passenger flow forecast of rail transit cannot be equated with the demand forecast. It is necessary to pay attention to the differences in passenger flow composition characteristics of different circles, different sections, and different service groups.

[0003] At present, the research on integrated rail transit planning in domestic metropolitan areas is in full swing, but the current integrated rail transit planning is still in the exploratory stage. It is urgently needed to carry out relevant theoretical and technical research to fill the relevant gaps and guide the planning practice of integrated rail transit. The passenger flow prediction of integrated rail transit is one of the core contents of integrated rail transit planning. It is the basis and prerequisite for the evaluation of urban rail transit integration plans and policy formulation. Therefore, the prediction method for the multi-level integrated rail transit demand of metropolitan areas has high theoretical and practical research value. Therefore, the present invention proposes an integrated passenger flow prediction method for metropolitan rail transit based on spatial circle demand to solve the problems existing in the prior art. Summary of the invention

[0004] In response to the above problems, the present invention proposes an integrated passenger flow prediction method for urban rail transit based on spatial circle demand. The method establishes a model to predict relevant demand according to the spatial scope of the prediction object, superimposes it with the passenger flow outside the circle and the passenger flow through the circle, and distributes it in the integrated rail transit network to obtain the passenger flow distribution indicators of rail transit lines at all levels, so as to guide the next step of multi-level rail transit integration planning practice.

[0005] To achieve the purpose of the present invention, the present invention is implemented by the following technical scheme: a method for predicting integrated passenger flow of metropolitan rail transit based on spatial circle demand, comprising the following steps:

[0006] The metropolitan area is divided into high-speed rail area, intercity area, urban area and urban rail area, and the structure and characteristic requirements of travel demand are analyzed from the inside to the outside according to the area;

[0007] Construct a traffic travel model for each circle to predict passenger flow, including prediction of passenger flow generation, passenger flow distribution and travel mode classification;

[0008] Predict the passenger flow of external channels of the circle, including the spatial distribution prediction of passenger flow of external rail transit and the classification prediction of external travel modes of the circle;

[0009] Carry out integrated planning for the external channels of the circle, including traffic corridor identification, passenger flow section distribution analysis, propose preliminary plans for tracks at all levels, conduct passenger flow tests on the preliminary plans, and carry out superimposed integrated planning for channels based on the adaptability of the plans;

[0010] Passenger flow is distributed to each circle. Relying on the prediction steps of the traffic travel model corresponding to the four circles, the passenger flow and passenger flow direction distribution corresponding to various rail transit modes are obtained, and then the passenger flow demand distribution is implemented.

[0011] Further improvements are as follows: the urban rail circle is used to predict the total passenger flow and distribution of urban travel, including the distribution of steady-state passenger flow at hubs within the city; the urban area circle is used to predict the total passenger flow and distribution of urban fringe travel, including the distribution of steady-state passenger flow at some hubs; the intercity circle is used to predict the total passenger flow and distribution of travel to surrounding cities within a specific radius of the city, including directional passenger flow within a limited range; the high-speed rail circle is used to predict the total passenger flow and distribution of travel within a specific coverage area, namely, the total passenger flow and distribution of cities within the metropolitan area and economic belt to the outside world and between long-distance provinces and cities.

[0012] Further improvement is to generate predictions for travel passenger flow, including the following steps:

[0013] Passenger flow generation model construction: In the prediction of passenger flow generation in each circle, the data fitting regression tool is used: SPSS software is used to detect the correlation between independent variables, so as to screen the independent variables, that is, to exclude one of the two related independent variables to ensure the validity of all independent variables involved in the regression analysis, and to introduce the accessibility of the traffic area as an influencing factor of travel generation and attraction:

[0014] Accessibility index benchmark = average travel distance (km) / average travel time (h),

[0015] Based on the above improvements, an improved passenger flow generation prediction model is obtained:

[0016]

[0017] in, : Traffic area the amount of traffic generated or attracted; , , : coefficient to be determined; : An indicator of zoning activity, that is, an independent variable with a high correlation coefficient with the dependent variable; : Accessibility index of the traffic community;

[0018] Prediction of periodic change trend of passenger flow generation: Fourier series is used to approximate the periodic change trend of passenger flow generation over days, weeks and months as the time series is included. The expression is:

[0019]

[0020] Where, p: time series length; : Fourier series;

[0021] The formula is:

[0022]

[0023] In the formula,

[0024]

[0025]

[0026] Where, X(t): passenger flow change period; : parameter vector, initialized to ~Normal (0, ), is the period influence coefficient;

[0027] Compensation for passenger flow generation during special periods: During holidays and special weather windows, passenger flow generation shows a trend different from usual. In order to reduce the prediction error, compensation is made for these special periods. The formula is as follows:

[0028]

[0029] Where, L: special period set; i: single special period; : The impact factor of the corresponding event; : The time included in the window period.

[0030] Further improvement is: passenger flow distribution prediction, including the following steps:

[0031] The gravity model is applied to predict the travel time between the two areas OD. The travel time between the two areas OD varies greatly due to different travel modes. Therefore, the logarithmic and composite impedance functions are introduced into the interval impedance of the gravity model:

[0032]

[0033] In this formula, represents the interval impedance, Indicates the number of transportation modes. Indicates the mode of transportation travel impedance;

[0034] In order to make the prediction results of the gravity model consistent with the prediction results of occurrence and attraction in future years, the gravity model is combined with the growth rate method. That is, after calculating the traffic volume between two areas, in order to ensure that the traffic volume between the two areas is consistent with the generated traffic volume, the average growth rate method is used to verify the data.

[0035] Further improvement is: travel mode classification prediction, including the following steps:

[0036] Based on the summary table of passenger flow direction distribution of all modes in each circle in the target year, the travel modes are divided to obtain the passenger flow demand of each mode of rail transit. The passenger flow travel mode division is applied to predict the travel proportion of different travel modes in the planned total transportation travel mode. Through the nested Logit model, the transportation mode is divided into two layers. The first layer is the main mode layer of residents' travel, that is, it is divided according to the three modes of urban bus, car and rail transit. The second layer is the travel sub-mode layer, which further divides the rail transit mode into two sub-modes: urban rail transit and urban rail transit.

[0037] In the nested Logit model, the last layer is two types of rail transit travel modes, and the conditional probability calculation formula for their selection is shown below:

[0038]

[0039] in, : The conditional probability of choosing urban rail transit travel; : The conditional probability of choosing the urban rail travel; :The utility of urban rail transit mode; : The effectiveness of the urban rail mode; : stratification coefficient of rail transit mode;

[0040] Stratification coefficient Indicates the degree of mutual substitution between sub-modes in the same level of rail transit mode, and its value range is , 0 value means that there is no difference between the sub-modes, and they can be completely replaced; 1 value means that there is a competitive relationship between the sub-modes. When the stratification coefficient is 1, the nested Logit model becomes a binary Logit model. When the utility of urban rail transit and the utility of urban rail transit are both clear, the utility of rail transit is expressed as the sum of the logarithms of the above two. The calculation formula is:

[0041]

[0042] In the formula The utility of rail transit modes in various circles;

[0043] In addition, the probability of each mode of transportation being selected in the first layer is calculated as follows:

[0044]

[0045] In the formula Represents the transportation mode in the first layer of the model The probability of being selected by the traveler.

[0046] Further improvement is: the prediction of the spatial distribution of passenger flow of rail transit external travel includes the following steps:

[0047] Comprehensively consider the external traffic demand of the circle, take key cities and provinces as destination units, use the elasticity coefficient of economic growth and external travel growth to predict the scale of passenger flow outside the circle, and implement the prediction of the spatial distribution of external travel in combination with the city's gravity;

[0048] In the travel scale forecast, historical data is combined to predict the frequency of outbound travel in the future. In the forecast, the elasticity coefficient method is used to predict the future travel rate through statistical analysis of outbound travel rate and historical GDP data;

[0049] The magnitude of gravity between regions is used to determine the scale of inter-regional travel, and then the spatial distribution of passenger flow of rail transit external travel is clarified:

[0050]

[0051]

[0052]

[0053]

[0054] in, : Gross regional product; : Regional resident population; : The time to reach the destination by a certain mode of transportation; : Spatial gravity correlation coefficient, obtained by fitting the existing rail transit spatial distribution data with the gravity relationship.

[0055] Further improvement is: the prediction of external travel mode division by circle includes the following steps:

[0056] According to the development trend characteristics of each mode, the predicted results of out-of-circle passenger flow are divided into the order of large categories first and small categories later;

[0057] Classification of outbound land transportation modes: forecasts are made based on two major categories: air transportation and land transportation. The multivariate LOGIT model based on travel price and convenience factors is used to predict the sharing of the modes.

[0058] Division of road and rail transportation: Based on the above prediction results, the total amount of land transportation outbound travel is obtained, and the division model of price factors and convenience factors is used to further divide the land transportation volume;

[0059] Rail transit sub-mode classification: Based on the prediction of rail transit passenger flow sharing based on travel distance, ticket price, radiating population, and departure frequency factors, rail transit passenger flow is divided into two sub-modes: intercity railway and trunk railway.

[0060] Further improvements are: identification of transportation corridors, specifically: from a regional level, judging the development trend of the total social passenger volume in the planning scope in the past, as well as the main distribution characteristics, and analyzing the function of rail transit in the comprehensive transportation system;

[0061] The passenger flow section distribution analysis is as follows: the passenger flow demand section of each channel is an important basis for the line layout and construction sequence within the channel, and thus the analysis is carried out at different levels, internally and externally;

[0062] Propose preliminary plans for each level of track, specifically: Propose preliminary plans for each level of track based on the positioning, reasonable division of labor, time goals, and coverage target requirements of each level of track. For existing planning plans, propose optimization plans based on the principles of track circle division and passenger flow predictions from a comprehensive transportation perspective.

[0063] Further improvements are: passenger flow testing of the preliminary plan, specifically: according to the characteristics of rail transit construction projects in urban agglomerations and metropolitan areas, the entire passenger flow forecasting work is divided into three stages. First, basic data on social economy, population, and existing comprehensive transportation construction development are collected. Then, the highway and railway survey data of previous years and mobile phone signaling data are used to analyze the travel characteristics of passengers. Then, travel generation, travel distribution, mode division, and travel allocation prediction on the intercity railway network are carried out to obtain the passenger flow of the intercity railway network and the annual passenger flow of recent construction projects.

[0064] The channels are planned to be superimposed and integrated according to the adaptability of the scheme. Specifically, the possibility of integration between the integrated lines is considered from the adaptability of the integrated scheme capacity and the adaptability to the time target value required for inter-city functions. The channels are divided layer by layer. For those layers where the capacity, function and timeliness of the single lines can be superimposed on each other, a plan is planned to integrate them into one line, or a plan of integration within a certain period of time and gradual division of the lines is planned.

[0065] Further improvement is: passenger flow distribution to each circle includes the following steps:

[0066] Relying on the prediction steps of the traffic travel model corresponding to the four circles, the above demand is superimposed with the passenger flow outside the circle and the passenger flow through the circle to obtain the passenger flow and passenger flow direction distribution corresponding to the four rail transit modes of urban rail transit, urban rail, intercity rail and trunk rail under the integrated condition;

[0067] Implement passenger flow demand allocation, and allocate the integrated rail transit passenger flow demand obtained in the above steps to each line of each mode according to the corresponding rules in combination with the actual line network conditions;

[0068] In the allocation process, non-equilibrium allocation method, equilibrium allocation method, random allocation method and dynamic allocation method are used for specific implementation according to the requirements of engineering practice;

[0069] For travel outside the circle in a fixed direction, the passenger flow will be allocated to each line according to the corresponding proportion based on the passenger flow scale forecast based on the radiating population of a certain sub-mode of rail transit and the gravity of the city connected by the line.

[0070] The beneficial effects of the present invention are:

[0071] 1. The present invention analyzes the structure and characteristic requirements of travel demand from the inside to the outside according to the circle, combines the technical characteristics of rails at each level, divides the urban circle into circles, and selects appropriate parameters in different circles according to the characteristics of each circle to construct different models to accurately predict the rail transit integration demand under each circle. For the external travel passenger flow of the circle, the operability of large-scale data units, the feasibility of large-scale division and the guidance of regional overall positioning are considered. In summary, a model is established according to the spatial scope of the prediction object to predict related demands, which are superimposed with the passenger flow outside the circle and the passenger flow through the circle, and distributed in the rail transit integrated network to obtain the passenger flow distribution indicators of rail transit lines at each level to guide the planning practice of multi-level rail transit integration in the next step.

[0072] 2. The present invention is applicable to the field of preliminary planning in the transportation industry, and provides an effective method and scientific theoretical guidance for multi-level rail transit integrated passenger flow forecasting, which has important application value in assisting relevant planning decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 is a flow chart of the present invention;

[0074] Figure 2 A schematic diagram of the transportation mode of the present invention;

[0075] Figure 3 This is a schematic diagram of passenger flow testing of the preliminary scheme of the present invention. DETAILED DESCRIPTION

[0076] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with examples. The examples are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0077] Embodiment 1

[0078] according to Figure 1 As shown, this embodiment proposes a method for predicting integrated passenger flow of metropolitan rail transit based on spatial circle demand, including the following steps:

[0079] The metropolitan area is divided into high-speed rail area, intercity area, urban area and urban rail area, and the structure and characteristic requirements of travel demand are analyzed from the inside to the outside according to the area;

[0080] Construct a traffic travel model for each circle to predict passenger flow, including prediction of passenger flow generation, passenger flow distribution and travel mode classification;

[0081] Predict the passenger flow of external channels of the circle, including the spatial distribution prediction of passenger flow of external rail transit and the classification prediction of external travel modes of the circle;

[0082] Carry out integrated planning for the external channels of the circle, including traffic corridor identification, passenger flow section distribution analysis, propose preliminary plans for tracks at all levels, conduct passenger flow tests on the preliminary plans, and carry out superimposed integrated planning for channels based on the adaptability of the plans;

[0083] Passenger flow is distributed to each circle. Relying on the prediction steps of the traffic travel model corresponding to the four circles, the passenger flow and passenger flow direction distribution corresponding to various rail transit modes are obtained, and then the passenger flow demand distribution is implemented.

[0084] The present invention's research on the integrated passenger flow prediction of metropolitan rail transit spans multiple spatial circles and multiple rail transit system formats, filling the technical gap in the planning and research of multi-level rail transit integration and providing theoretical support.

[0085] Traditional rail transit passenger flow prediction theories and methods are mostly applied to a single rail transit system. They are not targeted at multi-level rail transit systems in urban areas in terms of data survey and passenger flow division, and are not universal for the overall spatial area involved in rail transit lines. Therefore, based on the different circles inside and outside the urban circle, the present invention divides the integrated passenger flow prediction of rail transit into four corresponding levels, and proposes an integrated passenger flow prediction method and model system for urban circle rail transit based on space-circle-demand.

[0086] The present invention is based on the research on the integrated passenger flow prediction method of urban rail transit, relies on the four-stage method, adopts the idea of ​​system analysis and integration, takes the area involved in the rail transit line as the research area, studies the passenger flow prediction problem based on the characteristics of different circles, and establishes four main models according to the spatial scope of the prediction object to predict the relevant demand, which is superimposed with the passenger flow outside the circle and the passenger flow passing through the circle (path), and distributed in the integrated rail transit network to obtain the passenger flow distribution indicators of rail transit lines at all levels, so as to guide the planning practice of multi-level rail transit integration in the next step.

[0087] Embodiment 2

[0088] according to Figure 1 , 2 As shown in FIG. 3 , this embodiment proposes an integrated passenger flow prediction method for metropolitan rail transit based on spatial circle demand, including the following steps:

[0089] Divide the metropolitan area into different levels;

[0090] Divide the metropolitan area into layers, analyze the structure and characteristics of travel demand from the inside out, combine the technical characteristics of each layer of rail, and propose a method for dividing the rail layers with the goal of meeting the diverse travel needs at different levels;

[0091] Although urban agglomerations vary in size, most of them are consistent in terms of their circle structure levels. Combined with the analysis of the circle structure levels of typical urban agglomerations, urban agglomerations are divided into high-speed rail circles, intercity circles, urban circles, and urban rail circles based on travel radius.

[0092] Table 1 Requirements of rail systems for different traffic circles

[0093]

[0094] Different transportation circles have different requirements for travel time and space, comfort, fare sensitivity, and frequency, and therefore have different requirements for station locations.

[0095] Table 2 Requirements for rail supply in different traffic circles

[0096]

[0097] The concept of multi-level rail transit is closely related to the "urban circle", and different levels of rail transit serve different circles. According to the principle of "one network, multiple modes, wide coverage, and high concentration", a multi-mode public transportation consisting of trunk railways, intercity railways, urban express rails, and urban rail transit is constructed.

[0098] Table 3 Composition and characteristics of multi-level rail transit system

[0099]

[0100] Mainline railways mainly serve long-distance transportation between metropolitan areas and other cities in the country. They are composed of national-level high-speed railways and passenger and freight trunk railways, and are the main arteries for external exchanges in metropolitan areas. Intercity railways connect major node cities in urban agglomerations. Urban (suburban) railways and urban rail transit are used to support the development of metropolitan areas. Urban (suburban) railways serve the central urban area and peripheral clusters and towns. The central area rail system serves the central urban area, including the backbone large-capacity urban rail transit system, as well as local medium and low-capacity supplementary lines.

[0101] Based on this, the metropolitan area is divided into four levels: high-speed rail circle, intercity circle, urban area circle, and urban rail circle. The urban rail circle mainly predicts the total volume and distribution of passenger flow within the city, including the distribution of steady-state passenger flow at the hub within the city; the urban area circle traffic travel model mainly predicts the total volume and distribution of passenger flow within the edge of the city, including the distribution of steady-state passenger flow at some hubs; the intercity circle traffic travel model mainly predicts the total volume and distribution of passenger flow radiating from a specific radius of the city to surrounding cities, which is generally a directional passenger flow within a limited range; the high-speed rail circle traffic travel model mainly predicts the total volume and distribution of passenger flow from cities within a specific coverage area, that is, the metropolitan area, the economic belt, and the long-distance inter-provincial and inter-city travel.

[0102] Construct a traffic model for each circle to predict passenger flow;

[0103] Travel passenger flow generation prediction;

[0104] Passenger flow generation model: In the prediction of passenger flow generation in each circle, the correlation between independent variables is tested by using data fitting regression tools, such as SPSS software, so as to screen the independent variables, that is, to exclude one of the two significantly correlated independent variables to ensure the validity of all independent variables involved in the regression analysis. Introduce the accessibility of the traffic zone as an influencing factor of travel generation and attraction:

[0105] Accessibility index benchmark = average travel distance (Km) / average travel time (h).

[0106] Based on the above improvements, an improved passenger flow generation prediction model can be obtained:

[0107]

[0108] in, : Traffic area the amount of traffic generated or attracted; , , : coefficient to be determined; : An indicator of zoning activity, that is, an independent variable with a high correlation coefficient with the dependent variable; : Accessibility index of traffic community.

[0109] Periodic trend of passenger flow generation: At the same time, the Fourier series is used to approximate the periodic trend of passenger flow generation in the days, weeks, months, etc. contained in the time series. The expression is:

[0110]

[0111] Where, p: time series length; : Fourier series.

[0112] This formula can also be expressed as:

[0113]

[0114] In the formula,

[0115]

[0116]

[0117] Where, X(t): passenger flow change period; : parameter vector, initialized to ~Normal (0, ), is the period influence coefficient.

[0118] Compensation for passenger flow generation during special periods: In some individual time windows such as holidays and special weather conditions, passenger flow generation may be quite different from usual times. To reduce the prediction error in the above situations, compensation is required for these special periods. The formula is as follows:

[0119]

[0120] Where, L: special period set; i: single special period; :The impact factor of the corresponding event; : The time included in the window period.

[0121] Passenger flow distribution prediction;

[0122] The traditional gravity model is a common method in passenger flow distribution prediction, but the application of the gravity model has the following limitations: ① The travel time between OD cells will be different depending on the travel mode, while the gravity model uses the same time impedance; ② When the distance between OD cells tends to zero, the passenger flow tends to infinity, which is obviously inconsistent with the actual situation, that is, when the distance between OD cells is small, there is a risk of over-prediction data; ③ The predicted results may not be consistent with the value of traffic demand generated in the future. Therefore, the gravity model is applied and improved in the passenger flow distribution prediction, which specifically includes the following aspects.

[0123] Since the travel time between OD areas varies greatly due to different travel modes, the gravity model interval impedance introduces logarithmic and composite impedance functions.

[0124]

[0125] In this formula, represents the interval impedance, Indicates the number of transportation modes. Indicates the mode of transportation travel impedance.

[0126] In order to make the prediction results of the gravity model consistent with the prediction results of occurrence and attraction in future years, the gravity model is combined with the growth rate method. That is, after calculating the traffic volume between small areas, in order to ensure that the traffic volume between small areas is consistent with the generated traffic volume, the average growth rate method is used to verify the data.

[0127] Travel mode segmentation prediction;

[0128] Based on the summary table of passenger flow direction distribution in all modes in each circle in the target year, the travel modes are further divided to obtain the passenger flow demand of each mode of rail transit. The passenger travel mode division model is used to predict the travel proportion of different travel modes in the total planned transportation travel mode. Through the nested Logit model, the transportation modes are divided into two layers. The first layer is the main mode layer of residents' travel, that is, it is divided according to the three main modes of transportation: urban buses, cars (including taxis, online car-hailing) and rail transit. The second layer is the travel sub-mode layer, which further divides the rail transit mode into two sub-modes: urban rail transit and urban rail.

[0129] In the nested Logit model, the last layer is two types of rail transit travel modes, and the conditional probability calculation formula for their selection is shown below:

[0130]

[0131] in, : The conditional probability of choosing urban rail transit travel; : The conditional probability of choosing the urban rail travel; :The utility of urban rail transit mode; : The effectiveness of the urban rail mode; : Stratification coefficient of rail transit modes.

[0132] Stratification coefficient Indicates the degree of mutual substitution between sub-modes in the same level of rail transit mode, and its value range is . A value of 0 indicates that there is no difference between the sub-modes and they can be completely replaced; a value of 1 indicates that the sub-modes are in a competitive relationship. When the stratification coefficient is 1, the nested Logit model becomes a binary Logit model. When the utility of urban rail transit and the utility of urban rail transit are both clear, the utility of rail transit can be expressed as the sum of the logarithms of the above two, and the calculation formula is:

[0133]

[0134] In the formula The effectiveness of rail transit modes in various circles.

[0135] In addition, the probability of each mode of transportation being selected in the first layer is calculated as follows:

[0136]

[0137] In the formula Represents the transportation mode in the first layer of the model The probability of being selected by the traveler.

[0138] Traffic flow prediction of external channels in the circle;

[0139] For the outbound travel flow of the circle, since it involves a long spatial distance and covers a wide geographical span, the prediction accuracy is greatly affected by the division range of traffic zones and the parameters for calculating the utility of the mode, and the fluctuation of the prediction results is more obvious. The operability of the scaled data unit, the feasibility of the division of major modes, and the guidance of the overall regional positioning should be considered.

[0140] Passenger flow distribution prediction;

[0141] Considering the external traffic demand of the circle, taking key cities and provinces as destination units, the elasticity coefficient of economic growth and external travel growth is used to reasonably predict the scale of passenger flow outside the circle, avoiding the commonly used trend analysis method that ignores the impact of urban and regional development changes on travel scale. After that, the spatial distribution of external travel is predicted by considering the city's gravity.

[0142] In the travel scale forecast, historical data is combined to predict the frequency of outbound travel in the future. Due to the correlation between outbound travel rate and social and economic activities, the forecast is made through statistical analysis of outbound travel rate and historical GDP data, and the elasticity coefficient method is used to predict the future travel rate.

[0143] The magnitude of gravity between regions is used to determine the scale of travel between regions, and then the spatial distribution of passenger flow of rail transit external travel is clarified.

[0144]

[0145]

[0146]

[0147]

[0148] in, : Gross regional product; : Regional resident population; : The time to reach the destination by a certain mode of transportation; : Spatial gravity correlation coefficient, obtained by fitting the existing rail transit spatial distribution data with the gravity relationship.

[0149] Prediction of external travel modes by circle;

[0150] In the prediction of the classification of external travel modes by circle, the predicted results of the external passenger flow of the circle are divided into the order of large categories first and small categories later according to the development trend characteristics of each mode.

[0151] Classification of land transport modes for outbound travel

[0152] The forecast is made based on two categories: air transport and land transport. The multivariate LOGIT model based on travel price and convenience factors is used to predict the sharing of the implementation methods.

[0153] Road and rail transit division

[0154] Based on the above prediction results, the total amount of outbound land transport trips is obtained, and the land transport volume is further divided using the division model of price factors and convenience factors.

[0155] Rail transit sub-mode classification

[0156] Based on the prediction of rail transit passenger flow sharing taking into account travel distance, ticket prices, radiating population and departure frequency factors, rail transit passenger flow is divided into two sub-modes: intercity railway and trunk railway (including high-speed railway network).

[0157] Integration planning of external channels of circles;

[0158] Identification of transportation corridors;

[0159] From a regional perspective, the development trend of the total passenger volume in the planning scope in the past and its main distribution characteristics are analyzed, and the function of rail transit in the comprehensive transportation system is analyzed. Taking a certain metropolitan area as an example, external corridors are identified from a spatial perspective based on data, and the existing rail transit passenger flow operation data is analyzed in combination with different external corridors to determine the characteristics of channel passenger flow.

[0160] Analysis of passenger flow cross-sectional distribution;

[0161] The passenger flow demand section of each channel is an important basis for the layout of the lines and the construction sequence within the channel. It plays an extremely important role in the integration of the channel and needs to be analyzed internally and externally. Taking a certain city as an example, from the perspective of the passenger flow corridor composition and passenger flow demand section, the future passenger flow spatial distribution of the city is close to the network-type spatial structure, with the vertical axis strengthening and the horizontal axis showing, and the regional network center will be formed to the north and east of the city's central area.

[0162] Propose preliminary plans for each level of track;

[0163] Based on the positioning, reasonable division of labor, time goals and coverage target requirements of each level of track, preliminary track plans for each level are proposed. For existing planning plans, optimization plans are proposed based on the track circle division principles and passenger flow prediction results from a comprehensive transportation perspective.

[0164] Passenger flow test of preliminary plan;

[0165] In view of the characteristics of rail transit construction projects in urban agglomerations and metropolitan areas, the entire passenger flow forecasting work is divided into three stages. The first is the collection of basic information such as social economy, population, and existing comprehensive transportation construction and development. The second stage is to use historical highway and railway survey data and mobile phone signaling data to analyze passenger travel characteristics. The third stage is to predict trip generation, trip distribution, mode division, and trip allocation on the intercity railway network to obtain the intercity railway network passenger flow and the annual passenger flow of recent construction projects.

[0166] Channel superposition and fusion;

[0167] Considering the possibility of integration between comprehensive lines from the perspective of the adaptability of the integrated solution capacity and the adaptability to the time target value required for intercity functions, the solution of integrating into one line can be considered if the capacity, function and time efficiency of the single lines superimposed on each layer can be met. This can be a solution of integration within a certain period of time and gradual division of lines.

[0168] Distribution of passenger flow in each circle;

[0169] Relying on the prediction steps of the traffic travel model corresponding to the four circles, the above demand is superimposed with the passenger flow outside the circle and the passenger flow through the circle (path), and the passenger flow and passenger flow direction distribution corresponding to the four rail transit modes of urban rail transit, urban (suburban) railway, intercity railway and trunk railway (including high-speed rail network) under integrated conditions can be obtained. Then, the passenger flow demand distribution can be implemented, and the integrated rail transit passenger flow demand obtained in the previous steps is reasonably distributed to each line of each mode according to the corresponding rules in combination with the actual line network conditions. The traffic distribution model includes four categories: non-equilibrium distribution method, balanced distribution method, random distribution method and dynamic distribution method, which can be implemented according to the requirements of engineering practice. For fixed-direction out-of-circle travel, the passenger flow can be distributed to each line according to the corresponding proportion based on the passenger flow scale prediction of the radiation population of a certain sub-mode of rail transit and the gravity of the line connecting the city.

[0170] The present invention analyzes the structure and characteristic requirements of travel demand by circle from the inside to the outside, combines the technical characteristics of each level of rail, divides the urban circle into circles, and selects appropriate parameters in different circles according to the characteristics of each circle to construct different models to accurately predict the rail transit integration demand under each circle. For the external travel passenger flow of the circle, the operability of the large-scale data unit, the feasibility of the large-scale division and the guidance of the overall regional positioning are considered. In summary, a model is established according to the spatial scope of the prediction object to predict related demands, which are superimposed with the passenger flow outside the circle and the passenger flow through the circle, and distributed in the rail transit integrated network to obtain the passenger flow distribution indicators of each level of rail transit network, so as to guide the next step of multi-level rail transit integration planning practice. It is suitable for the early planning field of the transportation industry, and provides an effective method and scientific theoretical guidance for the multi-level rail transit integrated passenger flow prediction, which has important application value for assisting relevant planning decisions.

[0171] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An integrated passenger flow prediction method for metropolitan rail transit based on spatial circle demand, characterized in that: The following steps are involved: The metropolitan area is divided into high-speed rail area, intercity area, urban area and urban rail area, and the structure and characteristic requirements of travel demand are analyzed from the inside to the outside according to the area; Construct a traffic travel model for each circle to predict passenger flow, including prediction of passenger flow generation, passenger flow distribution and travel mode classification; Predict the passenger flow of external channels of the circle, including the spatial distribution prediction of passenger flow of external rail transit and the classification prediction of external travel modes of the circle; Carry out integrated planning for external channels for each circle, including traffic corridor identification, passenger flow section distribution analysis, propose preliminary plans for tracks at each level, conduct passenger flow tests on the preliminary plans, and carry out superimposed integrated planning for channels based on the adaptability of the plans; Based on the integrated planned routes and considering the external passenger flow of the circles, passenger flow is allocated to each circle. Relying on the prediction steps of the traffic travel model corresponding to the four circles, the passenger flow and passenger flow direction distribution corresponding to various rail transit modes are obtained, and then the passenger flow demand allocation is implemented; The urban rail circle is used to predict the total passenger flow and distribution of urban travel, including the distribution of steady-state passenger flow at hubs within the city; the metropolitan area circle is used to predict the total passenger flow and distribution of urban fringe travel, including the distribution of steady-state passenger flow at some hubs; the intercity circle is used to predict the total passenger flow and distribution of travel to surrounding cities within a specific radius of the city, including directional passenger flow within a limited range; the high-speed rail circle is used to predict the total passenger flow and distribution of travel to specific coverage areas, i.e., cities within metropolitan areas and economic belts and between provinces and cities for long distances; Generating a forecast for travel passenger flow includes the following steps: Passenger flow generation model construction: In the prediction of passenger flow generation in each circle, the data fitting regression tool is used: SPSS software is used to detect the correlation between independent variables, so as to screen the independent variables, that is, to exclude one of the two related independent variables to ensure the validity of all independent variables involved in the regression analysis, and to introduce the accessibility of the traffic area as an influencing factor of travel generation and attraction: Accessibility index benchmark = average travel distance (km) / average travel time (h), Passenger flow generation prediction model: in, : Traffic area the amount of traffic generated or attracted; , , : coefficient to be determined; : An indicator of zoning activity, that is, an independent variable with a high correlation coefficient with the dependent variable; : Accessibility index of the traffic community; Prediction of periodic change trend of passenger flow generation: Fourier series is used to approximate the periodic change trend of passenger flow generation over days, weeks and months as the time series is included. The expression is: Where, p: time series period length; : Fourier series; The formula is: In the formula, Where, X(t): passenger flow change period; : parameter vector, initialized to ~Normal (0, ), is the period influence coefficient; Compensation for passenger flow generation during special periods: During holidays and special weather windows, passenger flow generation shows a trend different from usual. In order to reduce the prediction error, compensation is made for these special periods. The formula is as follows: Where, L: special period set; i: single special period; :The impact factor of the corresponding event; : The time included in the window period; Identification of transportation corridors: Specifically, from a regional perspective, we will study the development trend of the total passenger volume in the planning area in the past, as well as its main distribution characteristics, and analyze the functions of rail transit in the comprehensive transportation system; The passenger flow section distribution analysis is as follows: the passenger flow demand section of each channel is an important basis for the line layout and construction sequence within the channel, and thus the analysis is carried out at different levels, internally and externally; Propose preliminary plans for each level of track, specifically: Propose preliminary plans for each level of track according to the positioning, reasonable division of labor, time target, and coverage target requirements of each level of track; for existing plans, propose optimization plans based on the principles of track circle division and passenger flow results predicted from a comprehensive transportation perspective; The preliminary plan was tested for passenger flow. Specifically, the entire passenger flow forecasting work was divided into three stages according to the characteristics of rail transit construction projects in urban agglomerations and metropolitan areas. First, basic data on social economy, population, and existing comprehensive transportation construction development were collected. Then, historical highway and railway survey data and mobile phone signaling data were used to analyze passenger travel characteristics. Then, travel generation, travel distribution, mode division, and travel allocation on the intercity railway network were predicted to obtain the intercity railway network passenger flow and the annual passenger flow of recent construction projects. According to the adaptability of the scheme, the channel superposition and fusion planning is carried out, specifically: based on the adaptability of the scheme capacity after fusion, the adaptability to the time target value required for inter-city functions, and the possibility of fusion between comprehensive lines, the channels are divided layer by layer. For those layers where the capacity, function, and time efficiency of superimposing single lines can be met, a plan is planned to merge them into one line, or a plan of fusion within a certain period and gradual division of lines; The passenger flow distribution for each circle includes the following steps: Relying on the prediction steps of the traffic travel model corresponding to the four circles, the structure and characteristic requirements of travel demand are analyzed from the inside to the outside by circle, and the passenger flow outside the circle and the passenger flow through the circle are superimposed to obtain the passenger flow and passenger flow direction distribution corresponding to the four rail transit modes of urban rail transit, urban rail, intercity rail and trunk rail under the integrated conditions; Implement passenger flow demand allocation, and allocate the obtained integrated rail transit passenger flow demand to each line of each mode according to the corresponding rules in combination with the actual line network conditions; In the allocation process, non-equilibrium allocation method, equilibrium allocation method, random allocation method and dynamic allocation method are used for specific implementation according to the requirements of engineering practice; For travel outside the circle in a fixed direction, the passenger flow will be allocated to each line according to the corresponding proportion based on the passenger flow scale forecast based on the radiating population of a certain sub-mode of rail transit and the gravity of the city connected by the line.

2. The integrated passenger flow prediction method for metropolitan rail transit based on spatial circle demand according to claim 1 is characterized by: Passenger flow distribution prediction includes the following steps: The gravity model is applied to predict the travel time between the two areas OD. The travel time between the two areas OD varies greatly due to different travel modes. Therefore, the logarithmic and composite impedance functions are introduced into the interval impedance of the gravity model: In this formula, represents the interval impedance, Indicates the number of transportation modes. Indicates the mode of transportation travel impedance; In order to make the prediction results of the gravity model consistent with the prediction results of occurrence and attraction in future years, the gravity model is combined with the growth rate method. That is, after calculating the traffic volume between two areas, in order to ensure that the traffic volume between the two areas is consistent with the generated traffic volume, the average growth rate method is used to verify the data.

3. The integrated passenger flow prediction method for metropolitan rail transit based on spatial circle demand according to claim 2 is characterized by: The travel mode classification prediction includes the following steps: Based on the summary table of passenger flow direction distribution of all modes in each circle in the target year, the travel modes are divided to obtain the passenger flow demand of each mode of rail transit. The passenger flow travel mode division is applied to predict the travel proportion of different travel modes in the planned total transportation travel mode. Through the nested Logit model, the transportation mode is divided into two layers. The first layer is the main mode layer of residents' travel, that is, it is divided according to the three modes of urban bus, car and rail transit. The second layer is the travel sub-mode layer, which further divides the rail transit mode into two sub-modes: urban rail transit and urban rail transit. In the nested Logit model, the last layer is two types of rail transit travel modes, and the conditional probability calculation formula for their selection is shown below: in, : The conditional probability of choosing urban rail transit travel; : The conditional probability of choosing the urban rail travel; :The utility of urban rail transit mode; : The effectiveness of the urban rail mode; : stratification coefficient of rail transit mode; Stratification coefficient Indicates the degree of mutual substitution between sub-modes in the same level of rail transit mode, and its value range is , 0 value means that there is no difference between the sub-modes, and they can be completely replaced; 1 value means that there is a competitive relationship between the sub-modes. When the stratification coefficient is 1, the nested Logit model becomes a binary Logit model. When the utility of urban rail transit and the utility of urban rail transit are both clear, the utility of rail transit is expressed as the sum of the logarithms of the above two. The calculation formula is: In the formula The utility of rail transit modes in various circles; In addition, the probability of each transportation mode being selected in the first layer is calculated as follows: In the formula Represents the transportation mode in the first layer of the model The probability of being selected by the traveler.

4. The integrated passenger flow prediction method for metropolitan rail transit based on spatial circle demand according to claim 3 is characterized by: The prediction of the spatial distribution of passenger flow of rail transit external travel includes the following steps: Comprehensively consider the external traffic demand of the circle, take key cities and provinces as destination units, use the elasticity coefficient of economic growth and external travel growth to predict the scale of passenger flow outside the circle, and implement the prediction of the spatial distribution of external travel in combination with the city's gravity; In the travel scale forecast, historical data is combined to predict the frequency of outbound travel in the future. In the forecast, the elasticity coefficient method is used to predict the future travel rate through statistical analysis of outbound travel rate and historical GDP data; The magnitude of gravity between regions is used to determine the scale of travel between regions, and then the spatial distribution of passenger flow of rail transit external travel is clarified.

5. The integrated passenger flow prediction method for metropolitan rail transit based on spatial circle demand according to claim 4 is characterized by: The prediction of external travel modes by circle includes the following steps: According to the development trend characteristics of each mode, the predicted results of out-of-circle passenger flow are divided into the order of large categories first and small categories later; Classification of outbound land transportation modes: forecasts are made based on two major categories: air transportation and land transportation. The multivariate LOGIT model based on travel price and convenience factors is used to predict the sharing of the modes. Division of road and rail transportation: Based on the above prediction results, the total amount of land transportation outbound travel is obtained, and the division model of price factors and convenience factors is used to further divide the land transportation volume; Rail transit sub-mode classification: Based on the prediction of rail transit passenger flow sharing based on travel distance, ticket price, radiating population, and departure frequency factors, rail transit passenger flow is divided into two sub-modes: intercity railway and trunk railway.