Auxiliary dispatching method and device for rail transit

Abstract

The application provides an auxiliary scheduling method and device for rail transit, wherein the method comprises the following steps: acquiring multi-source heterogeneous sensing data of a target line; analyzing the multi-source heterogeneous sensing data to determine a line operation situation map of the target line; determining at least one candidate scheduling strategy based on abnormal operation information in the line operation situation map, and simulating each candidate scheduling strategy in a digital twin decision sandbox to determine a simulation result of each candidate scheduling strategy; and screening an optimal target strategy based on the simulation result and an operation optimization target of the target line, and executing the optimal target strategy. By introducing the digital twin decision sandbox technology before decision making, the application performs parallel super real-time simulation deduction on a plurality of candidate scheduling strategies generated for abnormal situations in a virtual environment, quantitatively evaluates the influence of each strategy on future operation indexes in advance, ensures the decision accuracy, and improves the response efficiency of the rail transit line operation state.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and in particular to an auxiliary scheduling method and apparatus for rail transit. Background Technology

[0002] With the continuous expansion of the urban rail transit network and the sustained increase in passenger flow density, the operating environment of the rail transit system is becoming increasingly complex.

[0003] The existing rail transit dispatching and command mainly relies on basic monitoring data provided by the Automatic Train Supervision (ATS) system. When abnormal situations occur on the line, dispatchers usually make manual judgments based on established emergency plans or their own professional experience, and then choose a handling method based on experience, and manually issue control commands through the dispatching workstation to adjust train operation.

[0004] Scheduling methods that rely on human experience require a significant amount of time to analyze complex problems and formulate scheduling strategies, resulting in low efficiency in handling emergencies. Summary of the Invention

[0005] This invention provides an auxiliary scheduling method and apparatus for rail transit, which solves the problem that the existing rail transit scheduling relies on human experience, requires a lot of time to analyze complex problems and formulate scheduling strategies, and results in low efficiency in handling emergencies, thereby improving the scheduling efficiency of rail transit.

[0006] This invention provides an auxiliary scheduling method for rail transit, comprising the following steps: Acquire multi-source heterogeneous sensing data for the target line; the multi-source heterogeneous sensing data includes at least train operation data, signal equipment status data, passenger flow data, and line environment sensing data; The multi-source heterogeneous sensing data is analyzed to determine the current and future operating status of the target line, forming a line operation status map. Based on the abnormal operation information in the line operation status diagram, at least one candidate scheduling strategy is determined, and each candidate scheduling strategy is simulated in a digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy; the digital twin decision sandbox is a virtual simulation environment constructed for the target line; Based on the simulation results of each candidate scheduling strategy and the operation optimization objective of the target line, the optimal target strategy is selected from the candidate scheduling strategies and then executed.

[0007] According to the auxiliary scheduling method for rail transit provided by the present invention, the execution of the optimal target strategy includes: The optimal target strategy is decomposed into multiple atomic operation instructions over time; the atomic operation instructions correspond to the control interface of the Automatic Train Monitoring System (ATS) of the target line. The atomic operation instructions are subjected to logical conflict verification and security boundary verification. After the verification is passed, the atomic operation instructions are issued for execution based on the interface of the ATS system.

[0008] According to an auxiliary scheduling method for rail transit provided by the present invention, the step of issuing and executing the atomic operation command based on the interface of the ATS system includes: The simulation results of the optimal objective strategy and the atomization operation instructions are sent to the human-computer interaction interface. The system receives a confirmation instruction for the atomization operation instruction and, based on the interface of the ATS system, issues and executes the atomization operation instruction.

[0009] According to the auxiliary scheduling method for rail transit provided by the present invention, the digital twin decision sandbox is constructed as follows: A digital track topology network is constructed using the engineering data of the target line; In the digital track topology network, signal interlocking logic rules and control algorithms consistent with the CBTC system of the target line are loaded, and a train dynamics model is configured based on the train vehicle parameters in the target line to obtain an initial digital twin decision sandbox. The real-time operating status of the target line is mapped to the initial digital twin decision sandbox to obtain the digital twin decision sandbox.

[0010] According to the auxiliary scheduling method for rail transit provided by the present invention, the step of selecting the optimal target strategy from the candidate scheduling strategies based on the simulation results of each candidate scheduling strategy and the operation optimization objective of the target line includes: The target line is given multiple preset operational optimization objectives, including at least the shortest total passenger waiting time, the minimum total train delay time, and the lowest total line energy consumption. Calculate the objective function value of each candidate scheduling strategy under each of the operational optimization objectives based on the simulation results; Pareto optimal boundary analysis is performed on each of the objective function values ​​to determine the strategy with the highest score as the optimal objective strategy.

[0011] According to the auxiliary scheduling method for rail transit provided by the present invention, the step of simulating each candidate scheduling strategy in a digital twin decision sandbox and determining the simulation results of each candidate scheduling strategy includes: The current equipment status, train location, and passenger flow distribution of the target line are mapped to the digital twin decision sandbox; Each of the candidate scheduling strategies is loaded into the digital twin decision sandbox, and the simulation results of each candidate scheduling strategy are determined.

[0012] The present invention also provides an auxiliary dispatching device for rail transit, comprising the following modules: The data acquisition module is used to acquire multi-source heterogeneous sensing data of the target line; the multi-source heterogeneous sensing data includes at least train operation data, signal equipment status data, passenger flow data, and line environment sensing data; The situation analysis module is used to analyze the multi-source heterogeneous sensing data to determine the current and future operating status of the target line, forming a line operation situation diagram. The simulation module is used to determine at least one candidate scheduling strategy based on the abnormal operation information in the line operation status diagram, and to simulate each candidate scheduling strategy in a digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy; the digital twin decision sandbox is a virtual simulation environment constructed for the target line; The execution module is used to select the optimal target strategy from the candidate scheduling strategies based on the simulation results of each candidate scheduling strategy and the operation optimization target of the target line, and to execute the optimal target strategy.

[0013] The present invention also provides an auxiliary scheduling system for rail transit, which is connected to the Automatic Train Control (ATS) system of the target line and is used to execute the auxiliary scheduling method for rail transit as described above.

[0014] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the program to implement the auxiliary scheduling method for rail transit as described above.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the auxiliary scheduling method for rail transit as described above.

[0016] The auxiliary scheduling method and device for rail transit provided by this invention achieves in-depth understanding and prediction of the current and future operating status of the line by acquiring multi-source heterogeneous sensing data from multiple dimensions, including train operation, signal status, and passenger flow, in real time. Based on this, digital twin decision sandbox technology is introduced, enabling parallel ultra-real-time simulation and deduction of multiple candidate scheduling strategies generated for abnormal situations in a virtual environment. This allows for the pre-quantification and assessment of the impact of each strategy on future operating indicators, avoiding the lag and subjectivity of manual decision-making and ensuring decision accuracy through simulation optimization. This achieves second-level automatic decision-making for emergencies, improving the response efficiency of rail transit line operating status. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the scheduling structure of the relevant methods.

[0019] Figure 2 This is a flowchart illustrating the auxiliary scheduling method for rail transit provided by the present invention.

[0020] Figure 3 This is a schematic diagram of the intelligent collaborative decision-making auxiliary scheduling subsystem provided by the present invention.

[0021] Figure 4 This is a schematic diagram of the structure of the auxiliary dispatching system for rail transit provided by the present invention.

[0022] Figure 5 This is a schematic diagram of the structure of the auxiliary dispatching device for rail transit provided by the present invention.

[0023] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0025] Fully automated train operation (UTO) systems have been widely adopted in urban rail transit. These systems can autonomously manage train operation, precise station stops, automatic turnarounds, and coordinated control of multiple systems including signaling, power supply, and platform screen doors, without requiring manual intervention for routine operations. The specific implementation architecture is as follows: Figure 1 The scheduling structure diagram of the relevant method is shown. In the core scheduling stage, the system adopts a collaborative model of ATS system execution + scheduler decision-making, with the specific division of labor as follows: The functions of ATS (Automatic Train Service) are as the central hub for dispatching and execution, undertaking standardized dispatching tasks. It monitors key data such as the location, speed, and status of trains across the network in real time; automatically executes pre-programmed timetables to ensure trains run on schedule; can trigger control commands with preset logic such as train detaining and train stopping; and immediately sends alarm information to dispatchers when train malfunctions or equipment abnormalities are detected. However, it does not have the ability to handle complex problems independently.

[0026] The role of the human dispatcher: As the core of dispatching decisions, they are responsible for handling complex scenarios that exceed the system's preset logic. They need to continuously monitor the ATS (Automatic Train Service) interface 24 / 7. When unconventional situations occur, such as escalating equipment failures, widespread network-wide delays, or sudden surges in passenger flow at stations, dispatchers will combine their professional experience and operational knowledge to comprehensively analyze multi-dimensional information, including ATS alarm data, station video surveillance footage, and passenger information system feedback. Through the ATS interface, they will manually execute operations such as adjusting timetables, resetting train routes, instructing trains to clear passengers, or bypassing stations to quickly restore normal line operation.

[0027] Significant drawbacks of related methods that rely on human decision-making include: Response lag and inefficiency: Manual analysis of complex problems and formulation of scheduling strategies take time, making it difficult to make real-time optimal responses to rapidly changing operating conditions, and easily missing the best handling window, leading to the escalation of the situation; Highly dependent on personal experience: Dispatch quality is strongly correlated with the dispatcher's experience, skill level, and physical and mental state. The decision-making process is difficult to standardize, and the handling methods and effects of different dispatchers may vary greatly. Heavy workload and psychological pressure: In emergency situations, dispatchers need to process massive amounts of information and face enormous pressure in a short period of time, which can easily lead to human error and pose safety hazards. Lack of foresight and global optimization: Human decision-making usually focuses on solving current problems and makes it difficult to accurately assess the long-term and cascading effects of scheduling strategies. It also lacks global-level multi-objective optimization (such as taking into account efficiency, energy consumption and passenger experience).

[0028] To address the shortcomings of related methods, this invention provides an auxiliary scheduling method for rail transit. Figure 2This is a flowchart illustrating the auxiliary scheduling method for rail transit provided by the present invention, as shown below. Figure 2 As shown, the method includes the following: Step 210: Obtain multi-source heterogeneous sensing data of the target line; the multi-source heterogeneous sensing data includes at least train operation data, signal equipment status data, passenger flow data, and line environment sensing data; Step 220: Analyze the multi-source heterogeneous sensing data to determine the current and future operating status of the target line; Step 230: Based on the abnormal operation information in the line operation status diagram, determine at least one candidate scheduling strategy, and simulate each candidate scheduling strategy in the digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy; the digital twin decision sandbox is a virtual simulation environment constructed for the target line; Step 240: Based on the simulation results of each candidate scheduling strategy and the operation optimization objective of the target line, select the optimal target strategy from each candidate scheduling strategy and execute the optimal target strategy.

[0029] The executor of the auxiliary scheduling method for rail transit provided by this invention can be an electronic device, a component within the electronic device, an integrated circuit, or a chip. The electronic device can be a mobile electronic device or a non-mobile electronic device. For example, a mobile electronic device can be a mobile phone, tablet computer, laptop computer, PDA, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc., while a non-mobile electronic device can be a server, network attached storage (NAS), or personal computer (PC), etc. This invention does not impose specific limitations.

[0030] The following section uses the example of a computer executing the auxiliary scheduling method for rail transit provided by this invention to illustrate the technical solution of this invention in detail.

[0031] In step 210, multi-source heterogeneous sensing data of the target line is acquired; the multi-source heterogeneous sensing data includes at least train operation data, signal equipment status data, passenger flow data, and line environment sensing data.

[0032] In this embodiment, the target line refers to a rail transit line that requires scheduling and management. Multi-source heterogeneous sensing data refers to raw data from different systems, formats, and dimensions. Specifically, basic data from multiple underlying systems can be accessed and integrated in real time through standard interfaces.

[0033] Train operation data and signal equipment status data mainly come from automatic train monitoring systems, interlocking systems, and area controllers, including the location, speed, operating status, turnout status, and signal status of all trains in the network; passenger flow data comes from automatic fare collection systems, platform screen door systems, and passenger flow monitoring systems, used to reflect the degree of congestion in stations and carriages; and line environment perception data comes from onboard equipment, video analysis systems, power monitoring systems, and external information sources, such as meteorological data and information on large-scale events.

[0034] By acquiring massive amounts of multi-source heterogeneous data, the limitations of relying solely on ATS alarm data can be avoided, enabling a high-precision perception process for operational anomalies.

[0035] In step 220, the multi-source heterogeneous sensing data is analyzed to determine the current and future operating status of the target line, forming a line operation status diagram.

[0036] Based on the acquired data, in-depth analysis is performed using the built-in operational status awareness module. Specifically, knowledge graph and machine learning model technologies can be used to perform correlation analysis on the fused data.

[0037] The network operation status map not only provides a static display of the current network operation status, such as train location and congestion level, but also includes predictions of the operation status in future time periods. For example, it can not only identify the current event of train A malfunctioning at station B, but also predict its cascading effects based on historical data and models, such as predicting that it will cause congestion on the downstream line for more than 15 minutes in 5 minutes. Thus, a dynamic operation status map is constructed.

[0038] In step 230, based on the abnormal operation information in the line operation status diagram, at least one candidate scheduling strategy is determined, and each candidate scheduling strategy is simulated in the digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy.

[0039] A digital twin decision sandbox is pre-built. This sandbox is a virtual simulation environment highly mapped to the real physical line, which can be set up at the ground control center of the target line and connected to various systems within the target line. The digital twin decision sandbox is used to simulate the execution of the strategy before it is implemented, in order to determine the simulation execution result of the strategy.

[0040] Specifically, when abnormal operation information appears in the line operation status diagram, such as equipment failure, large passenger flow delays, and delays, the scheduling strategy generation mechanism will be triggered.

[0041] At least one candidate scheduling strategy is identified. This can be achieved through a built-in scheduling strategy generator, which combines a pre-defined expert rule base with a reinforcement learning algorithm. For the identified problem, multiple feasible solutions are automatically generated. For example, strategies might include reducing station dwell time, initiating short-route operations, or organizing small-scale train decommissioning or skipping stops.

[0042] Each generated candidate scheduling strategy is loaded into the digital twin decision sandbox for real-time simulation to determine the simulation results of each candidate scheduling strategy.

[0043] Simulation results can be obtained for each candidate scheduling strategy. These simulation results specifically predict key indicators of line operation over a future period, including the impact on punctuality, passenger flow management efficiency, and changes in train energy consumption.

[0044] In step 240, based on the simulation results of each candidate scheduling strategy and the operation optimization objective of the target line, the optimal target strategy is selected from each candidate scheduling strategy and the optimal target strategy is executed.

[0045] Based on the quantitative results derived from the simulation, and combined with the pre-set operational optimization goals, a comprehensive evaluation is conducted on each candidate strategy, and the strategy with the highest score or the one that best meets the current operational needs is selected as the optimal target strategy.

[0046] The optimal target strategy is executed. It should be noted that the execution process can be the process of translating the optimal target strategy into specific control commands and sending them to the underlying equipment, thereby realizing the automatic adjustment of train operation.

[0047] In the implementation of this invention, an intelligent collaborative decision-making auxiliary scheduling subsystem can be built on top of the ATS system. A schematic diagram of the intelligent collaborative decision-making auxiliary scheduling subsystem is shown below. Figure 3 The structural diagram of the intelligent collaborative decision-making auxiliary scheduling subsystem provided by this invention is shown. The intelligent collaborative decision-making auxiliary scheduling subsystem specifically includes a situational awareness and cognition layer, an intelligent decision-making and strategy generation layer, and a decision execution and interface layer.

[0048] The situation awareness and cognition layer includes a multi-source data fusion module and an operational situation cognition module.

[0049] Multi-source data fusion module: Through standard interfaces, it can access and fuse multi-source heterogeneous sensing data from the target line's ATS, signaling system, maintenance monitoring system, on-board equipment, video analysis, passenger flow monitoring, power monitoring system, and even external information (such as weather and large-scale events) in real time.

[0050] Operational Status Awareness Module: Based on knowledge graphs and machine learning models, it performs in-depth analysis of multi-source heterogeneous sensing data, not only identifying current events (such as train A malfunction) but also recognizing their impact (such as predicting that it will cause congestion on the downstream line for more than 15 minutes in 5 minutes), forming a line operational status map of the current and future operational status of the target line.

[0051] The intelligent decision-making and strategy layer includes a scheduling strategy generator, a digital twin decision sandbox, and a strategy evaluation and optimization module.

[0052] Scheduling strategy generator: Built-in rule base and reinforcement learning algorithm. Based on the abnormal operation information identified by the cognitive layer, it automatically generates multiple feasible candidate scheduling strategies (such as adjusting the timetable, starting short routes, organizing small-scale train decommissioning, etc.).

[0053] Digital twin decision sandbox: After multiple feasible candidate scheduling strategies are generated, they are immediately simulated and extrapolated in a digital twin decision sandbox built based on digital twin technology in real time to predict the impact of each strategy on key indicators such as on-time rate, passenger flow management, and energy consumption in the next 5-30 minutes.

[0054] Strategy Evaluation and Optimization Module: Based on simulation results, the module evaluates and selects the Pareto optimal strategy from multiple optimization objectives (such as minimizing total delay time, minimizing total passenger waiting time, and minimizing total energy consumption).

[0055] The decision execution and interface layer includes an instruction conversion and security verification module, as well as an instruction issuance and feedback interface.

[0056] The instruction conversion and safety verification module automatically decomposes the selected optimal strategy and converts it into a series of ATS-executable, atomic safety operation instructions (such as setting train T101 to stop at station S05, modifying the departure time of line R02), and performs logical conflict verification.

[0057] Command issuance and feedback interface: Through the enhanced ATS API, the command sequence is safely and reliably issued to the underlying ATS system for execution, and the execution status is monitored in real time, forming a closed loop of decision-making-execution-feedback.

[0058] The auxiliary scheduling method for rail transit provided by this invention achieves in-depth understanding and prediction of the current and future operational status of the line by acquiring multi-source heterogeneous sensing data from multiple dimensions, including train operation, signal status, and passenger flow, in real time. Based on this, digital twin decision sandbox technology is introduced, enabling parallel, ultra-real-time simulation and deduction of multiple candidate scheduling strategies generated for abnormal situations in a virtual environment. This allows for the pre-quantification and assessment of the impact of each strategy on future operational indicators, avoiding the lag and subjectivity of manual decision-making and ensuring decision accuracy through simulation optimization. This achieves second-level automatic decision-making for emergencies, improving the response efficiency of rail transit line operation.

[0059] In one embodiment, executing the optimal target strategy includes: The optimal target strategy is decomposed into multiple atomic operation instructions over time; the atomic operation instructions correspond to the control interface of the Automatic Train Monitoring System (ATS) of the target line. The atomic operation instructions are subjected to logical conflict verification and security boundary verification. After the verification is passed, the atomic operation instructions are issued for execution based on the interface of the ATS system.

[0060] Since the strategies generated at higher levels are typically macroscopic, the ATS system cannot directly recognize them. Therefore, an instruction conversion module is needed to automatically decompose the macroscopic strategies and convert them into a series of basic atomic operation instructions that can be executed by the ATS.

[0061] For example, the decision to stop train T101 at station S05 is broken down into specific route control instructions and stop time modification instructions. The generated instructions must conform to the standard format of the ATS API.

[0062] Atomic operation instructions undergo logical conflict checking and safety boundary checking. After passing the checks, the atomic operation instructions are issued for execution via the ATS system interface. Before being issued, instructions must undergo rigorous checks by the safety verification module. Logical conflict checking ensures that instructions do not contradict each other; safety boundary checking ensures that instructions do not violate interlocking safety principles. Only instruction sequences that pass all checks are sent to the ATS.

[0063] After the instruction verification is completed, the atomic operation instructions are issued and executed based on the interface of the ATS system.

[0064] In one embodiment, the execution of the atomic operation instruction based on the interface of the ATS system includes: The simulation results of the optimal objective strategy and the atomization operation instructions are sent to the human-computer interaction interface. The system receives a confirmation instruction for the atomization operation instruction and, based on the interface of the ATS system, issues and executes the atomization operation instruction.

[0065] During the command issuance and execution process, the simulation results of the optimal target strategy and the atomic operation commands are sent to the human-machine interface. The dispatcher at the ground control center can view the simulation results of the optimal target strategy and the atomic operation commands through the human-machine interface.

[0066] The dispatcher has the highest authority, allowing them to view the simulation prediction metrics and confirm them before clicking "confirm." Only after receiving manual confirmation does the system officially send the command to the underlying system for execution via the ATS API interface, and monitors the execution feedback in real time.

[0067] Understandably, relying on further human judgment not only utilizes automation technology to improve decision-making efficiency, but also ensures operational safety in complex scenarios through final human confirmation.

[0068] In one embodiment, the digital twin decision sandbox is constructed as follows: A digital track topology network is constructed using the engineering data of the target line; In the digital track topology network, signal interlocking logic rules and control algorithms consistent with the CBTC system of the target line are loaded, and a train dynamics model is configured based on the train vehicle parameters in the target line to obtain an initial digital twin decision sandbox. The real-time operating status of the target line is mapped to the initial digital twin decision sandbox to obtain the digital twin decision sandbox.

[0069] Specifically, a digital twin decision sandbox needs to be built in advance before auxiliary scheduling of rail transit.

[0070] A digital track topology network is constructed based on the engineering data of the target line. This engineering data includes fundamental geospatial data such as the line's physical layout, gradients, curve radii, station locations, and turnout configurations. Using this data, a static structure completely identical to the real line is built in virtual space.

[0071] In a digital track topology network, signal interlocking logic rules and control algorithms consistent with the communication-based train control system of the target line are loaded. A train dynamics model is configured based on the train vehicle parameters of the target line to obtain an initial digital twin decision sandbox. To ensure the accuracy of the simulation, the consistency of the internal logic must be guaranteed.

[0072] The signal interlocking logic and automatic train protection / automatic train driving algorithms of real CBTC systems are ported or replicated into the sandbox. Simultaneously, a high-precision train dynamics model is configured by incorporating train traction and braking characteristic curves, train formation information, and other data. This makes the virtual train's behavior in the sandbox highly similar to that of a real train.

[0073] The real-time operating status of the target line is mapped to an initial digital twin decision sandbox to obtain the digital twin decision sandbox. The real-world equipment status, such as turnout positioning, signal displays, train position and speed, and passenger flow distribution, is synchronized in real time to achieve state synchronization in the virtual environment.

[0074] In one embodiment, the step of selecting the optimal target strategy from the candidate scheduling strategies based on the simulation results of each candidate scheduling strategy and the operational optimization objective of the target line includes: The target line is given multiple preset operational optimization objectives, including at least the shortest total passenger waiting time, the minimum total train delay time, and the lowest total line energy consumption. Calculate the objective function value of each candidate scheduling strategy under each of the operational optimization objectives based on the simulation results; Pareto optimal boundary analysis is performed on each of the objective function values ​​to determine the strategy with the highest score as the optimal objective strategy.

[0075] The optimal target strategy is selected from among the candidate scheduling strategies, specifically including: The target route is given multiple preset operational optimization objectives, including at least the shortest total passenger waiting time, the minimum total train delay time, and the lowest total line energy consumption. Traditional scheduling often focuses on only a single indicator, while this invention adopts a multi-objective optimization perspective, comprehensively considering operational efficiency, passenger experience, and energy consumption.

[0076] Calculate the objective function value of each candidate scheduling strategy under each of the stated operational optimization objectives based on the simulation results. Based on the obtained simulation data, calculate the score or loss value of each strategy in each of the aforementioned dimensions.

[0077] Pareto optimal boundary analysis is performed on each of the stated objective function values ​​to determine the strategy with the highest score as the optimal objective strategy. Since there may be conflicts between the various optimization objectives—for example, reducing delays might require acceleration, thus increasing energy consumption—a multi-objective optimization algorithm is used to find the globally optimal solution that maximizes overall benefits.

[0078] In one embodiment, simulating each candidate scheduling strategy in a digital twin decision sandbox and determining the simulation results of each candidate scheduling strategy includes: The current equipment status, train location, and passenger flow distribution of the target line are mapped to the digital twin decision sandbox; Each of the candidate scheduling strategies is loaded into the digital twin decision sandbox, and the simulation results of each candidate scheduling strategy are determined.

[0079] For example, if strategy A skips a stop at a certain station, the virtual train in the sandbox will run according to the skipping logic, and the virtual passengers will replan their routes based on the skipping.

[0080] During the simulation, the evolution of the line's operating schedule, the dissipation of passenger congestion, and the cumulative energy consumption of trains are projected in the future timeframe following the implementation of the strategy. The final simulation output is a set of quantitative indicator data for subsequent evaluation.

[0081] The present invention also provides a schematic diagram of the structure of an auxiliary dispatching system for rail transit, as shown in the figure. Figure 4 As shown, the system communicates with the ATS system. The system specifically includes a situational awareness and cognition layer, an intelligent decision-making and strategy layer, and a decision execution and interface layer.

[0082] The situation awareness and cognition layer includes a multi-source data fusion module and an operational situation cognition module.

[0083] Multi-source data fusion module: Through standard interfaces, it can access and fuse multi-source heterogeneous sensing data from the target line's ATS, signaling system, maintenance monitoring system, on-board equipment, video analysis, passenger flow monitoring, power monitoring system, and even external information (such as weather and large-scale events) in real time.

[0084] Operational Status Awareness Module: Based on knowledge graphs and machine learning models, it performs in-depth analysis of multi-source heterogeneous sensing data, not only identifying current events (such as train A malfunction) but also recognizing their impact (such as predicting that it will cause congestion on the downstream line for more than 15 minutes in 5 minutes), forming a line operational status map of the current and future operational status of the target line.

[0085] The intelligent decision-making and strategy layer includes a scheduling strategy generator, a digital twin decision sandbox, and a strategy evaluation and optimization module.

[0086] Scheduling strategy generator: Built-in rule base and reinforcement learning algorithm. Based on the abnormal operation information identified by the cognitive layer, it automatically generates multiple feasible candidate scheduling strategies (such as adjusting the timetable, starting short routes, organizing small-scale train decommissioning, etc.).

[0087] Digital twin decision sandbox: After multiple feasible candidate scheduling strategies are generated, they are immediately simulated and extrapolated in a digital twin decision sandbox built based on digital twin technology in real time to predict the impact of each strategy on key indicators such as on-time rate, passenger flow management, and energy consumption in the next 5-30 minutes.

[0088] Strategy Evaluation and Optimization Module: Based on simulation results, the module evaluates and selects the Pareto optimal strategy from multiple optimization objectives (such as minimizing total delay time, minimizing total passenger waiting time, and minimizing total energy consumption).

[0089] The decision execution and interface layer includes an instruction conversion and security verification module, as well as an instruction issuance and feedback interface.

[0090] The instruction conversion and safety verification module automatically decomposes the selected optimal strategy and converts it into a series of ATS-executable, atomic safety operation instructions (such as setting train T101 to stop at station S05, modifying the departure time of line R02), and performs logical conflict verification.

[0091] Command issuance and feedback interface: Through the enhanced ATS API, the command sequence is safely and reliably issued to the underlying ATS system for execution, and the execution status is monitored in real time, forming a closed loop of decision-making-execution-feedback.

[0092] The auxiliary scheduling device for rail transit provided by the present invention will be described below. The auxiliary scheduling device for rail transit described below can be referred to in correspondence with the auxiliary scheduling method for rail transit described above.

[0093] like Figure 5 As shown, the device includes: The data acquisition module 510 is used to acquire multi-source heterogeneous sensing data of the target line; the multi-source heterogeneous sensing data includes at least train operation data, signal equipment status data, passenger flow data and line environment sensing data; The situation analysis module 520 is used to analyze the multi-source heterogeneous sensing data to determine the current and future operating status of the target line, and to obtain a line operation situation diagram. The simulation module 530 is used to determine at least one candidate scheduling strategy based on the abnormal operation information in the line operation status diagram, and to simulate each candidate scheduling strategy in a digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy; the digital twin decision sandbox is a virtual simulation environment constructed for the target line; The execution module 540 is used to select the optimal target strategy from the candidate scheduling strategies based on the simulation results of each candidate scheduling strategy and the operation optimization target of the target line, and to execute the optimal target strategy.

[0094] The auxiliary dispatching device for rail transit provided by this invention achieves in-depth understanding and prediction of the current and future operational status of the line by acquiring multi-source heterogeneous sensing data from multiple dimensions, including train operation, signal status, and passenger flow, in real time. Based on this, it introduces digital twin decision sandbox technology, enabling parallel ultra-real-time simulation and deduction of multiple candidate dispatching strategies generated for abnormal situations in a virtual environment. This allows for the pre-quantification and assessment of the impact of each strategy on future operational indicators, avoiding the lag and subjectivity of manual decision-making and ensuring decision accuracy through simulation optimization. This achieves second-level automatic decision-making for emergencies, improving the response efficiency of rail transit line operation.

[0095] In one embodiment, the execution module 540 is specifically used for: The execution of the optimal objective strategy includes: The optimal target strategy is decomposed into multiple atomic operation instructions over time; the atomic operation instructions correspond to the control interface of the Automatic Train Monitoring System (ATS) of the target line. The atomic operation instructions are subjected to logical conflict verification and security boundary verification. After the verification is passed, the atomic operation instructions are issued for execution based on the interface of the ATS system.

[0096] In one embodiment, the execution module 540 is specifically used for: Based on the interface of the ATS system, the atomic operation instructions are issued and executed, including: The simulation results of the optimal objective strategy and the atomization operation instructions are sent to the human-computer interaction interface. The system receives a confirmation instruction for the atomization operation instruction and, based on the interface of the ATS system, issues and executes the atomization operation instruction.

[0097] In one embodiment, the simulation module 530 is specifically used for: The digital twin decision sandbox is constructed as follows: A digital track topology network is constructed using the engineering data of the target line; In the digital track topology network, signal interlocking logic rules and control algorithms consistent with the CBTC system of the target line are loaded, and a train dynamics model is configured based on the train vehicle parameters in the target line to obtain an initial digital twin decision sandbox. The real-time operating status of the target line is mapped to the initial digital twin decision sandbox to obtain the digital twin decision sandbox.

[0098] In one embodiment, the execution module 540 is further configured to: The process of selecting the optimal target strategy from the candidate scheduling strategies based on the simulation results of each candidate scheduling strategy and the operational optimization objective of the target line includes: The target line is given multiple preset operational optimization objectives, including at least the shortest total passenger waiting time, the minimum total train delay time, and the lowest total line energy consumption. Calculate the objective function value of each candidate scheduling strategy under each of the operational optimization objectives based on the simulation results; Pareto optimal boundary analysis is performed on each of the objective function values ​​to determine the strategy with the highest score as the optimal objective strategy.

[0099] In one embodiment, the execution module 540 is further configured to: The process of simulating each candidate scheduling strategy in a digital twin decision sandbox and determining the simulation results of each candidate scheduling strategy includes: The current equipment status, train location, and passenger flow distribution of the target line are mapped to the digital twin decision sandbox; Each of the candidate scheduling strategies is loaded into the digital twin decision sandbox, and the simulation results of each candidate scheduling strategy are determined.

[0100] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include a processor 610, a communications interface 620, a memory 630, and a communication bus 640. The processor 610, communications interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions from the memory 630 to execute an auxiliary scheduling method for rail transit. This method includes: acquiring multi-source heterogeneous sensing data of the target line; the multi-source heterogeneous sensing data includes at least train operation data, signal equipment status data, passenger flow data, and line environment sensing data. The multi-source heterogeneous sensing data is analyzed to determine the current and future operating status of the target line, forming a line operation status map. Based on the abnormal operation information in the line operation status diagram, at least one candidate scheduling strategy is determined, and each candidate scheduling strategy is simulated in a digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy; the digital twin decision sandbox is a virtual simulation environment constructed for the target line; Based on the simulation results of each candidate scheduling strategy and the operation optimization objective of the target line, the optimal target strategy is selected from the candidate scheduling strategies and then executed.

[0101] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0102] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the auxiliary scheduling method for rail transit provided by the above methods. The method includes: acquiring multi-source heterogeneous sensing data of a target line; the multi-source heterogeneous sensing data includes at least train operation data, signal equipment status data, passenger flow data, and line environment sensing data. The multi-source heterogeneous sensing data is analyzed to determine the current and future operating status of the target line, forming a line operation status map. Based on the abnormal operation information in the line operation status diagram, at least one candidate scheduling strategy is determined, and each candidate scheduling strategy is simulated in a digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy; the digital twin decision sandbox is a virtual simulation environment constructed for the target line; Based on the simulation results of each candidate scheduling strategy and the operation optimization objective of the target line, the optimal target strategy is selected from the candidate scheduling strategies and then executed.

[0103] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the auxiliary scheduling method for rail transit provided by the above methods, the method comprising: acquiring multi-source heterogeneous sensing data of a target line; the multi-source heterogeneous sensing data including at least train operation data, signal equipment status data, passenger flow data, and line environment sensing data; The multi-source heterogeneous sensing data is analyzed to determine the current and future operating status of the target line, forming a line operation status map. Based on the abnormal operation information in the line operation status diagram, at least one candidate scheduling strategy is determined, and each candidate scheduling strategy is simulated in a digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy; the digital twin decision sandbox is a virtual simulation environment constructed for the target line; Based on the simulation results of each candidate scheduling strategy and the operation optimization objective of the target line, the optimal target strategy is selected from the candidate scheduling strategies and then executed.

[0104] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0105] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An auxiliary scheduling method for rail transit, characterized in that, include: Acquire multi-source heterogeneous sensing data of the target line; The multi-source heterogeneous sensing data includes at least train operation data, signal equipment status data, passenger flow data, and line environment sensing data. The multi-source heterogeneous sensing data is analyzed to determine the current and future operating status of the target line, forming a line operation status map. Based on the abnormal operation information in the line operation status diagram, at least one candidate scheduling strategy is determined, and each candidate scheduling strategy is simulated in the digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy. The digital twin decision sandbox is a virtual simulation environment constructed for the target line; Based on the simulation results of each candidate scheduling strategy and the operation optimization objective of the target line, the optimal target strategy is selected from each candidate scheduling strategy and the optimal target strategy is executed. The digital twin decision sandbox is constructed as follows: A digital track topology network is constructed using the engineering data of the target line; In the digital track topology network, signal interlocking logic rules and control algorithms consistent with the CBTC system of the target line are loaded, and a train dynamics model is configured based on the train vehicle parameters in the target line to obtain an initial digital twin decision sandbox. The real-time operating status of the target line is mapped to the initial digital twin decision sandbox to obtain the digital twin decision sandbox.

2. The auxiliary scheduling method for rail transit according to claim 1, characterized in that, The execution of the optimal objective strategy includes: The optimal target strategy is decomposed into multiple atomic operation instructions over time; the atomic operation instructions correspond to the control interface of the Automatic Train Monitoring System (ATS) of the target line. The atomic operation instructions are subjected to logical conflict verification and security boundary verification. After the verification is passed, the atomic operation instructions are issued for execution based on the interface of the ATS system.

3. The auxiliary scheduling method for rail transit according to claim 2, characterized in that, The interface based on the ATS system sends out and executes the atomic operation instructions, including: The simulation results of the optimal objective strategy and the atomization operation instructions are sent to the human-computer interaction interface. The system receives a confirmation instruction for the atomization operation instruction and, based on the interface of the ATS system, issues and executes the atomization operation instruction.

4. The auxiliary scheduling method for rail transit according to claim 1, characterized in that, The process of selecting the optimal target strategy from the candidate scheduling strategies based on the simulation results of each candidate scheduling strategy and the operational optimization objective of the target line includes: The target line is given multiple preset operational optimization objectives, including at least the shortest total passenger waiting time, the minimum total train delay time, and the lowest total line energy consumption. Calculate the objective function value of each candidate scheduling strategy under each of the operational optimization objectives based on the simulation results; Pareto optimal boundary analysis is performed on each of the objective function values ​​to determine the strategy with the highest score as the optimal objective strategy.

5. The auxiliary scheduling method for rail transit according to claim 1, characterized in that, The process of simulating each candidate scheduling strategy in a digital twin decision sandbox and determining the simulation results of each candidate scheduling strategy includes: The current equipment status, train location, and passenger flow distribution of the target line are mapped to the digital twin decision sandbox; Each of the candidate scheduling strategies is loaded into the digital twin decision sandbox, and the simulation results of each candidate scheduling strategy are determined.

6. An auxiliary dispatching device for rail transit, characterized in that, include: The data acquisition module is used to acquire multi-source heterogeneous sensing data of the target line; The multi-source heterogeneous sensing data includes at least train operation data, signal equipment status data, passenger flow data, and line environment sensing data. The situation analysis module is used to analyze the multi-source heterogeneous sensing data to determine the current and future operating status of the target line, forming a line operation situation diagram. The simulation module is used to determine at least one candidate scheduling strategy based on the abnormal operation information in the line operation status diagram, and to simulate each candidate scheduling strategy in the digital twin decision sandbox to determine the simulation results of each candidate scheduling strategy. The digital twin decision sandbox is a virtual simulation environment constructed for the target line; The execution module is used to select the optimal target strategy from the candidate scheduling strategies based on the simulation results of each candidate scheduling strategy and the operation optimization target of the target line, and to execute the optimal target strategy. The digital twin decision sandbox is constructed as follows: A digital track topology network is constructed using the engineering data of the target line; In the digital track topology network, signal interlocking logic rules and control algorithms consistent with the CBTC system of the target line are loaded, and a train dynamics model is configured based on the train vehicle parameters in the target line to obtain an initial digital twin decision sandbox. The real-time operating status of the target line is mapped to the initial digital twin decision sandbox to obtain the digital twin decision sandbox.

7. An auxiliary dispatching system for rail transit, characterized in that, The system is connected to the Automatic Train Control (ATS) system of the target line and is used to execute the auxiliary scheduling method for rail transit as described in any one of claims 1 to 5.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the auxiliary scheduling method for rail transit as described in any one of claims 1 to 5.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the auxiliary scheduling method for rail transit as described in any one of claims 1 to 5.

Images