Rail digital twin real-time online deduction method and system

Abstract

The application discloses a kind of track digital twin real-time online deduction method and system, belong to the traffic simulation technical field based on machine learning.Solve the problem that the conventional traffic simulation and track passenger flow dynamic allocation method in prior art is difficult to meet the demand of large-scale track traffic network dynamic passenger flow allocation due to low solving efficiency;The application constructs track network model;Generate road network passenger flow data and line schedule data;Further generate train and passenger simulation individual, carry out path trajectory matching based on time and space, complete passenger-train dynamic traffic allocation and passenger-train interactive operation deduction;According to real-time detection data and simulation result, adjust dynamic line schedule and passenger flow until the preset convergence condition is reached, complete dynamic traffic checking;Simulation is carried out, realize the restoration and prediction of passenger flow flow rate / flow direction of whole line network.The application effectively improves the efficiency of track passenger flow dynamic allocation, and can be applied to large-scale traffic network simulation.

Description

TECHNICAL FIELD

[0001] The present application relates to a kind of track digital twin real-time online deduction method and system, belong to the traffic simulation technical field based on machine learning. BACKGROUND

[0002] At present, there are many problems in the operation and dispatching of rail network, as follows: (1) After the operation of rail network, the planning of transportation, the pressure of operation and dispatching increases: In the stage of metro network operation, the sudden event is triggered by single point, and the whole body is activated, even affecting road traffic, causing regional traffic paralysis, for this, some cities have established network operation center NOCC, network control center NCC, transfer center TCC and other departments, and invested a lot of manpower and material resources to build infrastructure, but the investment in software construction is generally insufficient, which cannot meet the requirements of intelligent operation; (2) The train and station are crowded seriously due to peak large passenger flow, and the passenger travel experience is poor: With the rapid rise of network passenger flow, the normalization of peak large passenger flow has caused great pressure on metro operation, and the queuing phenomenon is serious in some stations, so that the large passenger flow is controlled, and the large passenger flow may also bring the risk of trampling, and the service level of metro is difficult to meet the increasing travel requirements of passengers (3) The operation and dispatching and passenger flow control lack real-time passenger flow basis, and still rely on "human sea tactics" to a great extent. Therefore, in terms of online network dispatching, resource allocation, station large passenger flow control, etc., due to the inability to accurately grasp the large passenger flow situation and space-time distribution, the train dispatching is difficult to accurately allocate the transport capacity, and the station control can only rely on manual experience to organize, lacking intelligent management means.

[0003] In the prior art, traffic simulation technology is one of the common technical means to analyze passenger flow situation and space-time distribution. Under the background of the increasing complexity of urban transportation system and the normalization of traffic congestion, with the rapid development of modern information technology such as big data, mobile internet and cloud computing, current mobile phone signaling, video and other multi-source big data can be accessed into the system in real time, how to utilize and mine the potential value of these multi-source heterogeneous big data and apply them to traffic simulation to analyze rail transit passenger flow situation and distribution, realize passenger portrait, line network intelligent dispatching, station passenger flow fine control, multi-level linkage intelligent emergency, intelligent travel navigation and other functions, so as to improve the organizational coordination ability of enterprise, assist enterprise to make operation decision, and serve passenger convenient travel, which is the front research focus in the international traffic engineering field, and has broad application prospect; But the existing traffic simulation software is mainly small-scale, single-field, offline and single-machine version deduction, which focuses on the simulation function of road traffic mode, mainly road traffic simulation, and lacks rail transit simulation, and the traffic simulation software has problems such as slow calculation speed in large-scale scene, difficult fusion and deduction in cross-field, and difficult to adapt to different scale elastic deployment to a certain extent.

[0004] Rail passenger flow dynamic distribution is to distribute passenger flow to each path and section of the rail network according to certain rules when the passenger flow demand is known, so as to obtain the dynamic distribution characteristics of traffic demand in time and space, and is the basis for formulating dynamic operation management strategies such as online operation adjustment strategy, passenger flow induction, emergency disposal measures and transport resource allocation strategy, and is the key to the implementation of network dynamic operation management. Under the condition of large-scale rail network, the number of passenger route choice behaviors is very large and the passenger route choice behaviors are complex, which causes difficulties in analyzing the dynamic distribution state of passenger flow. At present, there is still a lack of dynamic passenger flow distribution method for the characteristics of rail transport, and the existing dynamic traffic flow distribution model is mostly suitable for road traffic network, which has the disadvantages of complex model and low solving efficiency, and it is difficult to meet the actual needs of dynamic passenger flow distribution of large-scale rail network. Therefore, it is necessary to construct a practical and rapid dynamic passenger flow distribution method for the specific rail transport network containing transport schedule, so as to provide the most direct reference basis for formulating management strategies.

[0005] At the same time, it is only the basis for formulating operation management strategies to understand the passenger flow distribution law, and more importantly, it is to construct operation organization strategies on the basis of clear passenger flow distribution characteristics. Due to the short construction period of rail transit, the network structure is not perfect, and the short-term sharp increase of passenger flow demand leads to the prominent structural imbalance between transport demand and supply, especially the serious passenger flow congestion problem in peak hours, and the passenger flow safety risk is increased. The passenger flow demand management measures taking flow limiting as the means become one of the effective measures at present. However, there is still a lack of scientific theoretical basis and calculation method for formulating flow limiting measures in the current operation management process, and more experience is relied on. Therefore, for the typical problem of passenger flow inflow control and flow limiting in peak hours, it is necessary to establish a passenger flow inflow collaborative control model from the relationship between transport capacity and passenger flow demand, to collaboratively control the inflow passenger flow at the station, so as to reduce the passenger delay loss and improve the transport efficiency under the premise of ensuring the operation safety, thereby providing a scientific basis for formulating flow limiting strategies.

[0006] In summary, a rail digital twin real-time online deduction method and system are needed. SUMMARY

[0007] A brief summary of the application is given in the following to provide a basic understanding of some aspects of the application. It should be understood that this summary is not a comprehensive summary of the application. It is not intended to determine the key or important parts of the application, nor to limit the scope of the application. Its purpose is only to give some concepts in a simplified form as a prelude to the more detailed description discussed later.

[0008] In view of this, in order to solve the problem that the traditional traffic simulation and rail passenger flow dynamic allocation method in the prior art cannot meet the demand of large-scale rail transit network dynamic passenger flow allocation due to low solving efficiency, the present application provides a rail digital twin real-time online deduction method and system.

[0009] The technical solution one is as follows: a rail digital twin real-time online deduction method, comprising the following steps:

[0010] S1. Constructing a rail network model comprising a physical network and a space-time network;

[0011] S2. Based on the rail network model, generating road network passenger flow data and line schedule data according to historical passenger flow OD data, real-time card swiping data and planned departure schedule information, and then generating train and passenger simulation individuals respectively, and performing path search on each simulation individual;

[0012] S3. Based on the line schedule data and the road network passenger flow data, performing time and space-based path trajectory matching on the train simulation individuals and passenger simulation individuals after path search, and completing dynamic traffic allocation of passengers-trains;

[0013] S4. Simulating passengers and vehicles according to departure time, path and travel speed, and completing passenger-train interactive operation deduction;

[0014] S5. According to the simulation results obtained after real-time detection data, dynamic traffic allocation of passengers-trains and passenger-train interactive operation deduction, adjusting dynamic line schedule and passenger flow until the preset convergence condition is reached, and completing dynamic traffic checking;

[0015] S6. Based on the adjustment after dynamic traffic checking, using distributed simulation technology for simulation, and realizing accurate restoration and prediction of passenger flow volume / direction of the whole line network.

[0016] Further, in S1, for the construction of the physical network, the line network of the subway is constructed based on the existing line information and station information; for the construction of the space-time network, the "boarding" and "getting off" behaviors of passengers and vehicles are considered, and the space-time network is constructed according to the interactive behaviors between people and vehicles.

[0017] Further, in S2, the road network passenger flow data under different characteristic days and set time periods is generated according to the historical passenger flow OD data, the road network passenger flow data is preliminarily verified and checked in combination with the real-time station in-out passenger flow data, i.e. real-time card swiping data, and then the passenger simulation individuals are generated; the line schedule data is generated according to the planned departure schedule, and then the train simulation individuals are generated.

[0018] Further, in the S3, for a given historical passenger flow OD data, the departure station and the destination station, i.e., O-point and D-point, the earliest departure time EDT and the latest arrival time LAT of a simulation individual are obtained through real-time card swiping data;

[0019] According to the O-point and the D-point, a collection N of possible travel routes of the simulation individual is obtained, the individual path in the collection N is analyzed, all route service paths are traversed through spatial matching, and the stay time window of the individual path and the route service path at the O-point and the D-point is analyzed, whether the two paths can meet the time requirement is observed through time matching, the route service with the smallest time error is selected as the final result of the passenger-train, i.e., the individual path-route service path matching, all historical passenger flow OD data is traversed, and the individual path-route service path matching is completed, so as to realize the dynamic traffic distribution of the passenger-train.

[0020] Further, in the S4, the simulation process of the passenger includes the following behaviors: simulating the entrance into the station through the gate card swiping based on the time when the passenger enters the road network, simulating the walking to the platform based on the flow density speed relationship, simulating the platform queuing for the train based on the queuing model, simulating the boarding and alighting behaviors based on the carriage storage capacity, simulating the walking to the gate based on the flow density speed relationship, and simulating the exit from the station through the gate card swiping based on the time when the passenger leaves the road network.

[0021] The simulation process of the vehicle includes the following behaviors: simulating the departure of the vehicle based on the time when the vehicle enters the road network according to the planned route service, simulating the running according to the route based on the queuing model, simulating the arrival at the station, simulating the departure from the station, and simulating the arrival at the terminal of the route based on the time when the vehicle leaves the road network.

[0022] Further, in the S5, for the dynamic route service adjustment, the route service information of the vehicle is dynamically adjusted according to the arrival and departure data of the real-time route service in the real-time card swiping data, for the dynamic passenger flow OD estimation, the OD total amount and the path are dynamically adjusted based on the given OD path, and according to the adjusted route service and the adjusted passenger flow OD, the passenger-train dynamic traffic distribution in the step S4 and the passenger-train dynamic running deduction in the step S5 are re-performed to obtain the simulation result, until the simulation accuracy or the simulation iteration number requirement is met, and the whole dynamic traffic checking process is completed.

[0023] Further, in the S6, by using the distributed simulation technology, the internal transmission process of each road section and the transfer process of each node in the simulation process are independently processed, the transmission step of all road section sets in the road section and the transfer step of all node sets in the node are realized for the distributed processing of the sub-network.

[0024] Technical Solution 2: A real-time online deduction system for a track digital twin, used to implement the real-time online deduction method for a track digital twin described in Technical Solution 1, comprising a perception layer, a data layer, an algorithm layer, an output layer, and an application layer;

[0025] The perception layer is connected to the data layer, the data layer is connected to the algorithm layer, the algorithm layer is connected to the output layer, and the output layer is connected to the application layer;

[0026] The perception layer is used to store historical and real-time perception data;

[0027] The data layer is used to process and analyze the perception data;

[0028] The algorithm layer is used to call the algorithm to perform real-time online orbit deduction;

[0029] The output layer is used to output the full-time and space state of the train operation deduction in the line and the full-time and space state of the passenger travel process deduction;

[0030] The application layer is used to obtain multi-dimensional passenger flow analysis application results.

[0031] The beneficial effects of the present invention are as follows: The present invention provides a method for constructing a platform for line train operation simulation and passenger travel whole process simulation based on big data and fast calculation for large-scale rail network. Based on the spatial information of "line network-line-station" of rail transit, the invention considers the integrated analysis of rail transit multi-source historical and real-time data passenger card swiping data, real-time station entry and exit passenger flow data, real-time train weighing and location information data, video AI, train operation plan, etc., deeply explores the passenger travel pattern, and develops a platform that can accurately restore and deduce rail transit in real time through accurate simulation of individual passengers and individual trains. A real-time simulation system for digital twins of railway lines, which measures passenger flow and direction. This system uses network modeling, passenger flow and train simulation, dynamic passenger-train traffic allocation, passenger-train interactive operation simulation, dynamic line-passenger flow OD traffic verification, and distributed simulation technology to simulate the entire process of individual passengers entering the station, escalating, waiting on the platform, boarding and exiting the station. Through methods such as spatiotemporal matching of passengers and trains and cross-sectional passenger flow matching, the system dynamically verifies passenger travel spatiotemporal paths, especially transfer paths, ultimately achieving accurate restoration and prediction of passenger flow and direction across the entire railway network.

[0032] The beneficial effects brought about by the technology of the present invention also include the following:

[0033] (1) In response to the pressure of rail network operation, by accurately restoring and predicting the passenger flow / flow direction of the entire line network and combining the temporal and spatial distribution of passenger flow, it can support real-time monitoring of rail network passenger flow, and efficiently and reasonably allocate line network resources based on passenger flow monitoring, thus achieving efficient use of human, transportation and material resources;

[0034] (2) Based on the dynamic monitoring and deduction of the passenger flow of the whole network, the time and space distribution law of the passenger flow is accurately grasped, the existing scheduling scheme is evaluated and analyzed for the matching degree of the operation capacity and the traffic volume, the congestion sections and stations are identified, the operation cost and service quality are weighed, the optimization suggestions of the train operation scheme are proposed according to the passenger flow data, the accurate scheduling is realized, and the operation efficiency of the whole network is improved;

[0035] (3) For the evacuation of the station, the pedestrian simulation technology is applied, the scene reproduction of the pedestrian flow line of the station is realized, the evacuation bottleneck points are finely captured, and the optimization suggestions are provided, for the large passenger flow of the station, the passenger flow pressure test is carried out on the key stations through the micro-simulation, the service capacity of the station is obtained, the passenger flow control suggestions are provided based on the implementation of the passenger flow monitoring results, and the passenger flow of the station is controlled;

[0036] (4) Based on the network deduction and intelligent monitoring analysis, the subway whole trip chain information service can be provided for passengers, and the individual optimization and the system optimization are considered. Through the real-time deduction of the passenger flow, the multi-target path navigation service is provided, the congestion degree of the road network, the queuing length inside and outside the station and other information are accurately sent, and the "door-to-door" whole chain multi-mode trip planning service is provided for passengers. BRIEF DESCRIPTION OF DRAWINGS

[0037] The drawings described herein are used to provide further understanding of the present application, and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application, and do not constitute an improper limitation on the present application. In the drawings:

[0038] Figure 1 It is a flowchart of a kind of track digital twin real-time online deduction method;

[0039] Figure 2 It is a flowchart of an embodiment of a kind of track digital twin real-time online deduction method;

[0040] Figure 3 It is a structural schematic diagram of track network model;

[0041] Figure 4 It is a flowchart of passenger-train dynamic traffic distribution;

[0042] Figure 5 It is a flowchart of passenger-train interactive operation deduction;

[0043] Figure 6 It is a flowchart of dynamic traffic checking;

[0044] Figure 7 It is a structural schematic diagram of a kind of track digital twin real-time online deduction system;

[0045] Figure 8An embodiment structure schematic diagram of a rail digital twin real-time online deduction system;

[0046] Reference signs: 1. perception layer; 2. data layer; 3. algorithm layer; 4. output layer; 5. application layer. DETAILED DESCRIPTION

[0047] In order to make the technical solutions and advantages in the embodiments of the present application clearer, the exemplary embodiments of the present application are further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, and not an exhaustive enumeration of all embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0048] Embodiment 1: Reference Figures 1-8 In this embodiment, a rail digital twin real-time online deduction method is described in detail, which specifically includes the following steps:

[0049] S1. Construct a rail road network model containing a physical network and a space-time network;

[0050] S2. Based on the rail road network model, generate road network passenger flow data and line schedule data according to historical passenger flow OD data, real-time card swiping data and planned departure schedule information, and then generate simulation individuals of trains and passengers respectively, and perform path search on each simulation individual;

[0051] S3. Based on the line schedule data and the road network passenger flow data, perform time and space-based path trajectory matching on the train simulation individuals and passenger simulation individuals after path search, and complete dynamic traffic assignment of passengers-trains;

[0052] S4. Simulate passengers and vehicles according to departure time, path and travel speed, consider the deduction of the whole process of passengers entering station hall, gate, walking in station, building escalator and getting on and off the train, and perform fine simulation of the whole process of trains from starting station, station stop and terminal station, and complete passenger-train interactive operation deduction;

[0053] S5. According to the real-time detection data, the simulation results obtained after the dynamic traffic assignment of passengers-trains and the passenger-train interactive operation deduction, adjust the dynamic line schedule and passenger flow until the preset convergence condition is reached, and complete dynamic traffic checking;

[0054] S6. Based on the adjustment after dynamic traffic checking, use distributed simulation technology for simulation to realize accurate restoration and prediction of passenger flow volume / direction of the whole rail network;

[0055] Specifically, the present invention aims to provide a method for constructing a platform for line train operation simulation and passenger travel process simulation based on big data and rapid computing for large-scale rail networks. In the field of transportation, OD is the abbreviation of Origin-Destination, representing the departure and destination of a trip, and ODME is the abbreviation of Origin-Destination MartrixEstimation, which means dynamic OD estimation.

[0056] The present invention achieves the following technical effects: (1) Multi-dimensional rail passenger flow analysis based on multi-source big data fusion is carried out, so that the present invention can provide multi-dimensional passenger flow analysis such as entry and exit, station OD, section, line, line network, transfer, etc. under multiple time dimensions such as minutes, hours, days, months, years, characteristic days, holidays, etc., so as to meet the requirements of different business scenarios; (2) Accurate restoration of individual rail behaviors is carried out, providing individual-level and second-level fine simulation. For ultra-large-scale rail network simulation, the minimum statistical unit can be accurate to 30 seconds, with full coverage of trains and passengers in the entire network, and can accurately restore the individual travel chain of a single passenger and the driving and boarding and alighting conditions of a single train; (3) Individual-line dynamic passenger flow allocation is carried out according to the characteristics of rail transportation, so that individual passengers can swipe their cards when entering the station. The whole process of taking the train, escalator, waiting at the platform, getting on and off the train, and swiping the card at the exit is simulated, and the passenger travel time and space paths are dynamically checked by means of time and space matching between passengers and trains, cross-section passenger flow matching, etc., especially the accurate matching of transfer paths; (4) The present invention provides a rich set of scalable data interfaces, i.e., multi-source, open, and scalable data access and result output interfaces, which are convenient for model verification and update, result export, and visual display. When data and computing conditions permit, the expansion of multi-mode real-time online simulation can be carried out, and the mutual conversion of data and files of simulation software such as TransModeler, TransCAD, Vissim, and SUMO with the system provided by the present invention is supported, including data of road network, OD matrix, path, vehicle, and signal.

[0057] Furthermore, in S1, the construction of the physical network is the same as the construction of the traditional road network, and the subway line network is constructed based on the existing line information and station information; for the construction of the space-time network, the "boarding" and "getting off" behaviors of passengers and vehicles are taken into consideration, and the network is constructed based on the interaction between people and vehicles.

[0058] Furthermore, in S2, based on the historical passenger flow OD data, refined road network passenger flow data for different characteristic days (weekdays, weekends, holidays, etc.) and set time periods (5min / 15min) are generated. Combined with the real-time station entry and exit passenger flow data, i.e., real-time card swiping data, the road network passenger flow data is preliminarily verified and checked, and then passenger simulation individuals are generated; based on the planned departure timetable, line schedule data is generated, and then train simulation individuals are generated;

[0059] Reference Figure 3 , V1 represents a train carriage, P1 represents a first passenger, P2 represents a second passenger, P3 represents a third passenger, P4 represents a fourth passenger, P5 represents a fifth passenger, P6 represents a sixth passenger, A-E represent virtual platforms, represent real platforms;

[0060] Further, in S3, for a given historical passenger flow OD data, the departure station and the destination station, i.e., O point and D point, the earliest departure time EDT and the latest arrival time LAT of the simulation individual are obtained through real-time card swiping data;

[0061] According to the O point and the D point, a collection N of possible travel routes of the simulation individual is obtained, the individual path in the collection N is analyzed, all route service paths are traversed through spatial matching, and the stay time window of the individual path and the route service path at the O point and the D point is analyzed, whether the two paths can meet the time requirement is observed through time matching, the route service with the smallest time error is selected as the final result of the passenger-train, i.e., the individual path-route service path matching, all historical passenger flow OD data is traversed, and the individual path-route service path matching is completed, so as to realize the dynamic traffic distribution of the passenger-train.

[0062] Further, in S4, the simulation process of the passenger includes the following behaviors: simulating the entrance of the station based on the passenger entering the road network time and swiping the card, simulating the walking to the platform based on the flow density speed relationship and calculating the walking time, simulating the platform queuing for the train based on the queuing model, simulating the boarding and alighting behaviors based on the carriage storage capacity, simulating the walking to the gate based on the flow density speed relationship and calculating the walking time, and simulating the exit of the station based on the passenger leaving the road network time and swiping the card;

[0063] The simulation process of the vehicle includes the following behaviors: simulating the departure of the vehicle based on the planned route service and the vehicle entering the road network time, simulating the running according to the route based on the queuing model, simulating the arrival at the station, simulating the departure from the station, and simulating the arrival at the route terminal based on the vehicle leaving the road network time.

[0064] Further, in S5, for dynamic route service adjustment, the route service information of the vehicle is dynamically adjusted according to the arrival and departure data of the real-time route service in the real-time card swiping data, for dynamic passenger flow OD estimation, the OD total amount and the path are dynamically adjusted based on the given OD path, and according to the obtained adjusted route service and adjusted passenger flow OD, the passenger-train dynamic traffic distribution in step S4 and the passenger-train dynamic running deduction in step S5 are re-performed to obtain the simulation result, until the simulation accuracy or the simulation iteration number requirement is met, and the whole dynamic traffic checking process is completed;

[0065] Specifically, the deviation of the simulation passenger flow data from the actual passenger flow data on the OD path is determined and the deviation of the simulation train data from the actual train data on the OD path is determined The overall deviation value of the passenger flow on the OD path is calculated ;

[0066] The overall deviation value of the passenger flow on the OD path is represented as

[0067] ;

[0068] wherein, is the deviation of the simulation passenger flow data from the actual passenger flow data on the OD path is the weight parameter of the adjustment factor of the deviation of the simulation passenger flow data from the actual passenger flow data on the OD path, is the weight parameter of the adjustment factor of the deviation of the simulation train data from the actual train data on the OD path , and the weight parameter can be dynamically calibrated according to the confidence of the adjustment factor to meet different confidence requirements;

[0069] The actual flow on each path is adjusted according to the overall deviation value of the passenger flow on the different OD paths, and the passenger-train dynamic traffic distribution and passenger-train interactive operation deduction are performed again to calculate the simulation accuracy ;

[0070] The simulation accuracy is represented as

[0071] ;

[0072] wherein, is the traffic simulation value, is the actual observation value, and L is the number of all detector line segments;

[0073] If the total number n of simulation iterations reaches the preset total number of iterations, or the simulation accuracy has met the preset accuracy requirement, the entire dynamic adjustment process is completed, and the adjusted dynamic OD matrix is output; if not, the iteration number is increased by 1, and the step S3 is returned to continue adjusting the passenger flow on the OD path.

[0074] Further, in the S6, by using the distributed simulation technology, the internal transmission process of each road segment and the transfer process of each node in the simulation process are independently processed, the transmission step of all road segment sets in the road segment and the transfer step of all node sets in the node are distributedly processed in the branch network, and thus the engine running speed and efficiency are greatly improved.

[0075] Embodiment 2: refer to Figure 7 andFigure 8 In detail, the embodiment is a rail digital twin real-time online deduction system for implementing the rail digital twin real-time online deduction method in embodiment 1, which includes a perception layer 1, a data layer 2, an algorithm layer 3, an output layer 4 and an application layer 5;

[0076] The perception layer 1 is connected with the data layer 2 through a communication network and cloud computing, the data layer 2 is connected with the algorithm layer 3 through a communication network and cloud computing, the algorithm layer 3 is connected with the output layer 4 through a comprehensive transportation big data support platform, and the output layer 4 is connected with the application layer 5 through a communication network;

[0077] The perception layer 1 is used for storing historical and real-time perception data, mainly including card swiping data, train operation data and video data, supplemented by security check data, Internet data, GPS data and planning data;

[0078] The data layer 2 is used for processing and analyzing the perception data to obtain a road network basic database, a line network operation database, a historical / real-time card swiping database, a station / vehicle passenger flow database, a real-time database and a simulation database;

[0079] The algorithm layer 3 is used for calling algorithms to perform rail real-time online deduction, and the algorithm layer is the core of the system, which mainly includes a road network construction algorithm, a road network passenger flow generation algorithm, a line train generation algorithm, a passenger-train dynamic traffic distribution algorithm, a passenger-train dynamic traffic simulation algorithm and a dynamic schedule-passenger flow traffic checking algorithm;

[0080] The output layer 4 is used for outputting a train operation deduction full-time and space state and a passenger travel process deduction full-time and space state in a line;

[0081] The application layer 5 is used for obtaining multi-dimensional passenger flow analysis application results, that is, based on the data output by the output layer 4, obtaining in-station, out-station, station OD, section, line, line network and transfer in multiple time dimensions such as minute, hour, day, month, year, feature day and holiday;

[0082] Specifically, the embodiment is based on the spatial information of the “line network-line-station” of rail transit, considers the fusion analysis of the multi-source historical and real-time data of rail transit, such as passenger card swiping data, station real-time in-out passenger flow data, train real-time weighing and position information data, video AI and train operation plan, deeply mines passenger travel rules, accurately simulates passenger individuals and train individuals, and proposes a rail digital twin real-time online deduction system capable of accurately restoring and deducing rail line network passenger flow and direction in real time. Meanwhile, the rail digital twin real-time online deduction system can also consider information security technology, standard specification components, copyright information, data interface and other issues for independent design.

[0083] While the application has been described in accordance with a limited number of embodiments, these are merely illustrative of the many possible embodiments of the application. Other embodiments can be devised without departing from the scope of the application as described herein. Additionally, it is intended that the description set forth herein should not be construed as limiting but merely as illustrative of the presently preferred embodiments of the application. Many modifications and variations to the embodiments described herein will be apparent to those of ordinary skill in the art, and it is intended that the application encompass all such modifications and variations as fall within the scope of the appended claims. Accordingly, the application is not to be limited by the specific examples described herein, but only by the scope of the appended claims.

Claims

1. A track digital twin real-time online inference method, characterized in that, The method comprises the following steps: S1. Constructing a rail road network model comprising a physical network and a space-time network; S2. Based on the rail road network model, generating road network passenger flow data and line schedule data according to historical passenger flow OD data, real-time card swiping data and planned departure schedule information, and then generating simulation individuals of trains and passengers respectively, and performing path search on each simulation individual; S3. Based on the line schedule data and the road network passenger flow data, performing time and space-based path trajectory matching on the train simulation individuals and the passenger simulation individuals after path search, and completing dynamic traffic assignment of passengers-trains; S4. Simulating passengers and vehicles according to departure time, path and travel speed, and completing passenger-train interactive operation deduction; S5. According to real-time detection data, the simulation results obtained after the dynamic traffic assignment of passengers-trains and the passenger-train interactive operation deduction, adjusting the dynamic line schedule and passenger flow until the preset convergence condition is reached, and completing dynamic traffic checking; S6. Based on the adjustment after the dynamic traffic checking, performing simulation by using distributed simulation technology, and realizing accurate restoration and prediction of passenger flow volume and direction of the whole line network; In S3, for given historical passenger flow OD data, the departure station and the terminal station, i.e. O point and D point, the earliest departure time EDT and the latest arrival time LAT of the simulation individual are obtained through real-time card swiping data; According to the O point and the D point, a collection N of possible travel lines of the simulation individual is obtained, the individual path in the collection N is analyzed, all line schedule paths are traversed through space matching, and the stay time window of the individual path and the line schedule path at the O point and the D point is analyzed, whether the two paths can meet the time requirement is observed through time matching, the line schedule with the smallest time error is selected as the final result of the individual path-line schedule path matching, all historical passenger flow OD data are traversed, the individual path-line schedule path matching is completed, and thus the dynamic traffic assignment of passengers-trains is realized; In S5, for dynamic line schedule adjustment, the arrival and departure data of real-time line schedule in real-time card swiping data are used to dynamically adjust the line schedule information of the vehicle, for dynamic passenger flow OD estimation, the OD total amount and path are dynamically adjusted based on the given OD path, the simulation results are obtained by re-performing the passenger-train dynamic traffic assignment in S4 and the passenger-train dynamic operation deduction in S5 according to the adjusted line schedule and the adjusted passenger flow OD until the simulation accuracy or the simulation iteration number requirement is met, and the whole dynamic traffic checking process is completed In S6, by using the distributed simulation technology, the transmission process in each road segment and the transfer process at each node in the simulation process are independently processed, and the transmission steps of all road segment sets in the road segment and the transfer steps of all node sets in the node are independently processed.

2. The track digital twin real-time online inference method according to claim 1, wherein, In the S1, for the construction of the physical network, the subway line network is constructed based on the existing line information and station information; for the construction of the space-time network, the "boarding" and "alighting" behaviors of passengers and vehicles are considered, and the space-time network is constructed according to the interaction between people and vehicles.

3. The track digital twin real-time online inference method according to claim 2, wherein, In the S2, the road network passenger flow data under different characteristic days and set time periods is generated according to historical passenger flow OD data, the road network passenger flow data is preliminarily verified and checked in combination with real-time station entry and exit passenger flow data, that is, real-time card swiping data, and passenger simulation individuals are further generated; According to the planned departure timetable, line trip data is generated, and train simulation individuals are further generated.

4. The track digital twin real-time online inference method according to claim 3, wherein, In the S4, the simulation process of passengers includes the following behaviors: simulating the gate card swiping into the station based on the time when the passenger enters the road network, simulating the walking to the platform based on the flow density speed relationship, simulating the platform queuing for the train based on the queuing model, simulating the boarding and alighting behaviors based on the car storage capacity, simulating the walking to the gate based on the flow density speed relationship, and simulating the gate card swiping out of the station based on the time when the passenger leaves the road network; The simulation process of vehicles includes the following behaviors: simulating the vehicle departure based on the planned line trip and the time when the vehicle enters the road network, simulating the running according to the line based on the queuing model, simulating the arrival at the station, simulating the departure from the station, and simulating the arrival at the line terminal based on the time when the vehicle leaves the road network.

5. A rail digital twin real-time online inference system, characterized in that, A rail digital twin real-time online deduction method for realizing any one of claims 1-4, comprising a perception layer (1), a data layer (2), an algorithm layer (3), an output layer (4), and an application layer (5); The perception layer (1) is connected with the data layer (2), the data layer (2) is connected with the algorithm layer (3), the algorithm layer (3) is connected with the output layer (4), and the output layer (4) is connected with the application layer (5); The perception layer (1) is used for storing historical and real-time perception data; The data layer (2) is used for processing and analyzing the perception data; The algorithm layer (3) is used for calling algorithms to perform rail real-time online deduction; The output layer (4) is used for outputting the train running deduction full space-time state in the line and the passenger travel process deduction full space-time state; The application layer (5) is used for obtaining multi-dimensional passenger flow analysis application results.

Images