A method and system for estimating the dwell rate of a city rail transit station

Abstract

This invention discloses a method and system for estimating station dwell rate in urban rail transit, relating to the field of station dwell rate estimation technology. The method includes: acquiring the station to be tested and its dwell rate calculation direction; selecting passenger card-swiping data entering the station within a set time interval; calculating the dwell rate of non-transfer passengers when they remain on the station based on the passenger's theoretical maximum travel time and actual travel time; selecting all valid paths containing the station dwell rate calculation direction within the passenger's valid path based on the entry and exit card-swiping stations of each transfer passenger, and calculating the probability of the passenger choosing the valid path and remaining on the station or the probability of the passenger choosing the valid path but not remaining on the station, thus estimating the station dwell rate of transfer passengers; and estimating the dwell rate of passengers at the station to be tested based on the station dwell rates of non-transfer passengers and transfer passengers. This method improves the accuracy of station dwell rate estimation by calculating the dwell rates of non-transfer passengers and transfer passengers.

Description

Technical Field

[0001] This invention relates to the field of rail transit station dwell rate estimation technology, specifically to a method and system for estimating urban rail transit station dwell rate. Background Technology

[0002] Currently, rail transit systems operate using minimum train departure intervals, but this still falls short of fully meeting the enormous passenger demand. This supply-demand imbalance leads to a series of problems, such as overcrowded carriages and passengers being unable to board the first train after arriving at the platform. These issues severely impact urban transportation efficiency and passenger travel experience, and may even trigger operational safety accidents. Increasing rail transit infrastructure is one measure, but under current resource and environmental constraints, simply increasing supply to resolve the supply-demand imbalance seems unrealistic. Besides the huge financial and human resources required to build subway systems, peak-hour passenger flow is typically concentrated within specific time and spatial areas. Subway construction will lead to excessive density in the rail transit network, potentially further increasing the system's operational complexity and management difficulty. Therefore, under existing network conditions, from an operational optimization perspective, it is crucial to accurately predict passenger congestion at urban rail transit stations, considering transfer passenger flow, precisely calculate passenger congestion frequency and station dwell probability, achieve real-time monitoring and optimization of congestion phenomena, and formulate relevant measures to reduce the number of congested passengers and improve the passenger travel experience. This will not only improve the efficiency of public transportation systems but also promote the sustainable development of low-carbon transportation systems.

[0003] Current research on passenger dwell time is still in its early stages. Most scholars only mention it in describing passenger travel choices and passenger flow distribution. The few methods that analyze dwell time and estimate its probability only consider single-line, non-transfer passenger flow, assuming that cross-line transfer passengers and non-transfer passengers at the same station have the same dwell time probability distribution. Furthermore, they require the proportion of non-transfer direct passengers to reach the lower limit of the sampling principle, resulting in long study intervals and calculated dwell time probabilities that represent the average dwell time over that period, failing to pinpoint the situation for individual trains. With the continuous development of rail transit networks in recent years, the proportion of transfer passengers is increasing, making it difficult for existing methods to accurately calculate station dwell rates. This hinders operators from effectively planning train operation schemes and passenger flow control measures. Summary of the Invention

[0004] To address the shortcomings of existing technologies that only consider single-line passenger flow without transfers, making it difficult to accurately calculate station dwell rates and hindering operators from effectively planning train operation schemes and passenger flow control measures, this invention proposes a method and system for estimating station dwell rates in urban rail transit that considers transfer passenger flow. By combining a multi-path probability allocation model, a method for estimating station dwell rates within a small-granular time period that considers cross-line transfer passenger flow is designed. This enables the analysis of dwell status for each train operating within a small-granular time period, thereby solving the problems existing in the prior art.

[0005] A method for estimating the dwell rate at urban rail transit stations includes the following steps:

[0006] Obtain the entry and exit card swiping stations, entry and exit times, and on-train times of all passengers entering and exiting the urban rail transit station under test within a set time interval;

[0007] Based on the entry time, on-train time and exit time of each non-transfer passenger, the proportion of passengers who did not board the first train to complete their journey is determined out of the passengers who entered the station within the set time interval, and thus the retention rate of non-transfer passengers at the originating station is obtained.

[0008] Based on the entry and exit card swiping stations of each transfer passenger, the passenger's valid routes are filtered out. By selecting all valid routes that include the test station's route in a specific train direction, the probability of the passenger staying at the station if they choose a valid route is calculated, or the probability of the passenger traveling but not staying if they choose a valid route, is determined. Based on the probability of each transfer passenger staying if they choose a valid route that includes the test station's route in a specific train direction, and the probability of traveling but not staying if they choose a valid route that includes the test station's route in a specific train direction, the station dwell rate considering transfer passengers is estimated.

[0009] Based on the dwell rate of non-transfer passengers at the originating station and the dwell rate of transfer passengers at the station, the dwell rate of passengers at the station to be tested is estimated.

[0010] Furthermore, the process of determining the proportion of passengers who did not board the first train to complete their journey, based on the entry time, on-train time, and exit time of each non-transfer passenger, to obtain the non-transfer passenger congestion rate at the originating station, specifically includes the following steps:

[0011] Based on the AFC card swipe data of all passengers entering and exiting the urban rail transit station within a set time interval, the actual travel time of all passengers between the same origin and destination (OD) is obtained.

[0012] Sort all passengers’ actual travel times and select the minimum value as the minimum theoretical travel time for each passenger.

[0013] The passenger's walking time to and from the station is determined by the difference between the theoretical minimum travel time and the time spent in the vehicle;

[0014] The waiting time for passengers is determined by the difference between the arrival or departure times of trains arriving at or departing from the same station.

[0015] Based on the passenger's walking time to enter the station, walking time to exit the station, time on the train, and waiting time, the passenger's theoretical maximum travel time is determined;

[0016] Whether a passenger should remain on the train is determined by the ratio of their actual travel time to the theoretical maximum travel time. If a passenger remains on the train, the retention rate of non-transfer passengers at the originating station is calculated based on the proportion of passengers who did not board the first train to complete their journey to the number of passengers who entered the station within the set time interval.

[0017] Furthermore, the probability that a passenger will stay at a station by choosing the effective route is determined by the ratio of the probability of the passenger choosing the route to the number of stages in the passenger's complete journey.

[0018] Furthermore, the multi-path probability selection model (MNL) is used to calculate the probability of passengers choosing each path, specifically including the following steps:

[0019] The basic cost for each route chosen by the passenger is calculated based on the passenger's walking time from the entrance gate to the waiting platform, the passenger's walking time from the waiting platform to the exit gate, the passenger's travel time for choosing this route, and the total waiting time for this route.

[0020] Calculate the penalty cost for each route chosen by the passenger based on the transfer and congestion penalty time incurred when the passenger chooses this route;

[0021] Based on the basic cost and penalty cost of each route chosen by the passenger, the impedance function of each route is obtained;

[0022] Construct the utility function of each path based on its impedance function;

[0023] Based on the utility function of the route, the probability of passengers choosing each route is calculated using the MNL model.

[0024] Furthermore, the process of obtaining the passenger's exit time specifically includes the following steps:

[0025] For any station s, the travel direction d The exit times of passengers on board all follow the corresponding probability density function distribution. ;

[0026] The probability density function of passenger exit time is expressed using the conditional probability density function as follows:

[0027] ;

[0028] in, This is the station's theoretical minimum exit time. The passenger's exit time is the difference between the time the passenger swipes their card to exit and the train's arrival time. , It refers to the train departure interval; The conditional probability density function representing the passenger's exit time. Indicates passenger i The departure time is less than the station s The arrival train interval is greater than the station's s The probability of the theoretical minimum departure time;

[0029] ;

[0030] Then we get ;

[0031] The exit time probability of all passengers is obtained by multiplying the conditional probability functions of all passengers' exit times, and a log-likelihood function is established.

[0032] Based on the log-likelihood function, the maximum likelihood estimation method is used to estimate the passenger exit times for all passengers at all stations in all directions. .

[0033] Furthermore, the passenger's entry time includes the walking time to enter the station and the waiting time; the process of obtaining this information specifically includes the following steps:

[0034] Passengers arriving at any departure interval follow a uniform distribution. Then the probability density function of the passenger's waiting time is expressed as:

[0035] ;

[0036] in, , For passengers i The time of gathering, For passengers i Walking time to enter the station, For passengers i Waiting time for the bus;

[0037] For any station s their travel direction d The walking time of passengers entering the station all follow the corresponding probability density function distribution. For passengers with only one feasible travel option, their probability density function distribution is expressed as follows: In this case, and The joint probability of simultaneous occurrence:

[0038] ;

[0039] in The conditional probability density function represents the walking time for passengers to enter the station. This indicates that the passenger's walking time to the station is greater than or equal to the minimum passenger walking time, and less than or equal to the minimum passenger walking time. i The time for people to gather at the station; The probability that the sum of a passenger's walking time to enter the station and their waiting time is less than or equal to the gathering time is expressed as follows:

[0040] ;

[0041] ;

[0042] Then we get ;

[0043] The exit time probability of all passengers is obtained by multiplying the conditional probability functions of the entry time of all passengers, and a log-likelihood function is established.

[0044] Based on the log-likelihood function, the maximum likelihood estimation method is used to estimate the passenger arrival times of all passengers at all stations in all directions. .

[0045] Furthermore, after filtering out the valid routes for each transfer passenger based on their entry and exit card swiping stations, the system uses the transfer passenger's exit card swiping time and the train exit time threshold of the theoretical travel plan to determine the passenger's decision to stay on each valid route; wherein, the theoretical travel plan is the travel plan corresponding to the first train that the passenger can board after arriving at the waiting platform.

[0046] The present invention also includes a system for estimating the dwell rate of urban rail transit stations, comprising:

[0047] The data acquisition module is used to acquire the entry and exit card swiping stations, entry and exit times, and on-train time of all passengers entering and exiting the urban rail transit test station within a set time interval;

[0048] The non-transfer passenger retention rate calculation module is used to determine the proportion of passengers who did not board the first train to complete their journey to the number of passengers who entered the station within a set time interval, based on each non-transfer passenger's entry time, on-train time, and exit time, thereby obtaining the non-transfer passenger retention rate at the originating station.

[0049] The passenger delay rate calculation module is used to filter out the effective routes of each transfer passenger based on their entry and exit card swiping stations. By selecting all effective routes that include the test station in a specific train direction, and calculating the probability that the passenger will delay at the station if they choose an effective route or the probability that the passenger will travel but not delay if they choose an effective route, the probability that the passenger will choose an effective route that includes the test station in a specific train direction and delay is determined. Based on the probability that each transfer passenger will choose an effective route that includes the test station in a specific train direction and delay, and the probability that they will choose an effective route that includes the test station in a specific train direction and travel but not delay, the station delay rate considering transfer passengers is estimated.

[0050] The estimation module is used to estimate the passenger dwell rate of the station to be tested based on the dwell rate of non-transfer passengers at the originating station and the station dwell rate of transfer passengers.

[0051] This invention provides a method for estimating the dwell rate of urban rail transit stations, which has the following beneficial effects:

[0052] This invention combines card-swiping data of all passengers entering and exiting stations with train operation data to achieve granular time-period dwell rate analysis, enabling a more accurate reflection of real-time station dwell conditions. By considering the problem of calculating dwell rate at stations in a single direction of urban rail transit for cross-line transfer passenger flow, this invention calculates the dwell rates for both non-transfer and transfer passengers, analyzes the effective path selection behavior of transfer passengers, and calculates the probability of passengers choosing each path, significantly improving the accuracy of dwell rate estimation. By analyzing the time distribution of passengers entering, waiting, boarding, and transferring, this invention dynamically matches passenger flow with train operation, providing data support for optimizing train operation plans and passenger flow control measures. This method improves the accuracy of dwell rate estimation, helping operation managers to promptly identify congested stations and take targeted measures (such as adjusting departure intervals and guiding passenger flow) to reduce dwell time and improve service quality and safety. Attached Figure Description

[0053] Figure 1 This is a schematic diagram simulating the travel trajectory of direct passenger flow and cross-line transfer passenger flow in a certain station in a single direction in an embodiment of the present invention;

[0054] Figure 2 This is a schematic diagram of passenger travel in a single-track rail transit system according to an embodiment of the present invention;

[0055] Figure 3 This is a schematic diagram of passenger travel along a single-line given path in an embodiment of the present invention;

[0056] Figure 4This is a schematic diagram illustrating the traffic flow layout analysis of the passenger exit process in an embodiment of the present invention;

[0057] Figure 5 This is a schematic diagram of the evacuation time parameters in an embodiment of the present invention;

[0058] Figure 6 This is a schematic diagram of the spatiotemporal trajectory and itinerary division for passengers traveling without transfers in an embodiment of the present invention;

[0059] Figure 7 This is a schematic diagram of the spatiotemporal trajectory and itinerary division of passengers with transfers in an embodiment of the present invention;

[0060] Figure 8 This is a schematic diagram illustrating the analysis of passenger congestion at the originating station for non-transfer passengers in an embodiment of the present invention;

[0061] Figure 9 This is a flowchart of a station dwell rate estimation method that considers transfer passenger flow in an embodiment of the present invention. Detailed Implementation

[0062] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0063] With the development of information technology, data acquisition techniques and the availability of multi-source data are also constantly evolving. A large amount of data is generated during daily urban rail transit operations, such as passenger travel time data recorded by the Automatic Fare Collection (AFC) system, train timetable data, and passenger walking time data. Addressing the shortcomings of existing station dwell rate estimation methods in considering transfer passenger flow, this invention aims to integrate and utilize available multi-source data to conduct a detailed and accurate analysis of past passenger travel patterns, determine the spatiotemporal distribution of passenger flow and passenger dwell times at various stations during relevant time periods, and provide a preliminary foundation for train operation planning and train timetable design.

[0064] This invention addresses the problem of calculating station dwell rates in a single operating direction of urban rail transit, considering cross-line transfer passenger flow within a small time frame. Passenger dwelling behavior is categorized into two types: proactive delay and reactive delay. Proactive delay occurs when passengers choose to wait for a more comfortable travel environment, while reactive delay refers to passengers being unable to board due to train overcrowding. Furthermore, passenger flow entering the station within a given time period can be categorized into four types: direct passenger flow without transfers in either direction and cross-line transfer passenger flow in both directions. Figure 1 This is a schematic diagram simulating the travel trajectories of direct passenger flow and cross-line transfer passenger flow in a certain station in one direction.

[0065] Depend on Figure 1It can be seen that the travel itinerary of passengers with direct trains without transfers at the originating station is divided into four sub-processes: entering the station, waiting for the train, boarding the train, and exiting the station. The travel itinerary of passengers with cross-line transfers at the originating station includes not only the above four sub-processes, but also a transfer and the subsequent travel itinerary. For passengers with direct trains without transfers at the originating station, the phenomenon of staying on the train can only occur at the originating station, while for passengers with cross-line transfers at the originating station, the phenomenon of staying on the train can also occur at the transfer station. For direct passengers arriving at the originating station without transfers, the passenger travel plans can be deduced based on the entry and exit card swiping times recorded by the AFC system and train operation data. Assuming that theoretically all passengers entering the station within the current small-granularity time period can board Train 1_1, and their feasible travel plans also include Train 1_2 and Train 1_3, and considering the characteristics of morning rush hour passenger flow, this invention stipulates that passengers exit directly after reaching the final station. Therefore, the actual travel plan of the passenger can be determined by combining the exit card swiping time. If the deduction result shows that the passenger completes the journey by boarding Train 1_1, it means that the passenger did not linger at the originating station. Otherwise, the direct passengers arriving at the originating station without transfers have experienced lingering at the originating station and can only choose travel plans Train 1_2 and Train 1_3 to complete this journey, with a probability of... For cross-line transfer passengers at the originating station, the first step is to determine whether the effective path that meets the passenger's travel needs includes the route in the calculated direction. If it does, the passenger travel plan set for the effective path is deduced by combining the entry and exit card swiping times recorded by the AFC system with train operation data. The theoretical and actual travel times are then used to determine whether the passenger should stay on the train, and finally, the probability of the passenger staying on the train at the originating station is determined. Finally, the actual congestion situation at the station is calculated by integrating the congestion situation of passengers without transfers at the originating station and the congestion situation of passengers transferring across lines.

[0066] In summary, the problem of this invention can be described as follows: Given highly developed online network accessibility and a gradually increasing proportion of transfer passenger flow, how can we utilize existing multi-source data and passenger flow allocation models to propose a reasonable and effective method for calculating station dwell rates that considers cross-line transfer passenger flow, enabling analysis of station dwell rates within a small-granular time frame considering cross-line transfer passenger flow? To establish a generalized model, the following assumptions are further introduced:

[0067] Assumption 1: First, since the research object of this invention, the morning rush hour commuter passenger flow, has high requirements for timeliness, it is assumed that passengers will not linger after arriving at the destination station and will exit the station directly.

[0068] Assumption 2: The distribution of passenger walking time within the station follows a normal distribution. Existing studies have determined that passenger walking speed follows a normal distribution, and speed determines passenger walking time. Therefore, it can be assumed that passenger walking time also follows a normal distribution. This assumption is to facilitate the estimation of passenger walking time within the station.

[0069] Assumption 3: Assume that passengers arrive at the station at a uniform rate, that is, the waiting time of passengers follows a uniform distribution.

[0070] Relevant time threshold definition:

[0071] (1) Entry Time Threshold: For any operating train in urban rail transit, the earliest and latest card-swiping entry times that passengers can theoretically use to board the train are calculated. The set of these two time values ​​is called the entry time threshold. That is, it is the set of the current train departure time minus the average passenger walking time to enter the station and the previous train departure time minus the average passenger walking time to enter the station. .

[0072] (2) Arrival Time Threshold: For any passenger choosing urban rail transit as their mode of transportation, given their card-swiping time at the station, calculate the earliest and latest arrival times at the platform. The set of these two times is called the arrival time threshold. That is, the set of the earliest and latest arrival times at the platform, obtained by adding the passenger's card-swiping entry time to the minimum and maximum walking times. .

[0073] (3) Exit Time Threshold: For any operating train on urban rail transit, calculate the earliest and latest card-swiping exit times for passengers who theoretically alight from that train without lingering. The set of these two time values ​​is called the exit time threshold. That is, it is the set consisting of the train's arrival time plus the average walking time for passengers exiting the station, and the arrival time of the next train plus the average walking time for passengers exiting the station. .

[0074] (4) Departure Time Threshold: For any passenger choosing urban rail transit as their mode of transportation, given their card-swiping time at the exit, calculate the earliest and latest arrival times at the platform. The set of these two times is called the departure time threshold. That is, the set of the earliest and latest arrival times at the platform obtained by subtracting the maximum and minimum entry walking times from the passenger's card-swiping exit time. .

[0075] (5) Travel Time Threshold: For any feasible travel plan corresponding to a valid path between any origin and destination (OD), calculate the maximum and minimum travel times required for a passenger to complete the travel plan. The set of these two time values ​​is called the travel time threshold of the travel plan. That is, it is the set of values ​​formed by the difference between the maximum exit card swipe time and the minimum entry card swipe time, and the difference between the minimum exit card swipe time and the maximum entry card swipe time, corresponding to the passenger choosing this travel plan. .

[0076] Definition of relevant travel plans:

[0077] For a single-line urban rail transit system, after passengers swipe their cards (or scan codes) to enter the station, they can walk or take escalators to the waiting platform and board any train departing from the platform to complete their travel needs. In a single-line rail transit system, each train corresponds to one travel plan, and each passenger may have multiple feasible travel plans within their travel time threshold. To facilitate differentiation, this invention categorizes all feasible travel plans into three types: theoretical travel plans, unique travel plans, and other travel plans. Figure 2 The specific explanation is as follows.

[0078] (1) Theoretical travel plan: This refers to the travel plan corresponding to the first train that a passenger can board after arriving at the waiting platform; from Figure 2 (a) It can be seen that when the shaded area of ​​the triangle enclosed by the passenger's arrival time threshold and the area enclosed by the arrival time threshold of a certain train is greater than the area enclosed by the triangle enclosed by the arrival time threshold and the area enclosed by the arrival time threshold of other trains, the passenger can theoretically board the train to complete the journey, which is the passenger's theoretical travel plan; special cases such as Figure 2 As shown in (b), when the area of ​​the triangle enclosed by the passenger's arrival time threshold and the area enclosed by the arrival time threshold of a certain train are equal to the area enclosed by the triangle enclosed by the arrival time threshold and the arrival time threshold of another train, and the departure time of the current train is within the passenger's arrival time threshold, then the current train is the passenger's theoretical travel plan.

[0079] (2) Unique travel plan: Based on the previous assumptions, the unique travel plan is the travel plan corresponding to the exit time threshold of the passenger's card swiping exit time; by Figure 2 (a) It can be seen that when the area of ​​the triangle enclosed by the passenger's departure time threshold and the area enclosed by a certain train's departure time threshold overlaps with the area enclosed by the triangle enclosed by the departure time thresholds of other trains, then the current train is the passenger's only travel option, and the current passenger will definitely board this train to complete this journey; special cases such as Figure 2As shown in (b), when the area of ​​the triangle enclosed by the passenger's departure time threshold and the area enclosed by the departure time threshold of a certain train are equal to the area enclosed by the triangle enclosed by the passenger's departure time threshold and the departure time threshold of another train, and the arrival time of the current train is within the passenger's departure time threshold, then the current train is the passenger's only travel option.

[0080] (3) Other travel options: These are all feasible travel options other than the theoretical travel option and the unique travel option. They may be an empty set.

[0081] 1. Model building:

[0082] 1.1 Single-path, non-transfer passenger flow and train matching model:

[0083] A single-path, non-transfer passenger flow involves only four processes: entering the station, waiting for the train, being on the train, and exiting the station. The time spent on the train is a fixed value and does not change. Both entry and waiting times occur at the originating station, and the uncertainty of waiting times makes them difficult to accurately predict. Exit time, however, is relatively easy to predict; it can be determined by matching the passenger's exit card swipe time with the arrival time of the relevant train. This also allows for the determination of the passenger's travel plan. Therefore, for a single-path, non-transfer passenger flow, the passenger flow and train matching process can be determined based on the train's arrival time and the passenger's exit card swipe time. Further analysis of passenger dwell time, combined with the passenger's entry card swipe time, can then determine the remaining passenger stay. Figure 3 This is a diagram illustrating available trains for single-path passenger flow without transfers.

[0084] analyze Figure 3 It can be seen that if a passenger's exit time falls within the blue area, it means the passenger boarded... The journey was completed without any delays at the originating station; if the passenger's exit time falls within the yellow zone, it means the passenger boarded... The journey is complete, and one instance of passenger staying at the originating station has occurred. If the passenger's exit time falls within the purple zone, it indicates that the passenger boarded... The trip was completed, and there were two instances of passengers staying at the originating station.

[0085] 1.2 Urban rail transit passenger flow allocation model based on passenger route selection behavior:

[0086] Constructing a physical topology network for urban rail transit using graph theory and complex network theory, and selecting reasonable and effective passenger travel routes, is the prerequisite for studying passenger flow allocation problems under networked operation. This process will greatly simplify the spatiotemporal complexity of passenger flow allocation problems in networked rail transit.

[0087] 1.2.1 Valid Path Search Algorithm:

[0088] Currently, commonly used effective path search algorithms include Dijkstra's algorithm, Dail's algorithm, K-short circuit algorithm, depth-first search algorithm, and breadth-first search algorithm. While Dijkstra's algorithm can find the shortest path between origin-destination (OD) stations, it is clearly impractical for selecting effective paths in current networked rail transit systems. Dail's algorithm is not suitable for ring networks. The K-short circuit algorithm has an unpredictable K value. Depth-first and breadth-first algorithms are enumeration algorithms with high computational time complexity, but they guarantee traversal of all feasible paths, preventing omissions or incorrect selections. Therefore, this invention selects the depth-first search algorithm and incorporates impedance constraints and transfer number constraints to select effective paths for urban rail transit networks.

[0089] Impedance constraint: The generalized cost function of the path is constrained by setting absolute and relative thresholds. Let the absolute threshold be... The amplification factor relative to the threshold is The requirement is that the generalized cost must be less than the minimum between the absolute threshold and the relative threshold. (Setting...) It is 15%. Depending on the path, that is:

[0090] (1).

[0091] Transfer number constraint: Multiple transfers can significantly alter passenger travel comfort; therefore, the selected route must have fewer transfers than the maximum number of transfers. When the transfer threshold is set to 1 or 4 times, the feasibility of the MNL model parameter estimation results does not reach 95%, indicating that setting the transfer threshold to 1 or 4 times is unreasonable. The model with commuter passenger flow as input achieves the best accuracy when the transfer threshold is set to 3 times; the model with non-commuter passenger flow as input achieves the best accuracy when the transfer threshold is set to 2 times. This invention focuses on morning peak commuter passenger flow, therefore the transfer threshold is set to 3 times.

[0092] (2).

[0093] 1.2.2 Urban Rail Transit Passenger Flow Assignment Model Based on MNL:

[0094] Under the "one-ticket transfer" operation model, the path selection problem for passengers in the urban rail transit network can be abstracted as a path decision-making problem explained in behavioral science. To facilitate passenger path decision-making and differentiate between paths, the overall cost of passenger travel can be represented by analyzing the passenger travel process, quantifying the influencing factors of path decision-making, and establishing a generalized path travel cost function. However, the factors influencing passenger travel choices are numerous and complex, and stochastic factors must also be considered. Therefore, for a specific effective path between an OD pair, its utility function is shown in Equation 3.

[0095] (3)

[0096] in, For path k The utility value; for r，s Interval k The generalized travel cost of an effective route r、s For the origin and destination stations of an effective route; for urban rail transit, once the origin and destination are determined, the fare will also be a fixed value. Therefore, the impact of the fare on passengers' travel choices is not considered. Thus, the impedance function of each route only includes the basic cost and the penalty cost. It is a random item, and .

[0097] (1) Basic Fees

[0098] ① Walking time for entering and exiting the station: The walking time for passengers from the entrance gate to the waiting platform is called the entrance walking time, expressed as... The time it takes for a passenger to walk from the waiting platform to the exit gate is called the exit walking time, expressed as... It is indicated that the walking time for entering and exiting stations for each route in this invention is the average of the walking statistics time.

[0099] ② Travel time

[0100] (4)

[0101] In the formula: Select the travel time for the k-th route for the passenger; For the train in the section a The interval running time; For the train at the station s The stopping time. K is the set of paths, A is the set of segments, S KAS is the station collection point.

[0102] ③ Waiting time

[0103] (5)

[0104] In the formula: The total waiting time for path k; Let be the waiting time at station s. Assuming passenger arrivals follow a uniform distribution, its value can be expressed as half the departure frequency, 0. For train departure intervals, This is the correction factor.

[0105] (2) Penalty fees

[0106] ①Transfer fee

[0107] (6)

[0108] In the formula: This refers to transfer time; The penalty factor representing the transfer time is a parameter to be estimated, and its value is generally greater than 1. The penalty factor is the number of transfers. Transfer station Arrival and departure times for trains on different routes. Transfer station The symbols used to represent Line 1 and Line 2.

[0109] ② Penalty costs arising from overcrowding

[0110] (7)

[0111] In the formula: Penalties incurred due to overcrowding; This represents the congestion penalty factor, which is a parameter to be estimated. Select the first option for passengers k Travel time for each route.

[0112] Taking into account both basic costs and penalty fees, the first OD station inter-station k The generalized cost of a valid path can be expressed as:

[0113] (8)

[0114] Based on the above path impedance function, a utility function is constructed. The MNL model is selected to describe passenger route selection behavior. In order to avoid unreasonable allocation caused by the absolute difference of its utility function, the MNL model with relative utility improvement is used to describe the probability of passenger travel route selection between OD pairs.

[0115] (9)

[0116] In the formula: To select a path k The probability of; The passenger's familiarity with the network is the parameter to be estimated.

[0117] 1.3. Estimation model for the distribution of entry / exit times and transfer times:

[0118] Estimating the distribution of entry / exit times and transfer times is the core of identifying and calculating station dwell time. This invention proposes a method to estimate the distribution of passenger entry / exit times and transfer times using comprehensive data composed of passenger entry / exit time data estimated from AFC data and train operation data, and passenger entry / exit and transfer walking time data obtained from on-site surveys and statistical investigations. However, since neither the estimated data nor the survey data can cover all passengers, the obtained passenger travel time dataset is a subset of all passenger sets, representing a truncated sample from the passenger travel sample. This results in the estimated distribution not being completely identical to the actual situation, but it can ensure a high correlation with the actual situation, thereby improving the estimation accuracy of various time elements of passenger travel. At the same time, parametric distributions are used to describe the entry / exit and transfer time data to minimize the impact of incomplete sample information.

[0119] (1) Departure time estimation model:

[0120] Compared to the originating station and transfer station, the passenger's journey at the terminal station is relatively simple, only including the time to walk out of the station. Therefore, the passenger's exit time can be calculated by the difference between the passenger's card exit time and the train's arrival time. However, it is necessary to ensure that the passenger does not linger after arriving at the station. The morning rush hour commuter flow meets this characteristic, which can minimize the data error caused in the calculation process. The calculation formula is shown in Equation (10):

[0121] (10)

[0122] Assuming for any station s A certain travel direction d The exit times of passengers on board all follow the corresponding probability density function distribution. For each station, each passenger needs time to travel from the platform to the exit gate. Therefore, the exit time for all passengers should be greater than the station's theoretical minimum exit time. Passenger exit time includes walking time from the platform to the escalator, escalator time, and time from the escalator to the exit turnstile. Figure 4As shown, the passenger flow at the terminal station is determined by the train door stopping positions and the location of the exit gates. Passengers exiting from the train door closest to the escalator should have the shortest exit time, which corresponds to traffic flow line 2. The escalator riding time is fixed, while platform walking time and concourse walking time can be determined by walking distance and normal pedestrian walking speed. For morning peak passenger flow, the exit time is generally shorter than the train departure interval. Then the passenger's exit time probability density function can be represented by the conditional probability density function shown in equation (11).

[0123] (11).

[0124] In the formula, The conditional probability density function representing the passenger's exit time. Indicates passenger i The departure time is less than the station s The probability that the arrival train interval is greater than the theoretical minimum departure time of station s can be expressed as: To express.

[0125] (12)

[0126] Substituting equation 12 into equation 11 yields:

[0127] (13)

[0128] To estimate passengers' departure time First, we assume that the exit walking time of each passenger is independent of each other. Then, we multiply the conditional probability functions of the exit times of all passengers to obtain the exit time probability of all passengers in the observed sample. As shown in equation (14), the logarithmic likelihood function shown in equation (15) is then constructed using the logarithm. The maximum likelihood estimation method can be used to estimate the passenger exit time of all passengers at all stations in all directions. .

[0129] (14)

[0130] (15).

[0131] (2) Arrival time estimation model:

[0132] Unlike the final destination station, the travel time at the originating station also includes waiting time. Even passengers entering the station at the same time and boarding the same train may have different waiting times, influenced by their walking speed and the chosen traffic flow route. Therefore, for passengers with a single, non-transferring route and a single travel plan, the time obtained by subtracting the departure time of the single travel plan from the card-swiping entry time recorded by the AFC system is not the walking time to the station, but rather the sum of the walking time to the station and the waiting time—that is, the de-escalation time. Figure 5 As shown, its calculation formula is as shown in equation (16).

[0133] (16)

[0134] In the formula, For passengers i The time of gathering, For passengers i Walking time to enter the station, For passengers i The waiting time is assumed to be uniformly distributed for passengers arriving within any departure interval, i.e. Then the probability density function of the passenger's waiting time is shown in equation (17).

[0135] (17)

[0136] Assuming for any station s A certain travel direction d The walking time of passengers entering the station all follow the corresponding probability density function distribution. For passengers with only one feasible travel option, their probability density function distribution can be expressed as follows: In this case, and The joint probability of simultaneous occurrence can be used to derive the following equation.

[0137] (18)

[0138] In the formula The conditional probability density function represents the walking time for passengers to enter the station. This means that the passenger's walking time to the station must be greater than or equal to the minimum passenger walking time, and less than or equal to the passenger's walking time. i The time for people to gather at the station, using It can be expressed as equation (19); The probability that the sum of a passenger's walking time to the station and their waiting time is less than or equal to the gathering time can also be expressed as... It is represented as shown in equation (20).

[0139] (19)

[0140] (20)

[0141] By combining equations 17, 18, 19, and 20, we can obtain...

[0142] (twenty one)

[0143] To estimate passengers' walking time to the station First, we assume that the passenger's walking time to the station is independent for each passenger. Then, we multiply the conditional probability functions of the passenger's entry time by the conditional probability functions of all passengers to obtain the probability of the passenger's exit time in the observed sample. As shown in equation (22), the logarithm is then used to construct the log-likelihood function shown in equation (23). The maximum likelihood estimation method can be used to estimate the passenger arrival time of all passengers at all stations in all directions. .

[0144] (twenty two)

[0145] (twenty three).

[0146] (3) Transfer time estimation model

[0147] Generally, passenger transfer time is longer than the combined walking time to and from the station. Furthermore, due to differences in transfer methods at transfer stations, individual passenger heterogeneity significantly impacts walking time, making passenger transfer time difficult to estimate. Therefore, this invention employs on-site surveys and video surveillance to track passenger transfer processes and statistically analyzes the transfer times of a subset of passengers to estimate overall passenger transfer time. This assumes that for any given station... s The transfer walking time of passengers in a certain transfer direction d all follow the corresponding probability density function distribution. Considering that if passengers transfer during their journey, the transfer time is generally greater than the theoretical minimum transfer time, a conditional probability function for the transfer walking time can be established as shown in equation (24).

[0148] (twenty four)

[0149] In the formula The conditional probability density function represents the walking time for passengers to enter the station. The cumulative probability density function for passenger transfer walking time can be expressed as follows: To represent, as in equation (25).

[0150] (25)

[0151] By combining equations (24) and (25), we can obtain...

[0152] (26)

[0153] To estimate passengers' walking time to the station First, we assume that the transfer walking time for each passenger is independent of the others. Then, we multiply the conditional probability functions of the transfer walking times of all passengers to obtain the exit time probability for all passengers in the observed sample. As shown in equation (27), the logarithmic likelihood function shown in equation (28) is then constructed using the logarithm. The maximum likelihood estimation method can be used to estimate the passenger arrival time of all passengers at all stations in all directions. .

[0154] (27)

[0155] (28).

[0156] 1.4 Passenger Retention Phenomenon Identification Model

[0157] The phenomenon of passengers staying on the train mainly occurs at the originating station and transfer station. Based on the individual heterogeneity of passengers, the reasons for staying on the train can be divided into three categories: (1) The passenger flow entering the station is too large, and the supply and demand are not in balance, which makes it impossible for the train flow to match the passenger flow, that is, the supply is less than the demand, and passengers cannot board the train; (2) Overcrowding prevents passengers from boarding the train. Crowding is mainly divided into platform crowding and train crowding. Platform crowding may prevent passengers who enter the station later from reaching the train door and miss the opportunity to board the train. Train crowding may prevent even passengers at the front of the platform from boarding the train because there is no remaining capacity; (3) Passengers actively choose to wait for the next train to travel. These passengers generally have higher requirements for comfort, so they will choose to spend more time to find a relatively comfortable environment (with a seat or a high standing density per person). For the origin and destination points of a line, the only cause of passenger congestion is the volume of incoming passengers. For intermediate stations, the causes of congestion include not only incoming passengers but also passenger flow within the train and passenger flow at stations the train stopped at before arriving at the station. Congestion at transfer stations is related to incoming passengers, transfer passengers, and passenger flow at stations the train stopped at before arriving at the station. Congestion of passengers with no transfers can only occur at the originating station, and their travel process is relatively simple, which can be determined through train matching or the ratio of actual travel time to theoretical travel time. Congestion identification for passengers with transfers is more complex due to the greater uncertainty of transfer times, which may lead to omissions or errors in identifying passenger congestion by using the ratio of actual travel time to theoretical travel time; therefore, train matching is the only reliable method for determining congestion.

[0158] (1) Identification of passengers who do not transfer and remain on the train

[0159] The travel behavior of passengers without transfers is relatively simple, involving only four processes: entering the station, waiting for the train, being on the train, and exiting the station. Whether a passenger remains at the originating station can be determined by matching the previously defined unique travel plan with the theoretical travel plan. Alternatively, it can be determined by calculating the ratio of the passenger's theoretical maximum travel time to their actual travel time. Figure 6 Based on the spatiotemporal trajectory and itinerary division of passengers traveling without transfers, it can be seen that the theoretical travel time of passengers only includes three parts: gathering time, on-vehicle time, and evacuation time. The evacuation time is composed of two parts: entry time and waiting time. The occurrence of passengers staying on the train will inevitably lead to an increase in waiting time. Therefore, how to set the threshold for waiting time is a core issue in calculating the theoretical maximum travel time threshold for passengers.

[0160] a) Walking time to enter and exit the station

[0161] First, the actual travel time of all passengers between the same OD is obtained by data filtering. Then, the actual travel time of all passengers (with a large enough data sample) is sorted. Finally, the minimum value is taken as the minimum theoretical travel time of the passenger. For this passenger, it is assumed that he / she does not need to wait for the train after arriving at the platform and can directly board the current train to complete the journey. Therefore, his / her waiting time is zero. For passengers with the same OD, their time on the train is known and equal. Therefore, the passenger's walking time to and from the station can be determined by the difference between the theoretical minimum travel time and the time on the train, as shown in equation (29).

[0162] (29)

[0163] In the formula, This represents the sum of the walking time for passengers entering and exiting the station; This represents the theoretical minimum travel time for passengers between O and D stations. This represents the train travel time between OD pairs.

[0164] b) Waiting time

[0165] If no passengers stay on the train during their journey, their waiting time will not exceed the train departure interval at the current time. The extreme condition is that when a passenger arrives at the waiting platform, the previous train has just departed. At this time, the passenger's waiting time is the longest. This value can be determined by the difference between the arrival and departure times of the preceding and following trains at the same station, as shown in equation (30).

[0166] (30)

[0167] In the formula, This represents the theoretical maximum waiting time for passengers. For the station No. The arrival time of the train. For the station No. The arrival time of the train.

[0168] c) Theoretical maximum travel time

[0169] If no passengers remain on the train, the maximum theoretical travel time for passengers is [not specified]. It can be expressed by equation (31);

[0170] (31).

[0171] d) Actual travel time

[0172] By recording the times passengers swipe their cards to enter and exit the station through the AFC system, the actual travel time of passengers can be obtained. The difference between the two is shown in equation (32).

[0173] (32)

[0174] like If the value is positive, it indicates that the passenger stayed at the originating station; otherwise, the passenger did not stay at the originating station.

[0175] (2) Identification of passengers who remain on the train after transferring

[0176] Compared to passenger flows without transfers, passenger flows with transfers are more complex, making it difficult to calculate the theoretical maximum travel time. This is because the entry, exit, and transfer times all require precise estimation. Propagation errors in these three data points can lead to an overestimation or underestimation of the theoretical maximum travel time. Furthermore, the accuracy of entry, exit, and transfer time estimations using the trip estimation method is lower than that of the calculation method. Therefore, this invention uses the trip estimation method to determine whether transferring passengers should remain on the train. Figure 7 This refers to the spatial and temporal trajectories and itineraries of passengers with transfers.

[0177] The passenger's card swiping time for entering and exiting the station is known. At transfer stations, the passenger's complete journey is divided into two stages. Using the relevant time thresholds and travel plans defined earlier, the theoretical travel plan for the first stage and the unique travel plan for the second stage can be determined. For transfer stations, subway operators generally consider the train connections between different lines when compiling train timetables. Therefore, they consider that there is a matching train for the first and second stages of the journey, i.e., the first train that a passenger can board when transferring from one line platform to another. Figure 7 middle The matching train is ,Right now , The transfer time is designed to allow most passengers to complete the transfer process. Figure 7 Taking the passenger travel process shown as an example, when the passenger's card swipe time at the exit falls on the train... exit time threshold Inside, passengers have only one feasible itinerary. If the passenger's card swipe time falls on the train, then the passenger will not have any delays during this trip; exit time threshold Within this range, the passenger has two feasible itineraries, namely: and At this point, passengers will inevitably be stranded; similarly, the time when passengers swipe their cards to exit the station falls on the train. exit time threshold Inside, passengers will inevitably be stranded.

[0178] 1.5 Methods for estimating station dwell time

[0179] Existing studies have shown that non-transfer trips account for 30% of total subway trips. Therefore, this portion of passenger samples can be reasonably used to estimate station congestion based on sampling principles. However, for small-granularity time intervals, the sample size of non-transfer passengers is often insufficient to meet sampling requirements. Therefore, to address station congestion rates within small-granularity time intervals, it is also necessary to consider the congestion situation at the originating station for transferring passengers.

[0180] (1) A method for estimating the dwell rate of rail transit stations with direct passenger flow without transfers.

[0181] Considering the current station's no-transfer passenger flow congestion situation, which refers to the probability that a passenger who swipes their card to enter the station without transferring does not board the first train within a certain time range, the passenger's origin station is the current station, and the destination station could be any station on the line. Therefore, calculating the station congestion rate for no-transfer direct passenger flow requires considering the origin-destination (OD) of all passengers choosing to travel on this line. Figure 8 This is a diagram analyzing the congestion situation at the originating station for non-transfer passengers.

[0182] analyze Figure 8 It can be seen that on a single-track rail transit line with n stations, passengers from Departing from one station, it can arrive at any station on the line, when the passenger takes the first train after arriving at their destination. Upon arrival at the destination station, it indicates that the passenger has not been delayed; otherwise, it indicates that the passenger is at the station. This resulted in a "retention" phenomenon. This is represented in the diagram as... For all passengers entering the station within the purple area of ​​the train arrival time threshold, if their exit card swipe time falls within the green area of ​​the exit time threshold, it indicates that the passenger arrived at the originating station. There was no passenger remaining on the train; otherwise, if the exit card swipe time falls within the pink area, it means the passenger was at the originating station. This resulted in passengers remaining on the train. Therefore, non-transfer direct passengers were concentrated at the originating station. The dwell rate can be expressed as the proportion of passengers who enter the station within a given time interval t and exit in the pink area, or the proportion of passengers who complete their journey without boarding the first train out of the passengers who enter the station within a given time interval t. Its mathematical expression is defined as shown in equation (33).

[0183] (33)

[0184] (2) Calculation method for the dwell rate of rail transit stations considering cross-line transfer passenger flow.

[0185] With the construction mileage and passenger flow of rail transit networks increasing year by year, network operation has become an inevitable trend. The higher the accessibility of the network, the more transfer passengers will be. At the same time, it is becoming increasingly difficult to calculate the station dwell rate in the rail transit network. This mainly involves the identification and calculation of the dwell rate of transfer passengers. The dwell rate is an important indicator for evaluating the quality of rail transit services. Therefore, it is urgent to design a station dwell rate calculation method that takes into account the transfer passenger flow.

[0186] Given the known passenger travel delays, this invention assumes that the probability of a passenger being delayed at each stage of a complete journey is the same. Combining AFC card swipe data, it designs a clever method for estimating station delays in transferring passenger flow. The specific steps are as follows:

[0187] Step 0: Initialization, determining the computing site Determine the direction of calculation .

[0188] Step 1: Extract computing sites In time interval t Passengers transferring inside the station i The card swipe data.

[0189] Step 2: Passenger Screening i The effective paths are determined, and the probability of selecting each path is calculated based on the MNL multi-path probabilistic selection model.

[0190] Step 3: Extract a valid path k Determine the path k Does it include computing sites? In the direction of calculation If the path is on the set K, add it to set K and proceed to step 4; otherwise, the path is related to the computation station. The retention rate is irrelevant. Proceed to Step 7.

[0191] Step 4: Determine passengers based on the passenger flow retention phenomenon identification method. i Select path k Will there be any passengers left behind? If so, proceed to Step 5; otherwise, proceed to Step 6.

[0192] Step 5, Passengers i Select path k On the site Probability of being stranded Equal to the probability of choosing this path The ratio to the number of stages in the complete journey is then used to proceed to Step 7.

[0193] Step 6, Passengers i Select path k The probability of traveling but not staying is Proceed to Step 7.

[0194] Step 7: Is this the last valid path? If yes, proceed to Step 8; otherwise, proceed to Step 3.

[0195] Step 8: Confirm Passengers i Select Include Computing Sites In the direction of calculation Determine the path of the upper route and the probability of getting stuck. Proceed to Step 9.

[0196] Step 9: Determine the passenger i Is it a time interval? t If the last passenger is present, proceed to Step 10; otherwise, proceed to Step 2.

[0197] Step 10: Record the station selected by each transferring passenger. In the direction of calculation The probability of being stuck on the upper path. .

[0198] In summary, the station congestion calculation formula considering transfer passenger flow can be obtained as follows:

[0199] (34)

[0200] in, For passengers i Select path k The probability of traveling but not staying; Indicating the direction of travel; For the site; For passengers i Select path k On the site The probability of being stranded; For stranded passengers i Select path k The probability of; t For time intervals.

[0201] This invention enables the analysis of passenger retention for each train service within a small timeframe. Furthermore, the research findings provide more precise data support for the dynamic coupling of passenger and train flow, achieving refined allocation of passenger and train traffic. This provides a fundamental theoretical basis for rail transit operators to manage passenger flow and adjust service schedules.

[0202] Based on the same inventive concept, this invention also proposes an urban rail transit station occupancy rate estimation system, comprising:

[0203] The data acquisition module is used to acquire the entry and exit card swiping stations, entry and exit times, and on-train times of all passengers entering and exiting the urban rail transit test station within a set time interval.

[0204] The non-transfer passenger dwell rate calculation module is used to determine the proportion of passengers who did not board the first train to complete their journey to the number of passengers who entered the station within a set time interval, based on each non-transfer passenger's entry time, on-train time, and exit time, thereby obtaining the non-transfer passenger dwell rate at the originating station.

[0205] The passenger delay rate calculation module is used to filter out the effective routes of each transfer passenger based on their entry and exit card swiping stations. By selecting all effective routes that include the test station in a specific train direction, and calculating the probability that the passenger will delay at the station if they choose an effective route or the probability that the passenger will travel but not delay if they choose an effective route, the module determines the probability that the passenger will choose an effective route that includes the test station in a specific train direction and delay. Based on the probability that each transfer passenger will choose an effective route that includes the test station in a specific train direction and delay, and the probability that they will choose an effective route that includes the test station in a specific train direction and travel but not delay, the module estimates the station delay rate considering transfer passengers.

[0206] The estimation module is used to estimate the passenger dwell rate of the station to be tested based on the dwell rate of non-transfer passengers at the originating station and the station dwell rate of transfer passengers.

[0207] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for estimating dwell rate at an urban rail transit station, characterized in that, Includes the following steps: Obtain the entry and exit card swiping stations, entry and exit times, and on-train times of all passengers entering and exiting the urban rail transit station under test within a set time interval; Based on the entry time, on-train time and exit time of each non-transfer passenger, the proportion of passengers who did not board the first train to complete their journey is determined out of the passengers who entered the station within the set time interval, and thus the retention rate of non-transfer passengers at the originating station is obtained. Based on the entry and exit card swiping stations of each transfer passenger, the passenger's valid routes are filtered out. By selecting all valid routes that include the test station's path in a specific train direction, the probability of the passenger staying at the station due to choosing that valid route is calculated, as well as the probability of the passenger traveling but not staying due to choosing that valid route, is determined. Based on the probability of each transfer passenger staying due to choosing a valid route that includes the test station's path in a specific train direction, and the probability of traveling but not staying due to choosing a valid route that includes the test station's path in a specific train direction, the station dwell rate considering transfer passengers is estimated. The probability of a passenger staying at the station due to choosing that valid route is determined by the passenger's choice of that valid route. The probability of a path is determined by the ratio of the number of stages in the passenger's complete journey. A multi-path probability selection model (MNL) is used to calculate the probability of a passenger choosing each path. Specifically, the steps include: calculating the basic cost of each path based on the passenger's walking time from the entrance gate to the platform, the passenger's walking time from the platform to the exit gate, the passenger's travel time for choosing that path, and the total waiting time for that path; calculating the penalty cost of each path based on the penalty time caused by transfers and congestion; obtaining the impedance function of each path based on the basic cost and penalty cost; constructing the utility function of each path based on its impedance function; and calculating the probability of the passenger choosing each path using the MNL model based on the utility function of that path. Based on the dwell rate of non-transfer passengers at the originating station and the dwell rate of transfer passengers at the station, the dwell rate of passengers at the station to be tested is estimated.

2. The method according to claim 1, wherein, The process of determining the proportion of passengers who did not board the first train to complete their journey, based on each non-transfer passenger's entry time, on-train time, and exit time, out of the station, and thus obtaining the non-transfer passenger congestion rate at the originating station, specifically includes the following steps: Based on the AFC card swipe data of all passengers entering and exiting the urban rail transit station within a set time interval, the actual travel time of all passengers between the same origin and destination (OD) is obtained. Sort all passengers’ actual travel times and select the minimum value as the minimum theoretical travel time for each passenger. The passenger's walking time to and from the station is determined by the difference between the passenger's theoretical minimum travel time and the time spent in the vehicle; The waiting time for passengers is determined by the difference between the arrival or departure times of trains arriving at or departing from the same station. Based on the passenger's walking time to enter the station, walking time to exit the station, time on the train, and waiting time, the passenger's theoretical maximum travel time is determined; Whether a passenger should remain on the train is determined by the ratio of their actual travel time to the theoretical maximum travel time. If a passenger remains on the train, the retention rate of non-transfer passengers at the originating station is calculated based on the proportion of passengers who did not complete their journey on the first train to the number of passengers who entered the station within the set time interval.

3. The method of claim 1, wherein, The process of obtaining the passenger's exit time specifically includes the following steps: For any station s, the direction of travel d The exit times of passengers on board all follow the corresponding probability density function distribution. ; The probability density function of passenger exit time can be expressed as follows using the conditional probability density function: ； in, This is the station's theoretical minimum exit time. The passenger's exit time is the difference between the time the passenger swipes their card to exit and the train's arrival time. , It refers to the train departure interval; The conditional probability density function representing the passenger's exit time. Indicates passenger i The departure time is less than the station s The arrival train departure interval is greater than the station's departure interval. s The probability of the theoretical minimum departure time; ； Then we get ; The exit time probability of all passengers is obtained by multiplying the conditional probability density functions of all passengers' exit times, and a log-likelihood function is established. Based on the log-likelihood function, the maximum likelihood estimation method is used to estimate the passenger exit time for all passengers at all stations in all directions.

4. The method for estimating the dwell rate of urban rail transit stations according to claim 3, characterized in that, The passenger's entry time includes the walking time to enter the station and the waiting time; the process of obtaining this time specifically includes the following steps: Passengers arriving at any departure interval follow a uniform distribution. Then the probability density function of the passenger's waiting time is expressed as: ； in, , For passengers i The time of gathering, For passengers i Walking time to enter the station, For passengers i The waiting time for the bus Indicates the train's departure time. Indicates the time when the passenger swipes their card to enter the station; For any station s their travel direction d The walking time of passengers entering the station all follow the corresponding probability density function distribution. For passengers with only one feasible travel option, their probability density function distribution is expressed as follows: In this case, and The joint probability of simultaneous occurrence: ； in The conditional probability density function represents the walking time for passengers to enter the station. This indicates that the passenger's walking time to the station is greater than or equal to the minimum passenger walking time, and less than or equal to the minimum passenger walking time. i The time for people to gather at the station; The probability that the sum of a passenger's walking time to the station and their waiting time is less than or equal to the gathering time is expressed as follows: ； ； Then we get ; in, Indicates the maximum passenger walking time. Indicates the minimum passenger walking time. This indicates the integral variable and its upper limit for waiting; The conditional probability density functions of the arrival times of all passengers are multiplied together to obtain the probability of the departure times of all passengers, and a log-likelihood function is established. Based on the log-likelihood function, the maximum likelihood estimation method is used to estimate the passenger arrival time of all passengers at all stations in all directions.

5. The method for estimating the dwell rate of urban rail transit stations according to claim 1, characterized in that, It also includes, after filtering out the valid routes for each transfer passenger based on their entry and exit card swiping stations, using the transfer passenger's exit card swiping time and the train exit time threshold of the theoretical travel plan to determine the passenger's choice of each valid route for boarding; wherein, the theoretical travel plan is the travel plan corresponding to the first train that the passenger can board after arriving at the waiting platform.

6. A system for estimating the dwell rate of urban rail transit stations, characterized in that, include: The data acquisition module is used to acquire the entry and exit card swiping stations, entry and exit times, and on-train time of all passengers entering and exiting the urban rail transit test station within a set time interval; The non-transfer passenger retention rate calculation module is used to determine the proportion of passengers who did not board the first train to complete their journey to the number of passengers who entered the station within a set time interval, based on each non-transfer passenger's entry time, on-train time, and exit time, thereby obtaining the non-transfer passenger retention rate at the originating station. The passenger delay rate calculation module is used to filter out the effective routes for each transfer passenger based on their entry and exit card swiping stations. It selects all effective routes that include the test station's path in a specific train direction and calculates the probability of the passenger delaying at the station by choosing that effective route, or the probability of the passenger traveling on that effective route but not delaying. This determines the probability that the passenger will choose an effective route containing the test station's path in a specific train direction and delay. Based on the probability of each transfer passenger choosing an effective route containing the test station's path in a specific train direction and delaying, and the probability of choosing an effective route containing the test station's path in a specific train direction but not delaying, the station delay rate considering transfer passengers is estimated. The probability of the passenger choosing the effective route and delaying at the station is... The probability of a passenger choosing a route is determined by the ratio of the probability of that route to the number of stages in the passenger's complete journey. A multi-path probability choice model (MNL) is then used to calculate the probability of a passenger choosing each route. Specifically, the steps include: calculating the basic cost of each route based on the passenger's walking time from the entrance gate to the platform, the passenger's walking time from the platform to the exit gate, the passenger's travel time for choosing that route, and the total waiting time for that route; calculating the penalty cost of each route based on the penalty time caused by transfers and congestion; obtaining the impedance function for each route based on the basic cost and penalty cost; constructing the utility function of each route based on its impedance function; and using the MNL model to calculate the probability of a passenger choosing each route based on the utility function of that route. The estimation module is used to estimate the passenger dwell rate of the station to be tested based on the dwell rate of non-transfer passengers at the originating station and the station dwell rate of transfer passengers.

Images