An arbitrary time period OD passenger flow estimation method for urban rail transit based on station passenger flow fitting

Abstract

The application discloses a kind of urban rail transit OD passenger flow estimation method of arbitrary time period based on station passenger flow fitting, comprising the following steps: basic data acquisition;Calculate the cumulative in-and-out station passenger flow of each station;Fitting cumulative in-and-out station passenger flow function;Based on OD inter-station time consumption passenger flow time period projection;Based on the correlation of time period projection and station OD passenger flow estimation;Based on the passenger flow accuracy evaluation and correction of Frator model;Output each station arbitrary time period OD passenger flow result.The application solves the current urban rail transit OD passenger flow estimation is subject to fixed time granularity, discretization processing, data highly dependent, etc. Limit, difficult to meet the differentiated station, arbitrary time granularity and high-fidelity OD passenger flow real demand, and in low data dependence, continuous fitting and space-time alignment Limited in technology, the problem of insufficient practicality of estimation result, can break through the fixed time limit, without a large number of historical data support, realize continuous passenger flow fitting and space-time accurate matching OD passenger flow estimation.

Description

Technical Field

[0001] This invention belongs to the field of urban rail transit safety technology, specifically a method for urban rail transit safety based on station passenger flow fitting. Method for estimating origin-destination (OD) passenger flow at any time of day. Background Technology

[0002] Urban rail transit is an important component of urban public transportation. In the planning, design, operation, and management of urban rail transit lines, origin-destination (OD) passenger flow demand is a key focus for decision-makers, providing crucial support for capacity resource allocation, operation plan development, train dispatching, and passenger flow control analysis. OD passenger flow exhibits both temporal and spatial characteristics, with these two aspects interacting and influencing each other.

[0003] In recent years, with the development of refined and customized rail transit operations, decision-makers have put forward higher requirements for OD passenger flow. The calculation of OD passenger flow has also evolved from the traditional "hourly and fixed time granularity" to "arbitrary time granularity that can be determined according to needs", in order to meet the diverse operation and management needs such as passenger flow splitting, temporary scheduling adjustment, sudden passenger flow source analysis and control at different times.

[0004] Currently, while there has been some progress in data collection and model building for estimating OD (Original Departure) passenger flow in urban rail transit, it still primarily relies on fixed time periods. This makes it difficult to meet the passenger flow characteristics of different stations at any given time granularity, and it also suffers from inherent flaws such as high data dependence and a disconnect between the estimated results and actual operations. These flaws are specifically manifested in the following aspects (e.g., CN117475636, CN 103984993 B): (1) For prediction technology of passenger flow in the station, passenger flow out of the station and cross-sectional passenger flow, the prediction results are mainly used for macro-level route planning, road network layout, station scale determination and train route design. It is macro-level and coarse in time granularity. It cannot reflect the structural characteristics and travel patterns of urban rail transit passenger flow, and it is difficult to meet the refined needs of operation-level users such as operation organization, train dispatching and passenger flow control. It is urgent to further refine it from the OD passenger flow level. (2) OD passenger flow prediction technology based on historical OD passenger flow mostly relies on passenger flow feature extraction methods such as neural networks and machine learning. However, it requires a large amount of historical passenger flow in and out of the station and the corresponding OD passenger flow. It also has high requirements for data length and time continuity. It has problems such as difficulty in obtaining basic data, high data processing cost, and long time consumption for passenger flow feature extraction. It is difficult to apply to new routes, application scenarios without historical passenger flow data, or insufficient data accumulation.

[0005] (3) Existing technologies all rely on historical passenger flow data for fixed time periods, and the output OD passenger flow can only be presented in terms of passenger flow at the corresponding granularity of time period, and it is assumed that the passenger flow within the same time period changes uniformly. However, actual passenger flow has characteristics such as nonlinearity, short-term abrupt changes, and station differences. A uniform fixed time granularity cannot reflect the differences in passenger flow fluctuations at each station, which can easily lead to distortion of OD passenger flow results and inaccurate time and space matching. How to calculate the OD passenger flow at any time granularity for each station according to the actual needs of users is still a blank in existing technologies.

[0006] (4) Most existing OD passenger flow estimation or prediction technologies are based on discrete, fixed-time historical passenger flow data. They do not perform continuous fitting processing on passenger flow, and cannot restore the real continuous change pattern of passenger flow over time. At the same time, they cannot segment passenger flow at any time, which makes it difficult to accurately map and align the inbound passenger flow with the corresponding outbound passenger flow in the spatiotemporal dimension. Ultimately, this causes a deviation in the spatiotemporal correspondence of OD passenger flow, which is inconsistent with the actual OD passenger flow.

[0007] In summary, current urban rail transit OD passenger flow estimation technologies are constrained by core limitations such as fixed time granularity, discretized processing, and high data dependence, making it difficult to meet the actual needs of station-differentiated, arbitrary time granularity, and high-fidelity OD passenger flow in operational scenarios. Furthermore, they struggle to achieve technological breakthroughs in areas such as low data dependence, continuous fitting, and spatiotemporal alignment, resulting in limited applicability and insufficient practicality of the estimation results. Therefore, there is an urgent need to propose an OD passenger flow estimation method that can overcome the limitations of fixed time periods, requires no large amount of historical data support, and achieve continuous passenger flow fitting and accurate spatiotemporal matching. Summary of the Invention

[0008] Based on the above, the purpose of this invention is to provide a method for estimating OD passenger flow in urban rail transit at any time period based on station passenger flow fitting.

[0009] To achieve its purpose, the present invention adopts the following technical solution: A method for estimating OD passenger flow in urban rail transit at any time period based on station passenger flow fitting is proposed. This method relies solely on hourly passenger flow at each station, the time distribution of OD stations, user-defined passenger flow time periods at each station, and the correlation between OD stations within that time period. The process involves cumulative passenger flow calculation → cumulative passenger flow function fitting → projection of passenger flow time periods at each station → OD passenger flow estimation within the passenger flow time period → accuracy correction, achieving OD passenger flow estimation for any station at any time period and with any granularity. The specific steps are as follows: Step 1: Acquiring basic data: (1) Obtain a complete list of all stations along the rail transit line and determine the operating hours; (2) Obtain the hourly passenger flow in and hourly passenger flow out of each station; (3) Obtain the time distribution between OD stations, that is, the travel time distribution between each OD station; (4) Users can divide the passenger flow time periods of each station into any number, any start and end, and any granularity according to the passenger flow characteristics or estimated needs of each station, and support uniform granularity across the entire line or differentiated granularity for each station. (5) Obtain the passenger flow correlation between each OD station during the passenger flow period to characterize the passenger flow attraction intensity between stations; Step 2: Calculate the cumulative inbound passenger flow and cumulative outbound passenger flow for each station: For any station, based on the known hourly inbound and outbound passenger flows, calculate the cumulative inbound and outbound passenger flows at the end of each hour: (1) Cumulative passenger flow = the sum of all passenger flows entering the station from the start of the operating period to the end of the current hour; (2) Cumulative outbound passenger flow = the sum of all outbound passenger flows from the start of the operating period to the end of the current hour; Step 3: Fit the cumulative inbound passenger flow function and the cumulative outbound passenger flow function: Based on the cumulative passenger flow sequence obtained in step 2, a function fitting strategy is used to construct the following for each station: (1) Cumulative passenger flow function; (2) Cumulative outbound passenger flow function; The above function is a continuous function with respect to time. It can take any time as input and output the cumulative passenger flow at the corresponding time, realizing the conversion of passenger flow from discrete hour level to continuous time level, so as to accurately reflect the real changes of passenger flow over time. Step 4: Projecting passenger flow time periods based on OD station inter-station time consumption: For any site and any time period of passenger flow According to passenger flow time period Start and end times, and station Arrive at the station The total travel time between stations is used to obtain passenger flow time periods. On the site Time period projection on: (1) Projection start time = Time period Start time + time spent between OD stations; (2) Projection end time = Time period End time + OD station time; By using time projection, the arrival time and departure time can be mapped and aligned in the time dimension. Step 5: Initial OD passenger flow estimation based on time period projection and site correlation: Based on time-period projection, and combining the cumulative inbound passenger flow function, cumulative outbound passenger flow function, and passenger flow correlation of OD stations within the passenger flow period, the OD passenger flow of each station within its passenger flow period is estimated: (1) Calculate the station's passenger flow using the cumulative passenger flow function. During the period Total passenger flow entering the station; (2) Calculate the station's cumulative outbound passenger flow function. During the period projection period Total outbound passenger flow within the station; (3) Combined with the site During the period Total passenger flow into the station, stations During the projection period Total outbound passenger flow and inbound time period The site and sites Based on the correlation of passenger flow, the Logit model is used to estimate the station's passenger flow correlation. Arrive at the station During the period Initial OD passenger flow within the area; Using this method, the initial OD passenger flow matrix for all stations along the entire line can be obtained during any time period; Step 6: Output the OD passenger flow results for each station at any time period.

[0010] As a further preferred embodiment of the technical solution of the present invention, in order to ensure the accuracy of the calculation, the step between step 5 and step 6 further includes: a passenger flow accuracy evaluation and iterative correction step based on the Frator model, that is: based on the calculated initial OD passenger flow, the total inbound passenger flow and total outbound passenger flow of each station during the operating period are calculated in reverse, and the passenger flow accuracy is evaluated and corrected using the Frator model based on the total inbound passenger flow and total outbound passenger flow in the basic data.

[0011] Furthermore, the passenger flow accuracy evaluation and iterative correction based on the Frator model specifically includes the following steps: (1) Calculate the ratio of the total passenger flow entering and exiting each station during the operating period calculated in step 5 to the baseline total passenger flow entering and exiting the station in step 1, and determine whether the indicator meets the accuracy requirements. If it does, the correction ends. (2) If the indicator does not meet the accuracy requirements, the Frator model is used to correct the passenger flow during the calculated operating period. Based on this, according to the calculated OD passenger flow of each station during its passenger flow period, the Logit model is used to allocate the OD passenger flow of each station during its operating period to each passenger flow period, so as to obtain the OD passenger flow of each station during its passenger flow period after the current correction, and to evaluate the accuracy and determine whether to make correction again. (3) Repeat the correction steps until the index value meets the requirements, and the correction ends.

[0012] By adopting the above technical solution, the present invention solves the following problem: (1) Existing OD prediction / estimation methods rely heavily on a large amount of historical OD passenger flow data, passenger flow survey data, machine learning training samples, etc. To achieve the goal of estimating OD passenger flow within a preset time period for each station by relying only on hourly passenger flow, travel time distribution between OD stations, and correlation between stations within the passenger flow period, the data threshold can be reduced and the versatility can be improved. (2) The existing technology only supports fixed time granularity. It enables users to customize the OD passenger flow at any start and end time granularity according to the station passenger flow characteristics, so as to meet the diverse needs of urban rail transit such as refined operation, temporary scheduling, and passenger flow control. (3) Existing technology has problems with discrete passenger flow statistics and weakening the real fluctuation of passenger flow. Based on the hourly passenger flow of each station, a function fitting strategy is used to fit the cumulative passenger flow function and the cumulative passenger flow function of each station respectively, so as to restore the real change law of passenger flow of each station over time to the greatest extent, and lay the foundation for OD passenger flow at any time granularity; (4) The problem of mismatch between the time dimension of inbound and outbound passenger flow. Based on the cumulative inbound and outbound passenger flow function of each station, a time period projection mechanism between stations is established according to the time consumption distribution between each OD station to realize the mapping alignment of inbound and outbound passenger flow in the time dimension, so as to improve the spatiotemporal matching accuracy of OD passenger flow; (5) Existing OD prediction methods are only for future prediction and cannot restore historical / current OD passenger flow problems. To realize the estimation of OD passenger flow at any time granularity of the past period, it can be directly used for customized train scheduling, passenger flow review analysis and cross-sectional passenger flow verification and other actual operation scenarios.

[0013] Its beneficial effects are as follows: (1) Supports OD passenger flow estimation for any passenger flow period / any time granularity: Breaks through the limitation of fixed time periods, supports users to customize any number, any start and end and any granularity of passenger flow periods for each station according to passenger flow characteristics. It can be set uniformly or differentiated by station, and is highly adaptable to diverse needs such as refined operation, temporary scheduling and passenger flow control.

[0014] (2) Extremely low data dependence and extremely high versatility: Only hourly passenger flow, time distribution between OD stations, and correlation between OD stations are needed to complete the calculation. No historical OD passenger flow or machine learning training is required. It can be applied to new lines, new stations, and scenarios with insufficient data.

[0015] (3) Continuous passenger flow fitting, with higher estimation accuracy: The cumulative inbound passenger flow function and the cumulative outbound passenger flow function are obtained by fitting the cumulative passenger flow, restoring the nonlinear change law of passenger flow, avoiding the distortion of passenger flow smoothing caused by fixed granularity, and making the passenger flow estimation more accurate for any time period.

[0016] (4) Time-sharing projection mechanism for more accurate time-space matching: Based on the time consumption distribution between OD stations, the time-sharing projection of passenger flow time periods at each station on other stations is calculated, so as to achieve accurate alignment between the entry time period and the exit time period, and solve the problem of time-space misalignment of OD passenger flow from the root.

[0017] (5) With accuracy correction, the results are reliable and usable: The Frator model is used for accuracy evaluation and iterative correction to ensure the stability and reliability of passenger flow estimation results. It can be directly used in actual operation scenarios such as train scheduling, capacity configuration, cross-section verification, and passenger flow analysis.

[0018] (6) Simple calculation and strong engineering: No complex model or large amount of computing power is required. The process is clear and easy to program. It can be quickly embedded into the existing rail transit passenger flow analysis system and has low promotion and application costs. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method for calculating the OD passenger flow of urban rail transit at any time period based on station passenger flow fitting, according to the present invention. Detailed Implementation

[0020] The present invention will now be described in detail with reference to the accompanying drawings.

[0021] Taking a certain urban rail transit line as an example, the technical solution of the present invention will be further described in detail: (1) Basic data The assembly point of all stations along the line is denoted as Its operating hours are (e.g., 6:00~24:00); Get Station In the Passenger flow within one hour Passenger flow at the station ; by For the boarding station, For the destination station, take the total travel time between its corresponding OD stations. ; Users can divide a station into any number, granularity, and start / end times based on the station's passenger flow characteristics or their own needs. Specifically, let's assume that... Station operating hours Divided into During peak passenger flow periods, and Passenger flow time periods The start and end times, for Station No. The time granularity of each passenger flow period is as follows, and there is a recursive relationship between each passenger flow period and the operating period.

[0022] (1) (2) (3) (4) It should be noted that the above division can support a unified division of time periods and its granularity, i.e. and .

[0023] Enter the correlation coefficients of each OD station during the corresponding passenger flow period. , used to characterize Station and The intensity of passenger flow attraction between stations.

[0024] (2) Calculate the cumulative passenger flow in and out of the station For any station The number of passengers entering the station each hour is accumulated sequentially to obtain the station's [statistic / statistic]. Cumulative passenger flow at the end of the hour ,Right now: (5) Similarly, the cumulative outbound passenger flow at the end of the hour for any station can be obtained, i.e. (6) Based on this, any station can be formed. Cumulative passenger flow sequence and cumulative outbound passenger flow sequence , Operating hours The number of hours included.

[0025] (3) Fitting the cumulative passenger flow function Based on the cumulative inbound passenger flow sequence and the cumulative outbound passenger flow sequence, a function fitting strategy (which can be implemented using Python library functions) is employed to fit the cumulative inbound passenger flow function. and cumulative outbound passenger flow function , .

[0026] (4) Passenger flow time period projection based on OD station time according to Station No. Start time of each passenger flow period End time as well as Station and Travel time between stations ,Will Station No. Projecting passenger flow periods to the destination station Station, get Start time of the station projection period and end time They are respectively: (7) (8) By projecting time periods, precise alignment of arrival and departure times in the time dimension can be achieved.

[0027] (5) Initial OD passenger flow estimation based on time period projection and station correlation Using the cumulative passenger flow function ,calculate Standing at the Passenger flow period Total number of passengers entering the station ,Right now: (9) Using the cumulative outbound passenger flow function ,calculate During the projection period Total number of passengers leaving the station ,Right now (10) according to and Using the Logit model, we obtain Standing at its first Initial OD passenger flow for each passenger flow period ; (11) (6) Passenger flow accuracy evaluation and iterative correction based on the Frator model according to Station No. Od-out passenger flow during a specific passenger flow period Calculate the station's operating hours Internal OD Total OD passenger flow ,Right now (12) In the formula, Indicates the number of corrections, when When the initial calculated value is used, .

[0028] Recalculate Total passenger flow during its operating hours ,Right now (13) Recalculate Total outbound passenger flow during its operating hours ,Right now (14) calculate Total passenger flow into the station Accuracy evaluation index of the original total passenger flow ,Right now (15) Similarly, calculation Total passenger flow exiting the station Accuracy evaluation index compared with the original total outbound passenger flow ,Right now (16) judge and If all values ​​are close to 1, the iterative correction ends; otherwise, update the OD passenger flow in the correction formula (11) according to formula (17). and reset Then, execute equations (12) to (16).

[0029] (17) In the formula, For OD after one correction Total OD passenger flow As shown in equation (18).

[0030] (18) (7) Result Output Output the OD passenger flow results and other relevant information for all stations along the entire line during any estimated time period.

[0031] Examples of applications of this invention are as follows: Step 1: Basic Data Taking a rail transit line with 6 stations as an example, the operating hours are selected as 6:00 AM to 12:00 PM on a certain day, corresponding to 0 to 240 minutes. In addition, the time distribution between OD stations, the hourly entry volume of each station, the hourly exit volume of each station, and the passenger flow time period division results of each station are shown in Tables 1 to 4, respectively, assuming that each station has the same passenger flow attraction intensity.

[0032] Table 1. Time distribution between OD sites (min) Table 2. Hourly Inbound Volume at Each Station Table 3. Hourly Outbound Volume of Each Station Table 4. Passenger Flow Time Period Division Results for Each Station Step 2: Calculate the cumulative inbound passenger flow and cumulative outbound passenger flow for each station. According to Table 2, the cumulative passenger flow at the end of each hour is calculated using Equation (5). The calculation results are shown in Table 5.

[0033] Table 5: Cumulative Passenger Flow at Each Station at the End of Each Hour Similarly, according to Table 3, the cumulative outbound passenger flow at each station at the end of each hour is calculated using Equation (6), and the calculation results are shown in Table 6.

[0034] Table 6. Cumulative outbound passenger flow at each station at the end of each hour Step 3: Fit the cumulative passenger flow function. Based on the data in Tables 5 and 6, a function fitting strategy was adopted to fit the cumulative inbound passenger flow function and the cumulative outbound passenger flow function for each station. Various function fitting methods were used; among them, the polyfit function from the numpy library in Python was used to obtain the cumulative inbound passenger flow function and the cumulative outbound passenger flow function based on polynomial fitting, as shown in Tables 7 and 8, respectively.

[0035] Table 7. Cumulative Passenger Flow Function Based on Polynomial Fitting Table 8. Cumulative Outbound Passenger Flow Function Based on Polynomial Fitting It should be noted that when high accuracy is required, this invention does not recommend using polynomial fitting. The main reason is that the fitted polynomial function is not strictly monotonically increasing within a given time domain, which slightly contradicts the increasing characteristics of cumulative passenger flow. Therefore, this invention uses Python to call the PchipInterpolator library function to implement a strategy of preserving monotonically piecewise cubic Hermitian interpolation, thereby fitting the cumulative passenger flow function and the cumulative passenger flow into the station. Because the expression of this function is relatively complex, it will not be shown in detail here.

[0036] Step 4: Projecting passenger flow time periods at each station to other stations. Based on the passenger flow time period division results of each station in Table 4 and the time consumption distribution between OD stations in Table 1, the time period projection of each station's passenger flow time period on other stations can be obtained by using formulas (7) and (8), as shown in Table 9.

[0037] Table 9. Time projection of passenger flow periods at each station on other stations. Step 5: OD passenger flow prediction for each station during different time periods Based on the cumulative passenger flow function and formula (9), the passenger flow of each station during its corresponding passenger flow period (i.e., the period shown in Table 4) is obtained, as shown in Table 10.

[0038] Table 10 Passenger Flow at Each Station During Different Time Periods Similarly, based on the time projection of passenger flow periods at each station onto other stations in Table 9, the outbound passenger flow within the corresponding projection time periods at other stations can be obtained by using the fitted cumulative outbound passenger flow function and formula (10). On this basis, the initial OD passenger flow within the time period of each station can be calculated using formula (11), as shown in Table 11.

[0039] Table 11 Initial OD Passenger Flow for Each Station During Passenger Flow Periods Step 6: Evaluation and correction of passenger flow accuracy based on the Frater model For the initial OD passenger flow of each station during the passenger flow period in Table 11, the passenger flow accuracy is evaluated using equations (15) and (16). In this case, the accuracy is set to 0.005, that is, the range in which the evaluation index meets the requirements is [0.995~1.005]. At this time, the initial calculation results in Table 11 do not meet the accuracy requirements, and the maximum deviation is , so it needs to be corrected. Accordingly, based on the correction strategy, after 4 iterations of the Frator model, the OD passenger flow of each station during the passenger flow period that meets the accuracy requirements is shown in Table 12.

[0040] Table 12 Final Origin Passenger Flow for Each Station During Different Time Periods Furthermore, without altering the core innovative logic, the same inventive objective can be achieved by making equivalent substitutions only in the specific implementation methods. The specific substitution schemes are as follows: (1) Alternative to the cumulative passenger flow fitting algorithm The original solution used polynomial fitting, but alternative solutions include: cubic spline interpolation fitting, linear interpolation fitting, Lagrange interpolation fitting, piecewise polynomial fitting, and Gaussian process fitting. All of these methods can achieve the conversion of discrete cumulative passenger flow into a continuous-time function.

[0041] (2) Alternative solutions for cumulative passenger flow calculation The original plan used the accumulation method from the start of operation to the current time. The alternative plan can be: accumulation from 0:00 on the same day, accumulation from the morning peak, accumulation in segments according to natural days, or accumulation according to the operation cycle.

[0042] (3) Alternative solutions for time projection direction and calculation method The original scheme projected the total travel time plus the arrival time onto the departure side. An alternative scheme could be to project the total travel time minus the departure time onto the arrival side; or to use equivalent expressions and calculation methods such as time offset, time alignment, and time mapping.

[0043] (4) Alternative solutions for OD passenger flow allocation The original scheme used the OD site correlation coefficient to allocate passenger flow. The alternative scheme can be replaced by: site attractiveness weight, distance decay weight, time cost weight, passenger flow proportion weight, and interval passenger flow ratio.

[0044] (5) Alternative solutions for accuracy evaluation and correction models The original scheme used the Frator model for evaluation and iterative correction. The alternative scheme can be replaced by: mean square error correction, proportional normalization correction, cross-sectional passenger flow constraint correction, entropy weight correction, and least squares correction.

[0045] (6) Alternative solutions for time granularity setting methods The original solution supports custom granularity at any station level. Alternative solutions include: custom granularity at any interval, custom granularity at any origin-destination (OD) pair, custom granularity at any time period, and custom granularity at any passenger flow level. All of these fall within the equivalent protection scope of any time granularity.

[0046] It should be noted that the above-mentioned equivalent replacement solutions, without requiring creative effort, all fall within the protection scope of this invention.

Claims

1. A method for estimating OD passenger flow in urban rail transit at any time period based on station passenger flow fitting, characterized in that, Includes the following steps: Step 1: Acquiring basic data: (1) Obtain a complete list of all stations along the rail transit line and determine the operating hours; (2) Obtain the hourly passenger flow in and hourly passenger flow out of each station; (3) Obtain the time distribution between OD stations, that is, the travel time distribution between each OD station; (4) Users can divide the passenger flow time periods of each station into any number, any start and end, and any granularity according to the passenger flow characteristics or estimated needs of each station, and support uniform granularity across the entire line or differentiated granularity for each station. (5) Obtain the passenger flow correlation between each OD station during the passenger flow period to characterize the passenger flow attraction intensity between stations; Step 2: Calculate the cumulative inbound passenger flow and cumulative outbound passenger flow for each station: For any station, based on the known hourly inbound and outbound passenger flows, calculate the cumulative inbound and outbound passenger flows at the end of each hour: (1) Cumulative passenger flow = the sum of all passenger flows entering the station from the start of the operating period to the end of the current hour; (2) Cumulative outbound passenger flow = the sum of all outbound passenger flows from the start of the operating period to the end of the current hour; Step 3: Fit the cumulative inbound passenger flow function and the cumulative outbound passenger flow function: Based on the cumulative passenger flow sequence obtained in step 2, a function fitting strategy is used to construct the following for each station: (1) Cumulative passenger flow function; (2) Cumulative outbound passenger flow function; The above function is a continuous function with respect to time. It can take any time as input and output the cumulative passenger flow at the corresponding time, realizing the conversion of passenger flow from discrete hour level to continuous time level, so as to accurately reflect the real changes of passenger flow over time. Step 4: Projecting passenger flow time periods based on OD station inter-station time consumption: For any station and any passenger flow period, based on passenger flow period k The start and end times, and the total travel time between stations, are used to obtain the time projection of passenger flow periods on the stations: (1) Projection start time = Time period k Start time + time spent between OD stations; (2) Projection end time = Time period k End time + OD station time; By using time projection, the arrival time and departure time can be mapped and aligned in the time dimension. Step 5: Initial OD passenger flow estimation based on time period projection and site correlation: Based on time-period projection, and combining the cumulative inbound passenger flow function, cumulative outbound passenger flow function, and passenger flow correlation of OD stations within the passenger flow period, the OD passenger flow of each station within its passenger flow period is estimated: (1) Calculate the total passenger flow of the station during the time period using the cumulative passenger flow function; (2) Calculate the total outbound passenger flow of the station during the projected time period using the cumulative outbound passenger flow function; (3) Combining the total inbound passenger flow of the station during the time period, the total outbound passenger flow of the station during the projected time period, and the passenger flow correlation between stations within the time period, the Logit model is used to estimate the initial OD passenger flow from station to station during the time period. Using this method, the initial OD passenger flow matrix for all stations along the entire line can be obtained during any time period; Step 6: Output the OD passenger flow results for each station at any time period.

2. The method for estimating OD passenger flow in urban rail transit at any time period based on station passenger flow fitting as described in claim 1, characterized in that, Between steps 5 and 6, there is also a passenger flow accuracy evaluation and iterative correction step based on the Frator model, that is: based on the calculated initial OD passenger flow, the total inbound passenger flow and outbound passenger flow of each station during the operating period are calculated in reverse, and the passenger flow accuracy is evaluated and corrected using the Frator model based on the total inbound passenger flow and total outbound passenger flow in the basic data.

3. The method for estimating OD passenger flow in urban rail transit at any time period based on station passenger flow fitting as described in claim 2, characterized in that, The passenger flow accuracy evaluation and iterative correction based on the Frator model specifically includes the following steps: (1) Calculate the ratio of the total passenger flow entering and exiting each station during the operating period calculated in step 5 to the baseline total passenger flow entering and exiting the station in step 1, and determine whether the indicator meets the accuracy requirements. If it does, the correction ends. (2) If the indicator does not meet the accuracy requirements, the Frator model is used to correct the passenger flow during the calculated operating period. Based on this, according to the calculated OD passenger flow of each station during its passenger flow period, the Logit model is used to allocate the OD passenger flow of each station during its operating period to each passenger flow period, so as to obtain the OD passenger flow of each station during its passenger flow period after the current correction, and to evaluate the accuracy and determine whether to make correction again. (3) Repeat the correction steps until the index value meets the requirements, and the correction ends.

Images