Multi-source data fusion-based passenger aggregation prediction method and system for passenger transport hub

Abstract

The invention relates to the technical field of passenger transport traffic scheduling, in particular to a passenger transport hub passenger aggregation prediction method based on multi-source data fusion, which comprises the following steps: acquiring multi-source data; constructing an operation tracking model of passenger aggregation of the passenger transport hub based on the multi-source data; and adopting the operation tracking model for passenger aggregation of the passenger transport hub to predict the passenger aggregation condition of the passenger transport station. According to the method, a multi-source passenger gathering data set is collected, a passenger station passenger flow gathering rule is described from different angles, and a passenger transportation hub passenger gathering operation and tracking model is established by utilizing station entering gate data, ticket checking gate data, a passenger station train working diagram, actual arrival and departure time and ticket selling data; the method can predict passenger transport hub waiting hall on-site gathering people number conditions in real time, can adjust and predict future waiting hall on-site gathering people number change conditions according to a subsequent passenger transport work chart, thus providing scientific references for passenger transport hub daily and emergency passenger evacuation work.

Description

Technical Field

[0001] The present invention relates to the technical field of passenger transportation scheduling, and more specifically, to a passenger gathering prediction method and system for a passenger transportation hub based on multi-source data fusion. Background Art

[0002] At present, passenger stations are indispensable infrastructure for passenger transportation, especially railway passenger stations, which are important places for passengers to travel. With the continuous increase in railway operating mileage and the continuous growth of railway passenger volume, the scale of passenger station buildings with corresponding capacity is an indispensable link in the development of railway passenger transportation. The study of the maximum number of people gathered at railway passenger stations is of great practical significance for judging the gathering of railway passenger flow during major holidays (National Day, Spring Festival, etc.), formulating passenger evacuation emergency plans, and organizing and managing the railway waiting hall on site; it also has important scientific reference significance for the scale prediction of the new and renovated railway passenger station waiting hall buildings.

[0003] Current domestic and international situation:

[0004] (1) He Yuqiang et al. proposed three methods to calculate the maximum number of passengers at railway passenger stations (clustering coefficient method, one-way train frequency method and graphical method). The clustering coefficient method is simple to calculate and widely used, and is a more commonly used calculation method. The one-way train frequency method is mainly suitable for calculating the maximum number of passengers at small railway passenger stations. The traditional graphical method predicts the maximum number of passengers by drawing the passenger cluster curve during the peak train departure period.

[0005] (2) Oded et al. used computer simulation modeling to visualize the passenger entry process and activities within railway passenger stations; Chen Xichun et al. realized the fitting of the two-parameter Weibull distribution curve of passenger arrival time and proposed a systematic mathematical prediction model for the number of passengers gathering at railway passenger stations.

[0006] (3) Ingvardson et al. classified the early arrival behavior of railway passengers into two types: one is random arrival, and the other is arriving according to the train schedule and minimizing waiting time as much as possible so as not to miss the train. By comprehensively considering these two types of behavior, a passenger waiting time calculation model was established using the uniform distribution and beta distribution to describe the mixed distribution of passenger arrivals. Guo et al. analyzed the arrival patterns of passengers on the Beijing subway and proposed that the lognormal distribution and the gamma distribution can be used to describe the early arrival time distribution of direct transfer and non-direct transfer passengers, respectively, and based on this, proposed an average waiting time model for transfer passengers.

[0007] (4) Yang et al. pointed out that there are differences in the arrival patterns of passengers between the starting and ending trains and the passing trains, and between trains with different destinations and different train models, and conducted a binary correlation analysis. Zhang Tianwei et al. pointed out that passengers traveling at different times are affected by their travel habits, and their respective early arrival patterns are different.

[0008] Analysis of the current domestic and international research and technology status is as follows:

[0009] First, in terms of model research. Regarding the prediction model of the maximum number of passengers gathered at railway passenger stations, the research content in the direction of probability theory is relatively rich. Domestic and foreign scholars have successively proposed using different functions to describe the distribution law of passengers' advance arrival time, mainly including normal distribution, negative exponential distribution, Weibull distribution, Gaussian distribution and Gamma distribution, etc.:

[0010] Second, in terms of data sources, the data on passenger arrival time distribution, train departure time, ticket checking cycle and time distribution used in the passenger aggregation model of railway passenger transport hubs are mainly derived from train timetables, mobile phone signaling data, on-site survey data, etc.

[0011] Third, in terms of real-time detection, real-time queue length detection methods are mainly based on collecting signaling data and actual on-site investigations, which have high data costs and research results, and have a certain lag.

[0012] Compared with traditional passenger data collection methods, the passenger gathering calculation and tracking model of the railway passenger hub based on multi-source data fusion uses railway station gate data, railway ticket gate data, railway passenger station train operation diagram and actual arrival and departure time, ticket sales data, etc. It can predict the number of people gathering in the waiting hall of the railway hub in real time, and can also adjust and predict the changes in the number of people gathering in the waiting hall in the future according to the subsequent railway operation diagram, providing a scientific reference for passenger evacuation work in daily and emergency situations. Summary of the invention

[0013] In order to better solve the shortcomings of traditional methods, the present invention proposes a passenger gathering prediction method for passenger transport hubs based on multi-source data fusion. The present invention collects multi-source passenger gathering data sets, describes the passenger flow gathering rules of passenger stations from different angles, establishes a passenger gathering calculation and tracking model for passenger transport hubs, measures the number of people gathered in the waiting hall of the passenger transport hub in real time, and adjusts and predicts the changes in the number of people gathered in the waiting hall in the future according to the subsequent passenger operation diagram, providing a scientific reference for the daily and emergency passenger evacuation work of passenger transport hubs.

[0014] The present invention also proposes a system for applying a passenger hub passenger gathering prediction method based on multi-source data fusion.

[0015] The technical solution of the present invention is:

[0016] A passenger gathering prediction method for a passenger transport hub based on multi-source data fusion includes the following steps:

[0017] S1: Collect multi-source data;

[0018] S2: Construct a computational tracking model for passenger gathering at passenger transport hubs;

[0019] S3: Use the computational tracking model of passenger gathering at passenger transport hubs to predict passenger gathering at passenger stations.

[0020] Preferably, the multi-source data in step S1 include: vehicle operation diagram and punctuality rate, vehicle capacity and passenger boarding rate, and station gate data and ticket gate data, and statistical analysis is performed on the multi-source data respectively;

[0021] By combining the vehicle operation diagram and punctuality statistics, vehicle capacity and passenger boarding rate statistics, and station entrance gate data and ticket gate data statistics, we statistically analyze the patterns and data of passenger arrivals.

[0022] Preferably, the specific method of constructing the calculation tracking model of passenger gathering at the passenger transport hub in step S2 is:

[0023] S201: Construct the passenger arrival distribution density function; let is the passenger's early arrival distribution density function, x is the passenger's early arrival time, The maximum advance arrival time for passengers, exceeding The passengers missed the train and are invalid data. The passenger early arrival distribution density function is transformed into the passenger arrival distribution density function through axisymmetric transformation. ;

[0024]

[0025] S202: constructing a passenger arrival probability distribution function; combining the vehicle operation diagram and the punctuality rate, vehicle capacity and passenger loading rate; using the Gaussian kernel model L2 constrained least squares learning method fitting algorithm, by introducing the Gaussian kernel function, mapping the original data to a high-dimensional feature space, and then fitting using the L2 constrained least squares method in the feature space;

[0026] First, we get the probability distribution function of passenger arrival of the first vehicle:

[0027]

[0028] Determine the positive parameters , , , , , },in are the rated passenger weight constants of the originating vehicle and the passing vehicle, respectively, which are used to comprehensively describe and measure the probability distribution function of different types of vehicles. , Describe the expectation and tolerance of the Gaussian distribution function of the originating vehicle respectively, , Describe the expectation and tolerance of the Gaussian distribution function of the passing vehicles respectively, and comprehensively reflect the shape characteristics of the distribution function;

[0029] Assume t i is the departure time difference between the ith vehicle and the first vehicle, x is the early arrival time of the passenger, and the arrival probability distribution function of the passenger of the ith vehicle is:

[0030]

[0031] Let Q i is the vehicle capacity of the i-th vehicle; P i is the passenger boarding rate of the i-th vehicle; the passenger arrival probability distribution function is:

[0032]

[0033] S203: Constructing the passenger evacuation probability distribution function H(x); combining the ticket checking time period at the station ticket gate -t 1 The actual situation: the evacuation process of passengers checking in and entering the station for the originating vehicles and passing vehicles is uniformly distributed;

[0034] The probability distribution density function of passenger evacuation on the first vehicle is:

[0035] }

[0036] Passenger evacuation probability distribution function:

[0037]

[0038] S204: the function of the number of people gathering over time is the passenger arrival probability distribution function minus the passenger evacuation probability distribution function;

[0039]

[0040] Assume T is the time axis, T takes the value of (0:1440) minutes, the maximum number of people waiting in the waiting room at any time, that is, the function of the number of people gathering over time: .

[0041] A system for applying the above prediction method comprises the following modules:

[0042] Collect multi-source data module;

[0043] Model construction module: construct a computational tracking model for passenger gathering at passenger transport hubs;

[0044] Prediction module: The passenger gathering situation at the passenger station is predicted by using the calculation and tracking model of passenger gathering at the passenger hub.

[0045] Preferably, the multi-source data collection module includes collecting vehicle operation diagrams and punctuality rates, vehicle capacity and passenger boarding rates, and station gate data and ticket gate data, and performing statistical analysis on the multi-source data respectively;

[0046] By combining the vehicle operation diagram and punctuality statistics, vehicle capacity and passenger boarding rate statistics, and station entrance gate data and ticket gate data statistics, we statistically analyze the patterns and data of passenger arrivals.

[0047] Preferably, the specific manner of constructing the model module is:

[0048] Construct the passenger arrival distribution density function unit; let is the passenger's early arrival distribution density function, x is the passenger's early arrival time, The maximum advance arrival time for passengers, exceeding The passengers missed the train and are invalid data. The passenger early arrival distribution density function is transformed into the passenger arrival distribution density function through axisymmetric transformation. ;

[0049]

[0050] Construct a passenger arrival probability distribution function unit; combine the vehicle operation diagram and punctuality rate, vehicle capacity and passenger loading rate; use the Gaussian kernel model L2 constrained least squares learning method fitting algorithm, introduce the Gaussian kernel function, map the original data to a high-dimensional feature space, and then use the L2 constrained least squares method to fit in the feature space;

[0051] First, we get the probability distribution function of passenger arrival of the first vehicle:

[0052]

[0053] Determine the positive parameters , , , , , },in are the rated passenger weight constants of the originating vehicle and the passing vehicle, respectively, which are used to comprehensively describe and measure the probability distribution function of different types of vehicles. , Describe the expectation and tolerance of the Gaussian distribution function of the originating vehicle respectively, , Describe the expectation and tolerance of the Gaussian distribution function of the passing vehicles respectively, and comprehensively reflect the shape characteristics of the distribution function;

[0054] Assume t i is the departure time difference between the ith vehicle and the first vehicle, x is the early arrival time of the passenger, and the arrival probability distribution function of the passenger of the ith vehicle is:

[0055]

[0056] Let Q i is the vehicle capacity of the i-th vehicle; P i is the passenger boarding rate of the i-th vehicle; the passenger arrival probability distribution function is:

[0057]

[0058] Construct a passenger evacuation probability distribution function unit; combine the ticket checking time period at the station ticket gate -t 1 The actual situation: the evacuation process of passengers checking in and entering the station for the originating vehicles and passing vehicles is uniformly distributed;

[0059] The probability distribution density function of passenger evacuation on the first vehicle is:

[0060] }

[0061] Passenger evacuation probability distribution function:

[0062]

[0063] The function of the number of people gathering over time is the passenger arrival probability distribution function minus the passenger evacuation probability distribution function;

[0064]

[0065] Assume T is the time axis, T takes the value of (0:1440) minutes, the maximum number of people waiting in the waiting room at any time, that is, the function of the number of people gathering over time: .

[0066] Compared with the prior art, the present invention has the following characteristics:

[0067] (1) Having a good practical mechanism;

[0068] (2) It can save data and labor costs, realize the real-time tracking of the number of people gathered in the waiting hall of the railway hub, and make scientific predictions on the subsequent changes in the number of people gathered;

[0069] (3) The method is simple and has good processing efficiency while meeting a certain accuracy requirement. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 The figure is a flow chart of the method of the present invention.

[0071] Figure 2 Schematic diagram of passenger arrival patterns at Guangzhou South Railway Station in July 2024.

[0072] Figure 3 This is a diagram showing how the net increase in the number of people gathered changes over time.

[0073] Figure 4 This is a graph showing how the number of people gathered changes over time. DETAILED DESCRIPTION

[0075] The drawings are only used for illustrative purposes and should not be construed as limiting the present invention. In order to better illustrate the present embodiment, some parts of the drawings may be omitted, enlarged, or reduced, and do not represent the size of the actual product.

[0076] For those skilled in the art, it is understandable that some well-known structures and their descriptions may be omitted in the drawings. The technical solution of the present invention is further described below in conjunction with the drawings and embodiments.

[0077] The present invention improves, amends and supplements the probability calculation method of the maximum number of people gathered. In this embodiment, the passenger station is selected as a high-speed railway station.

[0078] Specifically, it collects multi-source passenger gathering data sets, describes the passenger flow gathering patterns of railway passenger stations from different angles, and uses railway entrance gate data, railway ticket gate data, railway passenger station train operation diagram and actual arrival and departure times, and ticket sales data to establish a railway passenger hub passenger gathering calculation and tracking model, and predict the number of people gathering in the waiting hall of the railway hub in real time. It also adjusts and predicts the changes in the number of people gathering in the waiting hall in the future based on subsequent railway operation diagrams, providing scientific reference for daily and emergency passenger evacuation work in railway passenger hubs.

[0079] Multivariate data collection and fusion

[0080] 1. Train operation diagram and punctuality statistics analysis Collect and collect statistics on the National Day peak train operation data at Guangzhou South Railway Station and other stations from September 30 to October 8, 2024, and count the number of train runs, ticket quotas and punctuality rates.

[0081] According to statistics, during the National Day in 2024:

[0082] (1) Guangzhou South Railway Station operates 644 pairs of trains to and from Guangzhou South Railway Station, including 364 pairs of trains that depart and arrive at the railway station, and 280 pairs of trains that pass through the railway station;

[0083] (2) The overall punctuality rate was 98.3%, of which the punctuality rate of originating and terminating trains was 99.1% and the punctuality rate of transit trains was 97.3%.

[0084] (3) The average ticket quota is 67.2%, of which 80% is for departure trains and 8.25% is for transit trains.

[0085] 2. Statistical analysis of train capacity and passenger loading rate

[0086] Collect and count the National Day peak train operation data, train capacity and passenger boarding rate at Guangzhou South Station and other stations from September 30 to October 8, 2024. The train capacity is the number of tickets allocated to Guangzhou South Station, and the passenger boarding rate is predicted by dividing the actual number of people by the number of allocated tickets.

[0087] According to statistics, during the National Day in 2024:

[0088] (1) The average number of passengers on trains departing from or arriving at Guangzhou South Station is 653, of which the average number of passengers on trains departing or arriving at Guangzhou South Station is 778, and the average number of passengers on trains passing through Guangzhou South Station is 247.

[0089] (2) The overall average occupancy rate at Guangzhou South Railway Station was approximately 107.5%, of which the occupancy rate for trains with the highest departure and arrival points was 111.2%, and the occupancy rate for trains passing through was 102.4%.

[0090] 3. Statistical analysis of gate arrival data Collect the passenger flow data and ticket gate data of Guangzhou South Station in mid-July 2024, and count the changes in the number of people entering Guangzhou South Station 24 hours a day. Combined with the train capacity and passenger boarding rate statistics of the train operation diagram and the station gate data of Guangzhou South Station, statistically analyze the passenger arrival patterns and data, and provide a data basis for the subsequent passenger arrival pattern and gathering number prediction model, such as Figure 2 The figure shows the arrival pattern of passengers at Guangzhou South Railway Station.

[0091] The train operation diagram, punctuality rate, train capacity and passenger boarding rate are comprehensively utilized to describe the passenger flow aggregation pattern of railway passenger stations from different angles, build a data foundation for subsequent model construction, and use the gate arrival data to debug the model parameters to improve the accuracy and practicality of the model.

[0092] Building calculation and tracking models

[0093] Step 1: Construct the time distribution function of passenger arrival. is the passenger's advance arrival time distribution function, x is the advance arrival time, is the maximum advance arrival time of passengers. In this embodiment The value is 120 minutes. The passengers missed the train and are invalid data. The passenger early arrival time distribution function is transformed into the passenger arrival time distribution function through axisymmetric transformation. ;

[0094]

[0095] Step 2: Construct the probability distribution function of passenger arrival. Combined with the train operation diagram and punctuality rate, train capacity and passenger boarding rate and ticket purchase data, the Gaussian kernel model L2 constrained least squares learning method fitting algorithm is used. By introducing the Gaussian kernel function, the original data is mapped to a high-dimensional feature space, and then the L2 constrained least squares method is used for fitting in the feature space. Taking the passenger entry data statistics of G818 at Guangzhou South Station as an example, the Gaussian distribution of the number of arrivals of this train is obtained as shown in the following table.

[0096]

[0097] The fitting results of Gaussian distribution parameters of the number of arrivals are as follows:

[0098] General model Gauss2:

[0099] f(x)=a1*exp(-((x-b1) / c1)^2)+a2*exp(-((x-b2) / c2)

[0100] Coefficients (with 95% confidence bounds):

[0101] a1 = 0.09136 (0.059, 0.1237)

[0102] b1 = 17.94 (17.62,18.27)

[0103] c1 = 3.749 (3.118, 4.379)

[0104] a2 = 0.03391(0.02284,0.04497)

[0105] b2 = 11.61 (7.343,15.88)

[0106] c2 = 6.782 (3.311,10.25)

[0107] Goodness of fit:

[0108] SSE: 0.0002241

[0109] R-square: 0.9914

[0110] Adjusted R-square: 0.989

[0111] RMSE: 0.003529

[0112] First, we get the probability distribution density function of the passengers arriving on the first train:

[0113]

[0114] Determine the positive parameters , , , , , },in are the rated passenger weight constants of the originating vehicle and the passing vehicle, respectively, which are used to comprehensively describe and measure the probability distribution function of different types of vehicles. , Describe the expectation and tolerance of the Gaussian distribution function of the originating vehicle respectively, , Describe the expectation and tolerance of the Gaussian distribution function of the passing vehicles respectively, and comprehensively reflect the shape characteristics of the distribution function;

[0115] Assume t i is the departure time difference between the ith vehicle and the first vehicle, x is the early arrival time of the passenger, and the arrival probability distribution function of the passenger of the ith vehicle is:

[0116]

[0117] Let Q i is the train capacity of the i-th train; Pi is the occupancy rate of the i-th train. The overall probability distribution function of passenger arrival is:

[0118]

[0119] Step 3: Construct the probability distribution function of passenger evacuation. Combined with the ticket checking time period at the station ticket gate -t 1 Actual data show that, in this embodiment, the ticket inspector starts checking tickets 20 minutes before the departure of the departure train, the ticket checking time is 15 minutes, and the ticket checking is stopped 5 minutes before departure. The evacuation process of passengers checking tickets and entering the station can be approximately regarded as obeying uniform distribution, and the same is true for passing trains.

[0120] The first train passenger evacuation distribution density function:

[0121] ,

[0122] Passenger evacuation probability distribution function:

[0123]

[0124] Step 4: The function of the number of people gathering over time is the passenger arrival distribution function minus the passenger evacuation distribution function.

[0125]

[0126] Assume T is the time axis, T takes the value of (0:1440) minutes, the maximum number of people waiting in the waiting room at any time, that is, the function of the number of people gathering over time:

[0127]

[0128] Gathering crowd prediction

[0129] Using Matlab combined with the actual train operation plan and evacuation plan of the high-speed railway station, taking the train capacity and occupancy rate as the basic data, combined with the calculation and tracking model of the number of people gathered, the function of the change of the number of people gathered over time is obtained, and the curve of the change of the net increase of the number of people gathered over time is shown as follows Figure 3 shown.

[0130] % -------------Distribution function of passenger arrivals--------

[0131] TrainData = xlsread('filename') % Train data input

[0132] [TrM,TrN] = size(TrainData) % Number of rows and columns of the train matrix

[0133] ssem=zero(TrM,30*24*60) % Time / train-number of arrivals matrix, assigned 0

[0134] Eva=zero(TrM,30*24*60) % Time / train-evacuation number matrix, assigned 0

[0135] for i=1:TrM t=1:25

[0136] AssemDen = a1.*exp(-((t-b1) / c1)^2)+a2.*exp(-((t-b2) / c2)^2) % Passenger arrival density function

[0137] AssemTrain = TrainData(i, capacity).*TrainData(i, occupancy).*AssemDen % Passenger arrival function

[0138] Assem(i,(TrainData(i,time)-25):(TrainData(i,time)-1)) = AssemTrain% Time / train-arrival matrix

[0139] EvaDen = ones(1,15).*(1 / 15) % Passenger evacuation density function (uniform distribution)

[0140] EvaTrain = TrainData(i, capacity).*TrainData(i, occupancy rate).*EvaDen % Passenger evacuation function

[0141] Eva(i,(TrainData(i,time)-20):(TrainData(i,time)-6)) = EvaTrain %Time / train-evacuation number matrix

[0142] End

[0143] Using Matlab, we can combine the function of the number of people gathering over time, the maximum number of people gathering at the bus station, and the gathering time. The curve of the number of people gathering over time is as follows: Figure 4 As shown:

[0144] NetAss = sum(Assem)-sum(Eva) % Time-net inflow matrix

[0145] for i = 1:(30*24*60-9)

[0146] TotNum(i) = sum(NetAss(i:(i+9)))

[0147] MaTot = max(TotNum)

[0148] The calculation results show that the number of people gathered in the waiting hall of Guangzhou South Station from 2:00 a.m. to 5:00 a.m. in the morning is relatively small, and the overall number is maintained below 2,000. It starts to increase to 3,000 from 7:00 a.m. and continues to grow. From 8:00 a.m. to 16:00 p.m., the number of people gathered basically remains above 5,000. Among them, the peaks are reached at 8:30 a.m. and 10:30 a.m., which are 7,462 and 7,121 respectively. The fluctuation of the real-time number of people gathered in the railway station for 24 hours a day is consistent with the train operation plan and ticket inspection plan of Guangzhou South Station throughout the day. The mathematical model is indeed effective and can be used in abnormal situations of railway hubs (natural disasters such as freezing rain, line failures, etc. cause some trains to be delayed). By solving the speed and number of passenger flow gathering during major holidays, emergency passenger evacuation plans at all levels can be formulated and implemented.

[0149] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A passenger gathering prediction method for passenger transport hubs based on multi-source data fusion, characterized in that: The following steps are involved: S1: Collect multi-source data; S2: Construct a computational tracking model for passenger gathering at passenger transport hubs; S3: Use the computational tracking model of passenger gathering at passenger transport hubs to predict passenger gathering at passenger stations.

2. The prediction method according to claim 1, characterized in that: The multi-source data in step S1 include: vehicle operation diagram and punctuality rate, vehicle capacity and passenger boarding rate, and station gate data and ticket gate data, and statistical analysis is performed on the multi-source data respectively; By combining the vehicle operation diagram and punctuality statistics, vehicle capacity and passenger boarding rate statistics, and station entrance gate data and ticket gate data statistics, we statistically analyze the patterns and data of passenger arrivals.

3. The prediction method according to claim 2, characterized in that: The specific method of constructing the calculation tracking model of passenger gathering at the passenger transport hub in step S2 is: S201: Construct the passenger arrival distribution density function; let is the passenger's early arrival distribution density function, x is the passenger's early arrival time, The maximum advance arrival time for passengers, exceeding The passengers missed the train and are invalid data. The passenger early arrival distribution density function is transformed into the passenger arrival distribution density function through axisymmetric transformation. ; ； S202: constructing a passenger arrival probability distribution function; combining the vehicle operation diagram and the punctuality rate, vehicle capacity and passenger loading rate; using the Gaussian kernel model L2 constrained least squares learning method fitting algorithm, by introducing the Gaussian kernel function, mapping the original data to a high-dimensional feature space, and then fitting using the L2 constrained least squares method in the feature space; First, we get the probability distribution function of passenger arrival of the first vehicle: ； Determine the positive parameters , , , , , },in are the rated passenger weight constants of the originating vehicle and the passing vehicle, respectively, which are used to comprehensively describe and measure the probability distribution function of different types of vehicles. , Describe the expectation and tolerance of the Gaussian distribution function of the originating vehicle respectively, , Describe the expectation and tolerance of the Gaussian distribution function of the passing vehicles respectively, and comprehensively reflect the shape characteristics of the distribution function; set up t i is the departure time difference between the ith vehicle and the first vehicle, x is the early arrival time of the passenger, and the arrival probability distribution function of the passenger of the ith vehicle is: ； Let Q i is the vehicle capacity of the i-th vehicle; P i is the passenger boarding rate of the i-th vehicle; the passenger arrival probability distribution function is: ； S203: Constructing the passenger evacuation probability distribution function H(x); combining the ticket checking time period at the station ticket gate -t1 Actual situation: The evacuation process of passengers checking in and entering the station for the originating vehicle and the passing vehicle is uniformly distributed; The probability distribution density function of passenger evacuation on the first vehicle is: }； Passenger evacuation probability distribution function: ； S204: the function of the number of people gathering over time is the passenger arrival probability distribution function minus the passenger evacuation probability distribution function; ； Assume T is the time axis, T takes the value of (0:1440) minutes, the maximum number of people waiting in the waiting room at any time, that is, the function of the number of people gathering over time: .

4. A system using any one of the methods of claims 1 to 3, characterized in that: Includes the following modules: Collect multi-source data module; Model construction module: construct a computational tracking model for passenger gathering at passenger transport hubs; Prediction module: The passenger gathering situation at the passenger station is predicted by using the calculation and tracking model of passenger gathering at the passenger hub.

5. The system according to claim 4, characterized in that The multi-source data collection module includes collecting vehicle operation diagrams and punctuality rates, vehicle capacity and passenger boarding rates, and station gate data and ticket gate data, and performing statistical analysis on the multi-source data respectively; By combining the vehicle operation diagram and punctuality statistics, vehicle capacity and passenger boarding rate statistics, and station entrance gate data and ticket gate data statistics, we statistically analyze the patterns and data of passenger arrivals.

6. The system according to claim 5, characterized in that The specific method of constructing the model module is: Construct the passenger arrival distribution density function unit; let is the passenger's early arrival distribution density function, x is the passenger's early arrival time, The maximum advance arrival time for passengers, exceeding The passengers missed the train and are invalid data. The passenger early arrival distribution density function is transformed into the passenger arrival distribution density function through axisymmetric transformation. ; ； Construct a passenger arrival probability distribution function unit; combine the vehicle operation diagram and punctuality rate, vehicle capacity and passenger loading rate; use the Gaussian kernel model L2 constrained least squares learning method fitting algorithm, introduce the Gaussian kernel function, map the original data to a high-dimensional feature space, and then use the L2 constrained least squares method to fit in the feature space; First, we get the probability distribution function of passenger arrival of the first vehicle: ； Determine the positive parameters , , , , , },in are the rated passenger weight constants of the originating vehicle and the passing vehicle, respectively, which are used to comprehensively describe and measure the probability distribution function of different types of vehicles. , Describe the expectation and tolerance of the Gaussian distribution function of the originating vehicle respectively, , Describe the expectation and tolerance of the Gaussian distribution function of the passing vehicles respectively, and comprehensively reflect the shape characteristics of the distribution function; set up t i is the departure time difference between the ith vehicle and the first vehicle, x is the early arrival time of the passenger, and the arrival probability distribution function of the passenger of the ith vehicle is: ； Let Q i is the vehicle capacity of the i-th vehicle; P i is the passenger boarding rate of the i-th vehicle; the passenger arrival probability distribution function is: ； Construct a passenger evacuation probability distribution function unit; combine the ticket checking time period at the station ticket gate -t1 Actual situation: The evacuation process of passengers checking in and entering the station for the originating vehicle and the passing vehicle is uniformly distributed; The probability distribution density function of passenger evacuation on the first vehicle is: }； Passenger evacuation probability distribution function: ； The function of the number of people gathering over time is the passenger arrival probability distribution function minus the passenger evacuation probability distribution function; ； Assume T is the time axis, T takes the value of (0:1440) minutes, the maximum number of people waiting in the waiting room at any time, that is, the function of the number of people gathering over time: .

Images