Virtual section-oriented rail transit path planning method and device and readable medium

By modeling the train routing problem in the rail transit network as multi-agent path planning problem, using virtual sections and conflict-based search algorithms, the safety and efficiency problems in train path planning are solved, and high precision and Efficient path planning is suitable for practical application scenarios in railway networks.

CN118940923APending Publication Date: 2024-11-12HUAQIAO UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410957618.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing train routing problem (TRP) is difficult to effectively solve train path planning in the railway network, especially due to the limitations of safe operation and the scale limitations of the rail-side train detection (TTD) section, resulting in system complexity and Maintenance requirements have increased, and multi-agent path planning (MAPF) has discrete time steps problems in actual scenarios.

Method used

A rail transit path planning method for virtual sections is proposed. By modeling the rail transit network, the TTD segment is divided into virtual sections, abstracting each train as an agent, using a conflict-based search algorithm (CBS) With the A* search algorithm, generates conflict-free path planning, resolves mobile conflicts, rear-end collisions and TTD conflicts, and improves the efficiency and security of path planning.

Benefits of technology

It realizes high-precision train path planning, meets safety and efficiency requirements, can effectively avoid collisions and other dangerous situations, improves system operation efficiency and stability, has a wider scope of application, and adapts to path planning needs in different complex scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940923A_ABST
    Figure CN118940923A_ABST
Patent Text Reader

Abstract

The invention discloses a virtual section-oriented rail transit path planning method and device and a readable medium, and aims to efficiently generate a conflict-free path through a two-stage structure of a global planning layer and a local planning layer by adopting a conflict-based search algorithm. According to the method, the TTD section is subdivided into smaller fragments by using a VSS virtual section division method, so that the calculation amount is reduced, and the path planning efficiency is improved. Factors such as safety interval and path continuity are considered in the path planning process, and safe operation of rail transit is reliably guaranteed. Through real-time adjustment and evaluation of the path, potential dangers can be effectively avoided, and the safety of vehicle operation is ensured. The idea of path planning based on multiple agents is adopted, and the method aims at efficiently solving the train running path planning problem in a rail transit system and being close to application in a more practical scene.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a field, and in particular to a rail transit path planning method, device and readable medium for virtual sections. Background Art

[0002] The Train Routing Problem (TRP) is a network optimization problem that aims to determine the best path for trains in a railway network to improve transportation efficiency and reduce costs. The efficiency of railway infrastructure can be improved by increasing railway throughput. This usually means increasing the number of trains, but the requirements for safe operation limit only one train to occupy a specific section at a time to avoid collisions.

[0003] Trackside Train Detection (TTD) technology is used to monitor the occupancy of a section and ensure that a section can only be entered by another train after a train has completely left. Although reducing the size of a TTD section can increase transportation capacity, it requires the installation of new TTD equipment, increasing the complexity and maintenance requirements of the system.

[0004] Previous studies have used multi-agent pathfinding (MAPF) to solve TRP, but its assumptions limit its application in practical scenarios and there is a discrete time step problem in MAPF. Summary of the invention

[0005] The purpose of this application is to propose a rail transit path planning method, device and readable medium for virtual sections in response to the above-mentioned technical problems.

[0006] In a first aspect, the present invention provides a rail transit path planning method for a virtual section, comprising the following steps:

[0007] The rail transit network is modeled on a map to obtain a rail transit network graph, which includes a number of nodes and edges between two nodes. The edge e represents a TTD segment. Each TTD segment is divided into a number of VSS virtual segments. Each train running on the rail transit network is modeled as an intelligent agent, and the position of the intelligent agent in the current state is defined as Pos = (e, a, b), where a and b represent the positions of the intelligent agent in the range of a to b in the TTD segment, and 0≤a≤b≤l(e), where l(e) is the length of the TTD segment.

[0008] Obtain the train's travel mission, perform independent path planning for each agent based on the rail transit network diagram and the train's travel mission, obtain the planned path of each agent, create a CT node based on the planned path of each agent and add it to the open list. Each CT node stores the planned path of an agent, the cost and constraints of the agent's planned path;

[0009] Select the CT node with the minimum cost of the agent's planned path in the open list, and determine whether there is a conflict in the planned path corresponding to the CT node with the minimum cost of the agent's planned path. The conflict includes movement conflict and TTD conflict. Movement conflict includes frontal collision conflict, overlapping collision conflict and rear-end collision conflict. If so, select a conflict, create a new CT node with corresponding constraints for each agent in the conflict and add it to the open list. The constraint is (tri, si, ti), which means that at time step ti, agent tri is not allowed to expand to state si. When the conflict is a rear-end collision conflict or a TDD conflict, si is extended to the VSS virtual section where the train rear-end collision conflict or TDD conflict occurs, and the independent path planning of the intelligent agent tri is re-performed and the constraints are complied with. This step is repeated until the planned path corresponding to the CT node with the minimum cost of the planned path of the intelligent agent has no conflict. The planned path corresponding to the CT node with the minimum cost and no conflict of the planned path of the intelligent agent is taken as the optimal planned path of the corresponding train.

[0010] As a preferred method, the frontal collision conflict and the overlapping collision conflict are resolved as follows:

[0011] Assume that agent trp expands to the next state sp at time step tp, and agent trq expands to the next state sq at time step tq. If agent trp and agent trq generate a head-on collision conflict or an overlapping collision conflict during the expansion process, generate constraints for agent trp or agent trq:

[0012] If the constraint (trp, sp, tp) is generated for the agent trp, (trp, sp, tp) means that at the time step tp, trp is not allowed to expand to the state sp, and the independent path planning of the agent trp is re-performed and complies with the constraint;

[0013] If the constraint (trq, sq, tq) is generated for the agent trq, (trq, sq, tq) means that at the time step tq, trq is not allowed to expand to the state sq, and the independent path planning of the agent trq is re-performed and complies with the constraint.

[0014] As a preferred solution, the rear-end collision conflict and TDD conflict are resolved as follows:

[0015] Assume that agent trx expands to the next state sx at time step tx, and agent try expands to the next state sy at time step ty, and the position of agent trx in the current state is ahead of the position of agent try. If agent trx and agent try generate rear-end collision conflict and TDD conflict during the expansion process, generate constraints for agent trx or agent try:

[0016] If constraints (try, sy, ty) are generated for the agent try, (try, sy, ty) means that at the time step ty, try is not allowed to expand to state sy, sy is the VSS virtual section where a train rear-end collision conflict or TDD conflict occurs, and the independent path planning of the agent try is re-performed and the constraints are observed.

[0017] As a preference, the steps of independent path planning are as follows:

[0018] Use the A* search algorithm to find a single planning path for each agent, ignoring all other agents; set the starting state Add to the open list, Start is the starting position of the train, and the g value of the starting state is set to 0;

[0019] Select the state s with the minimum cost from the open list to expand, if state s is the target state Dest is the destination of the train, Ttr is the time step for the train to reach the destination, then the search is successful, and the path from the starting state to the target state is used as the planning path of the agent, otherwise the state s is removed from the open list and added to the closed list;

[0020] For every successor state s of state s next , if the successor state s next In the closed list, it is ignored;

[0021] If the successor state s next If it is not in the open list, the successor state s next Add to the open list and calculate the successor state s next The path cost is g(s next )=g(s)+t', g(s=t is the path cost from the starting state sinit to state s, t represents the train from the starting state s init The time step to run to state s, t' represents the train moving from state s to the successor state s next time, calculate the successor state s next The target cost h(s next ): z is the distance the train moves from state s to the target state, Str is the running speed of the train, and the successor state s is calculated. next The cost F(s next ): F(s next )=g(s next )+h(s next );

[0022] If the successor state s next In the open list, the computation runs from state s to the successor state s next The path cost g(s next ): g(s next )=g(s)+t'; if it runs from state s to the successor state s next The path cost is less than running from any state other than state s to the successor state s next The path cost is , then the successor state s next The path cost is updated to g(s next )=g(s)+t', update the calculation of the successor state s next The target cost h(s next ) and cost F(s next ), and update the successor state s next The parent node of is in state s, otherwise it is not updated;

[0023] Repeat the above steps until the state with the minimum cost is the target state or the open list is empty. If the state with the minimum cost is the target state, backtrack along the parent node of each state until the starting state is reached, and construct the planning path for each agent.

[0024] Preferably, the length of the VSS virtual section is at least the sum of the train length and the safety distance. The maximum number of VSS virtual sections required is r=l(e) / (ltr+d), where ltr represents the train length and d represents the safety distance.

[0025] Preferably, whether there is a conflict is determined by:

[0026] Assume that the states of the agent Trp are:

[0027] Assume that the states of the agent Trq are:

[0028] By looping through the time, it is determined whether the positions of the agent Trp and the agent Trq overlap. If so, it is determined that there is a conflict, otherwise it is determined that there is no conflict.

[0029] In a second aspect, the present invention provides a rail transit path planning device for a virtual section, comprising:

[0030] A modeling module is configured to map the rail transit network to obtain a rail transit network graph, wherein the rail transit network graph includes a plurality of nodes and edges between two nodes, wherein an edge e represents a TTD segment, each TTD segment is divided into a plurality of VSS virtual segments, each train running on the rail transit network is modeled as an intelligent agent, and a position of the intelligent agent in a current state is defined as Pos=(e, a, b), wherein a and b respectively represent positions of the intelligent agent in the range of a to b in the TTD segment, and 0≤a≤b≤l(e), wherein l(e) is the length of the TTD segment;

[0031] The independent path planning module is configured to obtain the driving task of the train, perform independent path planning for each agent according to the rail transit network diagram and the driving task of the train, obtain the planned path of each agent, create a CT node according to the planned path of each agent and add it to the open list, and each CT node stores the planned path of an agent, the cost and constraints of the planned path of the agent;

[0032] The conflict resolution module is configured to select the CT node with the minimum cost of the agent's planned path in the open list, and determine whether there is a conflict in the planned path corresponding to the CT node with the minimum cost of the agent's planned path. The conflict includes movement conflict and TTD conflict. The movement conflict includes frontal collision conflict, overlapping collision conflict and rear-end collision conflict. If so, a conflict is selected, and a new CT node with corresponding constraints is created for each agent in the conflict and added to the open list. The constraints are (tri, si, ti), which means that at the time step ti, the agent tri is not allowed to expand to the state si.

[0033] When the conflict is a rear-end collision conflict or a TDD conflict, si is extended to the VSS virtual section where the train rear-end collision conflict or TDD conflict occurs, and the independent path planning of the intelligent agent tri is re-performed and the constraints are complied with. This step is repeated until the planned path corresponding to the CT node with the minimum cost of the planned path of the intelligent agent has no conflict. The planned path corresponding to the CT node with the minimum cost and no conflict of the planned path of the intelligent agent is taken as the optimal planned path of the corresponding train.

[0034] In a third aspect, the present invention provides an electronic device comprising one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation manner in the first aspect.

[0035] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any implementation manner in the first aspect.

[0036] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method described in any implementation manner in the first aspect.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] (1) The virtual-segment rail transit path planning method proposed in the present invention models the train routing problem in the rail network as an instance of a multi-agent path planning problem. Each train is abstracted into an agent through agent modeling, and the optimal path is effectively determined through a conflict-based search algorithm. A high-precision solution is generated to meet safety and efficiency requirements. That is, the path planning algorithm must ensure a safe interval between trains to avoid collisions, rear-end collisions and other dangerous situations, and should be able to calculate the shortest path so that passengers can reach their destination as quickly as possible, thereby improving the operating efficiency of the system.

[0039] (2) The virtual segment-oriented rail transit path planning method proposed in the present invention introduces the VSS virtual segment, which can divide the rail system without hardware and is easy to adjust the layout. At the same time, it can solve the discrete time step problem in MAPF and has stronger conflict resolution capabilities. It can more effectively resolve conflicts between vehicles in path planning. Through continuous conflict detection and path optimization, potential conflicts can be avoided and resolved in a timely and accurate manner, improving the operational stability and reliability of the system.

[0040] (3) The virtual segment-oriented rail transit path planning method proposed in the present invention adopts a conflict-based search algorithm through a two-level structure of a global planning layer and a local planning layer to efficiently generate a conflict-free path. The VSS virtual segment division method is used to subdivide the TTD segment into smaller segments, reducing the amount of calculation, thereby improving the efficiency of path planning. In the path planning process, factors such as safety intervals and path continuity are taken into account to reliably ensure the safe operation of rail transit. Through real-time adjustment and evaluation of the path, potential dangers can be effectively avoided and the safety of vehicle operation can be ensured.

[0041] (3) The virtual section rail transit path planning method proposed in the present invention combines the advantages of different algorithms and is closer to actual application scenarios. Due to its high flexibility, strong adjustability, wider application range, and ability to meet the path planning needs in different complex scenarios, it has a wider practical application prospect. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 A schematic diagram of a flow chart of a rail transit path planning method for a virtual section according to an embodiment of the present application;

[0044] Figure 2 A schematic diagram of a rail transit network map modeling and a train intelligent agent modeling of a rail transit path planning method for a virtual section according to an embodiment of the present application;

[0045] Figure 3 A flowchart of independent path planning of a rail transit path planning method for a virtual section according to an embodiment of the present application;

[0046] Figure 4 A schematic diagram of possible conflicts caused by train operation in a rail transit path planning method for a virtual section according to an embodiment of the present application;

[0047] Figure 5 A flow chart of path planning for all trains in a rail transit path planning method for a virtual section according to an embodiment of the present application;

[0048] Figure 6 A schematic diagram of a rail transit path planning device for a virtual section according to an embodiment of the present application;

[0049] Figure 7 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Figure 1 A virtual section-oriented rail transit path planning method provided by an embodiment of the present application is shown, comprising the following steps:

[0052] S1. Carry out map modeling on the rail transit network to obtain a rail transit network graph, which includes several nodes and edges between two nodes. Edge e represents a TTD segment. Each TTD segment is divided into several VSS virtual segments. Each train running on the rail transit network is modeled as an intelligent agent, and the position of the intelligent agent in the current state is defined as Pos=(e, a, b), where a and b respectively represent the position of the intelligent agent in the range of a to b in the TTD segment, and 0≤a≤b≤l(e), where l(e) is the length of the TTD segment.

[0053] In a specific embodiment, the length of the VSS virtual segment is at least the sum of the train length and the safety distance. The maximum number of VSS virtual segments required is r=l(e) / (ltr+d), where ltr represents the train length and d represents the safety distance.

[0054] Specifically, refer to Figure 2 ,In the process of map modeling, node representation and edge representation are required. Node representation is to abstract the key locations of the rail transit network, such as turnouts, stations, etc., into nodes of the rail transit network graph. Each node represents a specific location. There may be multiple nodes on a long track. Figure 2 In the figure, 2 and 3 are turnout nodes, and 0, 1, 4, 5, and 6 are nodes on the track. Edge representation is the use of edges in the rail transit network graph to represent the connection relationship between two nodes, usually representing that a train can reach one location from another in a certain period of time. Edge e describes the track section connected by vertices. Vertices represent trackside train detection (TTD) hardware. Each edge has a unique identifier, such as Figure 2 In the embodiment of the present application, edge e is also referred to as a TTD segment. Each TTD segment adopts the VSS virtual segment division method to improve the efficiency and safety of rail transit path planning. The maximum number r of VSS virtual segments is introduced to divide a TTD segment into r VSS virtual segments. The VSS virtual segment is composed of several fragments, so that a TTD segment can no longer generate any multiple VSS virtual segments, thereby reducing the amount of calculation.

[0055] Taking into account the characteristics of the rail transit network, it is necessary to recognize some of the limitations, especially in the case of three-way turnouts, where not all directions can be freely passed. Figure 2In the rail transit network diagram shown, there are three edges 1-2, 2-3, and 2-5 at node 2, but in the actual train operation scenario, path 3-2-5 may be a prohibited section. Similarly, at the node 3, the fork formed by 2-3, 3-4, and 3-6, path 2-3-6 may also be impassable.

[0056] These restricted sections need to be properly considered during route planning to ensure that the generated route planning scheme is consistent with the actual situation. During the route planning process, these inaccessible sections will be excluded from the rail transit network diagram so that the actual operating restrictions can be followed during route planning to generate a more reasonable and effective route. Through such optimization and consideration, it can be ensured that the planned route not only meets the structural characteristics of the network, but can also be successfully applied in actual train operations.

[0057] After the map modeling phase is completed, the next step is to model the train and adapt it to the rail transit network diagram. In order to accurately describe the position of the train in the rail transit network, the embodiment of the present application uses a triple TTD segment to represent the position of the train, specifically: POS = (e, a, b). Among them, e corresponds to a TTD segment in the rail transit network diagram, a and b respectively indicate that the position of the train occupies the range of a to b in the TTD segment e, and 0≤a≤b≤l(e), where l(e) is the length of the TTD segment. The travel task of the train is described by the tuple Tr = (ltr, v, Start, Dest), where ltr represents the length of the train, v represents the maximum speed of the train, Start is the starting position of the train, and Dest is the destination position of the train. For example, in Figure 2 In the figure, the current position of train Tr1 is (e1, 300, 500). Similarly, the current position of train Tr2 is (e4, 600, 1000). This triple representation can accurately describe the position of the train on the track. By matching the position of the agent with the edge of the rail transit network graph, the train can be accurately located during the path planning process and better meet the needs of rail transit path planning. The strategy of adapting agent modeling to the rail transit network graph can ensure the accuracy and practicality of path planning, laying a solid foundation for the successful implementation of rail transit path planning methods.

[0058] The movement of the train is represented by m, that is, m contains a series of TTD segments to represent the position changes of the train.

[0059] Specifically, there are two types of train actions in the rail transit network diagram:

[0060] 1. Stop: The train maintains the same state in a section, that is, m = ((e, i, j)), for example Figure 2 In the example, train Tr1 stops at the current position, which is represented by m=((e1, 300, 500));

[0061] 2. Movement: The train moves from one location to the next. If it crosses multiple TTD segments, then m can be expressed as m = ((ei, a, b), (ei+1, c, d)). The length of the movement may also be less than the length of the corresponding train. Figure 2 In the example, train Tr1 moves from the current position to node 2, which can be expressed as m=((e1, 300, 500), (e2, 800, 1000)).

[0062] S2, obtain the train's driving mission, plan the path of each agent independently according to the rail transit network diagram and the train's driving mission, obtain the planned path of each agent, create a CT node according to the planned path of each agent and add it to the open list. Each CT node stores the planned path of an agent, the cost and constraints of the planned path of the agent.

[0063] In a specific embodiment, the steps of independent path planning are as follows:

[0064] Use the A* search algorithm to find a single planning path for each agent, ignoring all other agents; set the starting state Add to the open list, Start is the starting position of the train, and the g value of the starting state is set to 0;

[0065] Select the state s with the minimum cost from the open list to expand, if state s is the target state Dest is the destination of the train, Ttr is the time step for the train to reach the destination, then the search is successful, and the path from the starting state to the target state is used as the planning path of the agent, otherwise the state s is removed from the open list and added to the closed list;

[0066] For every successor state s of state s next , if the successor state s next In the closed list, it is ignored;

[0067] If the successor state s next If it is not in the open list, the successor state s next Add to the open list and calculate the successor state s next The path cost is g(s next )=g(s)+t', g(s=t is the path cost from the starting state sinit to state s, t represents the train from the starting state s init The time step to run to state s, t' represents the train moving from state s to the successor state snext time, calculate the successor state s next The target cost h(s next ): z is the distance the train moves from state s to the target state, Str is the running speed of the train, and the successor state s is calculated. next The cost F(s next ): F(s next )=g(s next )+h(s next );

[0068] If the successor state s next In the open list, the computation runs from state s to the successor state s next The path cost g(s next ): g(s next )=g(s)+t'; if it runs from state s to the successor state s next The path cost is less than running from any state other than state s to the successor state s next The path cost is , then the successor state s next The path cost is updated to g(s next )=g(s)+t', update the calculation of the successor state s next The target cost h(s next ) and cost F(s next ), and update the successor state s next The parent node of is in state s, otherwise it is not updated;

[0069] Repeat the above steps until the state with the minimum cost is the target state or the open list is empty. If the state with the minimum cost is the target state, backtrack along the parent node of each state until the starting state is reached, and construct the planning path for each agent.

[0070] Specifically, the embodiments of the present application adopt a conflict-based search algorithm (Conflict-Based Search, CBS) and a train path planning solution for VSS virtual segment division, so as to efficiently solve the conflict problem between trains. The conflict-based search algorithm (Conflict-Based Search, CBS) is an algorithm for multi-agent path planning, which has a two-level structure: a global planning layer and a local planning layer. The global planning layer is responsible for generating an overall path planning solution. The algorithm considers the starting position, destination position and constraints of all agents, and tries to find a set of conflict-free paths. If a conflict is found, the conflict information is passed to the local planning layer for resolution. The local planning layer is responsible for resolving conflicts. When the global planning layer finds a conflict, it will pass the conflict information to the local planning layer. Based on the conflict information, the local planning layer will try to adjust the path of the relevant agents to resolve the conflict.

[0071] Train independent path planning is done in the global planning layer of the CBS algorithm. At this stage, the CBS algorithm attempts to generate initial paths for all agents so that conflict detection and resolution can be performed in subsequent searches. In the embodiment of the present application, the A* search algorithm is used as the global planning layer of the CBS algorithm. Define the state of the train Where Ttr is the time step for the train to reach the destination, or t if the train has not yet reached the destination. Target state The path cost of the train is g(s)=t, which calculates the cost of the train running from the starting state to the current state. The target cost is Calculate the cost of the train running from the current state to the target state, where z is the distance from the current state to the target state, and Str is the running speed of the train. The cost is F(s) = g(s) + h(s).

[0072] refer to Figure 3 At the beginning of the algorithm, the rail transit network diagram and the train driving task are input, and the A* search algorithm is used to find a single path planning solution for each agent, ignoring all other agents. init Add to the open list OPEN, and change the initial state g(s init ) value is set to 0. Select the state s with the smallest F(s) value from the open list for expansion. If state s is the target state, the search is successful and the found path is returned; otherwise, remove state s from the open list and add it to the closed list CLOSE. For each successor state s expanded from state s next , if s next In the closed list, it is ignored. nextIf s is not in the open list, add it to the open list and calculate its g, h, and F values. next If it is already in the open list and the path to it through the current state s is better than the previous one, update its g value, h value, F value and parent node. Repeat the above steps until the target state is found or the open list is empty. If the target state is found, start from the target state and backtrack along the parent node of each state until the initial state is reached, and construct the optimal path as the planning path for each agent.

[0073] S3, select the CT node with the minimum cost of the agent's planned path in the open list, and determine whether there is a conflict in the planned path corresponding to the CT node with the minimum cost of the agent's planned path. The conflict includes movement conflict and TTD conflict. Movement conflict includes frontal collision conflict, overlapping collision conflict and rear-end collision conflict. If so, select a conflict, create a new CT node with corresponding constraints for each agent in the conflict and add it to the open list. The constraints are (tri, si, ti), which means that at time step ti, agent tri is not allowed to expand to state si. When the conflict is a rear-end collision conflict or a TDD conflict, si is extended to the VSS virtual section where the train rear-end collision conflict or TDD conflict occurs, and the independent path planning of the intelligent agent tri is re-performed and the constraints are complied with. This step is repeated until the planned path corresponding to the CT node with the minimum cost of the planned path of the intelligent agent has no conflict. The planned path corresponding to the CT node with the minimum cost and no conflict of the planned path of the intelligent agent is taken as the optimal planned path of the corresponding train.

[0074] In a specific embodiment, the frontal collision conflict and the overlapping collision conflict are resolved as follows:

[0075] Assume that agent trp expands to the next state sp at time step tp, and agent trq expands to the next state sq at time step tq. If agent trp and agent trq generate a head-on collision conflict or an overlapping collision conflict during the expansion process, generate constraints for agent trp or agent trq:

[0076] If the constraint (trp, sp, tp) is generated for the agent trp, (trp, sp, tp) means that at the time step tp, trp is not allowed to expand to the state sp, and the independent path planning of the agent trp is re-performed and complies with the constraint;

[0077] If the constraint (trq, sq, tq) is generated for the agent trq, (trq, sq, tq) means that at the time step tq, trq is not allowed to expand to the state sq, and the independent path planning of the agent trq is re-performed and complies with the constraint.

[0078] In a specific embodiment, the rear-end collision conflict and the TDD conflict are resolved as follows:

[0079] Assume that agent trx expands to the next state sx at time step tx, and agent try expands to the next state sy at time step ty, and the position of agent trx in the current state is ahead of the position of agent try. If agent trx and agent try generate rear-end collision conflict and TDD conflict during the expansion process, generate constraints for agent trx or agent try:

[0080] If constraints (try, sy, ty) are generated for the agent try, (try, sy, ty) means that at the time step ty, try is not allowed to expand to state sy, sy is the VSS virtual section where a train rear-end collision conflict or TDD conflict occurs, and the independent path planning of the agent try is re-performed and the constraints are observed.

[0081] In a specific embodiment, whether there is a conflict is determined by:

[0082] Assume that the states of the agent Trp are:

[0083] Assume that the states of the agent Trq are:

[0084] By looping through the time, it is determined whether the positions of the agent Trp and the agent Trq overlap. If so, it is determined that there is a conflict, otherwise it is determined that there is no conflict.

[0085] Specifically, the primary goal of the multi-agent path planning (MAPF) problem is to find the path planning for all trains without conflicts. For this purpose, the concept of conflict is introduced. Figure 4 In the embodiments of the present application, there are two common conflict types.

[0086] The first type is motion conflict. If two movements overlap, a motion conflict occurs. Specifically, motion conflicts can include the following three situations:

[0087] Figure 4 (a) There is a "head-on collision conflict". When the two trains in the figure try to go from TTD segment e1 to TTD segment e2 and from TTD segment e2 to TTD segment e1 respectively, they must pass through the same edge, resulting in a conflict. This conflict cannot be resolved by introducing a VSS virtual segment.

[0088] Figure 4(b) There is an "overlap collision conflict". An overlap collision conflict occurs when two trains need to pass through the same track intersection or turning point at the same time. In this case, the trains may block each other, resulting in operation obstruction or collision. This conflict needs to be resolved to ensure that the trains can pass through the intersection smoothly. Similar to the conflict in 4(a), the overlap collision conflict cannot be resolved by splitting the TTD segment into VSS virtual segments. The conflict can only be avoided when one of the trains is waiting.

[0089] Figure 4 (c) There is a "rear-end collision conflict". The two trains move from TTD segment e1 to TTD segment e2 respectively. This situation may lead to a rear-end collision accident during train operation, that is, one train catches up with another train and may cause a collision. This type of conflict can be resolved by VSS virtual segments. At the red arrow of TTD segment e1, TTD segment e1 is divided into two VSS virtual segments to prevent train Tr2 from entering the VSS virtual segment of train TR1, thereby achieving conflict avoidance.

[0090] The second type is TTD conflict, which occurs if two movements contain the same TTD segment and no movement conflict occurs.

[0091] Figure 4 (d) is an example of TTD conflict. After both trains have completed their respective movements, train Tr1 and train Tr2 are both located in TTD segment e1, but no collision occurs. TTD conflict can also be resolved using VSS virtual segments. The TTD segment e1 is divided into two VSS virtual segments, which belong to train Tr1 and train Tr2 respectively, ensuring that train Tr1 and train Tr2 do not occupy the same VSS virtual segment at any time, thereby resolving the TTD conflict.

[0092] Resolve the above conflicts at the local planning level. Figure 5 , the specific process is as follows:

[0093] At the beginning of the algorithm, the rail transit network diagram and the starting and target positions of each train are input, and the rail transit path planning problem is abstracted into a MAPF problem through map modeling and agent modeling. The A* search algorithm is used to find a single agent path plan for each agent, ignoring all other agents. At this stage, the optimal plan of each agent only includes movement actions. This set of single agent path plans is used to create CT nodes and add them to the open list. The CT node is a node in the conflict tree, which stores a possible solution, including a set of single agent path plans for agents, the cost of path planning, and constraint information. Conflicts are gradually resolved by creating and expanding these CT nodes until a solution without conflict is found or it cannot be expanded further. The open list is used to store CT nodes to be processed. Each CT node represents a possible solution, sorted according to the cost of the CT node.

[0094] Then, in each iteration, a CT node N is extracted from the open list OPEN, which represents the solution with the minimum cost relative to other CT nodes in the open list, that is, the F value is the smallest. If there is no conflict in CT node N, the planned path of the train is obtained. Otherwise, a conflict is selected. For each agent in the conflict, a new CT node is created, which contains the corresponding constraints (constraints refer to a set of restrictions or rules that limit the actions or movements that the agent can take when planning a path) and the new single-agent path plan of the agent. This new CT node is added to the open list so that it can be selected for expansion in future iterations.

[0095] The constraints contained in the new CT node are generated by resolving conflicts. The specific steps are as follows:

[0096] like Figure 4 As shown in (a), assuming that train Tr1 expands to the next state s1 at time step t1, and train Tr2 expands to the next state s2 at time step t2, a head-on collision occurs between the two. A constraint is generated for train Tr1 or train Tr2. If constraint (Tr1, s1, t1) is generated for train Tr1, the tree (Tr1, s1, t1) indicates that at time t1, train Tr1 is not allowed to expand to state s1, and independent path planning of train Tr1 is performed again and the constraint is observed.

[0097] Figure 4 (b) The conflicting constraints generated and Figure 4 (a) Similar.

[0098] Figure 4(c) After splitting the VSS virtual segment, add constraints (Tr2, s2, t2), where s2 is the state corresponding to the VSS virtual segment that is extended to collide with Tr1, that is, at time t2, Tr2 is not allowed to expand to state s2, and the independent path planning of Tr2's train is re-performed and complies with the constraints.

[0099] Figure 4 (d) The conflicting constraints generated and Figure 4 (c) Similar.

[0100] Further references Figure 6 As an implementation of the methods shown in the above figures, the present application provides an embodiment of a rail transit path planning device for a virtual section, and the device embodiment is similar to Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0101] The embodiment of the present application provides a rail transit path planning device for a virtual section, including:

[0102] The modeling module 1 is configured to map the rail transit network to obtain a rail transit network diagram, wherein the rail transit network diagram includes a plurality of nodes and edges between two nodes, wherein the edge e represents a TTD segment, each TTD segment is divided into a plurality of VSS virtual segments, each train running on the rail transit network is modeled as an intelligent agent, and a position of the intelligent agent in the current state is defined as Pos=(e, a, b), wherein a and b respectively represent the positions of the intelligent agent in the range of a to b in the TTD segment, and 0≤a≤b≤l(e), wherein l(e) is the length of the TTD segment;

[0103] The independent path planning module 2 is configured to obtain the travel mission of the train, perform independent path planning for each agent according to the rail transit network diagram and the travel mission of the train, obtain the planned path of each agent, create a CT node according to the planned path of each agent and add it to the open list, and each CT node stores the planned path of an agent, the cost and constraints of the planned path of the agent;

[0104] The conflict resolution module 3 is configured to select the CT node with the minimum cost of the planned path of the agent in the open list, and determine whether there is a conflict in the planned path corresponding to the CT node with the minimum cost of the planned path of the agent. The conflict includes movement conflict and TTD conflict. The movement conflict includes frontal collision conflict, overlapping collision conflict and rear-end collision conflict. If so, a conflict is selected, and a new CT node with corresponding constraints is created for each agent in the conflict and added to the open list. The constraints are (tri, si, ti), which means that at the time step ti, the agent tri is not allowed to expand to the state si.

[0105] When the conflict is a rear-end collision conflict or a TDD conflict, si is extended to the VSS virtual section where the train rear-end collision conflict or TDD conflict occurs, and the independent path planning of the intelligent agent tri is re-performed and the constraints are complied with. This step is repeated until the planned path corresponding to the CT node with the minimum cost of the planned path of the intelligent agent has no conflict. The planned path corresponding to the CT node with the minimum cost and no conflict of the planned path of the intelligent agent is taken as the optimal planned path of the corresponding train.

[0106] Figure 7 Schematic diagram of the hardware structure of the electronic device provided by the embodiment of the present invention. Figure 7 As shown, the electronic device of this embodiment includes: a processor 701 and a memory 702; wherein the memory 702 is used to store computer-executable instructions; the processor 701 is used to execute the computer-executable instructions stored in the memory to implement the various steps performed by the electronic device in the above embodiment. For details, please refer to the relevant description in the above method embodiment.

[0107] Optionally, the memory 702 may be independent or integrated with the processor 701 .

[0108] When the memory 702 is independently provided, the electronic device further includes a bus 703 for connecting the memory 702 and the processor 701 .

[0109] The embodiment of the present invention further provides a computer storage medium, in which computer execution instructions are stored. When the processor 701 executes the computer execution instructions, the above method is implemented.

[0110] The embodiment of the present invention further provides a computer program product, including a computer program. When the computer program is executed by the processor 701, the above method is implemented.

[0111] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0112] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to implement the solution of this embodiment.

[0113] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each module may exist physically independently, or two or more modules may be integrated into one unit. The above-mentioned module-composed unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0114] The above-mentioned integrated module implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium, including a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor 701 to perform some steps of the methods of various embodiments of the present application.

[0115] It should be understood that the processor 701 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or the processor 701 may be any conventional processor 701, etc. The steps of the method disclosed in the invention may be directly embodied in the execution of the hardware processor 701, or may be executed by a combination of hardware and software modules in the processor 701.

[0116] The memory 702 may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk memory, and may also be a USB flash drive, a mobile hard disk, a read-only memory, a magnetic disk, or an optical disk.

[0117] The bus 703 may be an Industry Standard Architecture (ISA), a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus 703 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus 703 in the drawings of the present application is not limited to only one bus 703 or one type of bus 703.

[0118] The above storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The storage medium can be any available medium that can be accessed by a general or special purpose computer.

[0119] An exemplary storage medium is coupled to the processor 701, so that the processor 701 can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor 701. The processor 701 and the storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as ASIC). Of course, the processor 701 and the storage medium can also exist as discrete components in an electronic device or a main control device.

[0120] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0121] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A rail transit path planning method for virtual sections, characterized in that: The following steps are involved: The rail transit network is map-modeled to obtain a rail transit network graph, wherein the rail transit network graph includes a plurality of nodes and edges between two nodes, wherein the edge e represents a TTD segment, each TTD segment is divided into a plurality of VSS virtual segments, each train running on the rail transit network is modeled as an intelligent agent, and the position of the intelligent agent in the current state is defined as Pos=(e, a, b), wherein a and b respectively represent the positions of the intelligent agent in the range of a to b in the TTD segment, and 0≤a≤b≤l(e), wherein l(e) is the length of the TTD segment; Acquire the travel mission of the train, perform independent path planning for each agent according to the rail transit network diagram and the travel mission of the train, obtain the planned path of each agent, create a CT node according to the planned path of each agent and add it to the open list, each CT node stores the planned path of an agent, the cost and constraints of the planned path of the agent; The CT node with the minimum cost of the planned path of the agent is selected from the open list, and it is determined whether there is a conflict in the planned path corresponding to the CT node with the minimum cost of the planned path of the agent, wherein the conflict includes movement conflict and TTD conflict, and the movement conflict includes frontal collision conflict, overlapping collision conflict and rear-end collision conflict. If so, a conflict is selected, and a new CT node with corresponding constraints is created for each agent in the conflict and added to the open list. The constraint is (tri, si, ti), which means that at the time step ti, the agent tri is not allowed to expand to the state si. When the conflict is a rear-end collision conflict or a TDD conflict, si is extended to the VSS virtual section where the train rear-end collision conflict or TDD conflict occurs, and the independent path planning of the intelligent agent tri is re-performed and complies with the constraints. This step is repeated until the planned path corresponding to the CT node with the smallest cost in the planned path of the intelligent agent has no conflict, and the planned path corresponding to the CT node with the smallest cost and no conflict in the planned path of the intelligent agent is used as the optimal planned path of the corresponding train.

2. The virtual section-oriented rail transit path planning method according to claim 1, characterized in that: The frontal collision conflict and the overlapping collision conflict are resolved as follows: Assume that agent trp expands to the next state sp at time step tp, and agent trq expands to the next state sq at time step tq. If agent trp and agent trq generate a head-on collision conflict or an overlapping collision conflict during the expansion process, generate constraints for agent trp or agent trq: If the constraint (trp, sp, tp) is generated for the agent trp, (trp, sp, tp) means that at the time step tp, trp is not allowed to expand to the state sp, and the independent path planning of the agent trp is re-performed and complies with the constraint; If the constraint (trq, sq, tq) is generated for the agent trq, (trq, sq, tq) means that at the time step tq, trq is not allowed to expand to the state sq, and the independent path planning of the agent trq is re-performed and complies with the constraint.

3. The virtual section-oriented rail transit path planning method according to claim 1, characterized in that: The solutions to the rear-end collision conflict and TDD conflict are as follows: Assume that agent trx expands to the next state sx at time step tx, and agent try expands to the next state sy at time step ty, and the position of agent trx in the current state is ahead of the position of agent try. If agent trx and agent try generate rear-end collision conflict and TDD conflict during the expansion process, generate constraints for agent trx or agent try: If constraints (try, sy, ty) are generated for the agent try, (try, sy, ty) means that at the time step ty, try is not allowed to expand to state sy, sy is the VSS virtual section where a train rear-end collision conflict or TDD conflict occurs, and the independent path planning of the agent try is re-performed and the constraints are observed.

4. The virtual section-oriented rail transit path planning method according to claim 1, characterized in that: The steps of independent path planning are as follows: Use the A* search algorithm to find a single planning path for each agent, ignoring all other agents; set the starting state Add to the open list, Start is the starting position of the train, and the g value of the starting state is set to 0; Select the state s with the minimum cost from the open list to expand, if state s is the target state Dest is the destination of the train, Ttr is the time step for the train to reach the destination, then the search is successful, and the path from the starting state to the target state is used as the planned path of the agent, otherwise the state s is removed from the open list and added to the closed list; For every successor state s of state s next , if the successor state s next In the closed list, it is ignored; If the successor state s next If it is not in the open list, the successor state s next Add to the open list and calculate the successor state s next The path cost is g(s next )=g(s)+t', g(s=t is the path cost from the starting state sinit to state s, t represents the train from the starting state s init The time step to run to state s, t' represents the train moving from state s to the successor state s next time, calculate the successor state s next The target cost h(s next ): , z is the distance the train moves from state s to the target state, Str is the running speed of the train, and the successor state s is calculated. next The cost F(s next ): F(s next )=g(s next )+h(s next ); If the successor state s next In the open list, the computation runs from state s to the successor state s next The path cost g(s next ): g(s next )=g(s)+t'; if it runs from state s to the successor state s next The path cost is less than running from any state other than state s to the successor state s next The path cost is , then the successor state s next The path cost is updated to g(s next )=g(s)+t', update the calculation of the successor state s next The target cost h(s next ) and cost F(s next ), and update the successor state s next The parent node of is in state s, otherwise it is not updated; Repeat the above steps until the state with the minimum cost found is the target state or the open list is empty. If the state with the minimum cost found is the target state, backtrack along the parent node of each state until the starting state is reached, and construct the planning path for each agent.

5. The virtual section-oriented rail transit path planning method according to claim 1, characterized in that: The length of the VSS virtual section is at least the sum of the train length and the safety distance. The maximum number of VSS virtual sections required is r=l(e) / (ltr+d), where ltr represents the train length and d represents the safety distance.

6. The virtual section-oriented rail transit path planning method according to claim 1, characterized in that: Whether there is a conflict is determined by the following method: Assume that the states of the agent Trp are: Assume that the states of the agent Trq are: By looping through the time, it is determined whether the positions of the agent Trp and the agent Trq overlap. If so, it is determined that there is a conflict, otherwise it is determined that there is no conflict.

7. A rail transit path planning device for virtual sections, characterized in that: include: A modeling module is configured to map the rail transit network to obtain a rail transit network graph, wherein the rail transit network graph includes a plurality of nodes and edges between two nodes, wherein an edge e represents a TTD segment, each TTD segment is divided into a plurality of VSS virtual segments, each train running on the rail transit network is modeled as an intelligent agent, and a position of the intelligent agent in a current state is defined as Pos=(e, a, b), wherein a and b respectively represent positions of the intelligent agent in the range of a to b in the TTD segment, and 0≤a≤b≤l(e), wherein l(e) is the length of the TTD segment; The independent path planning module is configured to obtain the travel mission of the train, perform independent path planning for each agent according to the rail transit network diagram and the travel mission of the train, obtain the planned path of each agent, create a CT node according to the planned path of each agent and add it to the open list, each CT node stores the planned path of an agent, the cost and constraints of the planned path of the agent; The conflict resolution module is configured to select the CT node with the minimum cost of the planned path of the agent in the open list, and determine whether there is a conflict in the planned path corresponding to the CT node with the minimum cost of the planned path of the agent, wherein the conflict includes movement conflict and TTD conflict, and the movement conflict includes frontal collision conflict, overlapping collision conflict and rear-end collision conflict. If so, a conflict is selected, and a new CT node with corresponding constraints is created for each agent in the conflict and added to the open list, wherein the constraint is (tri, si, ti), which means that at the time step ti, the agent tri is not allowed to expand to the state si, When the conflict is a rear-end collision conflict or a TDD conflict, si is extended to the VSS virtual section where the train rear-end collision conflict or TDD conflict occurs, and the independent path planning of the intelligent agent tri is re-performed and complies with the constraints. This step is repeated until the planned path corresponding to the CT node with the smallest cost in the planned path of the intelligent agent has no conflict, and the planned path corresponding to the CT node with the smallest cost and no conflict in the planned path of the intelligent agent is used as the optimal planned path of the corresponding train.

8. An electronic device, comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.