Rail transit rapid scheduling method and system

Abstract

The present application relates to the technical field of rail transit scheduling, and discloses a rail transit rapid scheduling method and system, which comprises the following steps: collecting multi-dimensional data through real-time interaction between a vehicle-mounted sensor and a ground dispatching center, establishing a passenger flow change model to predict passenger flow density data; taking the multi-dimensional data as a data basis, generating a rail transit marshalling adjustment instruction through a scheduling unit, calling a previous vehicle power and a front road condition from a traffic database, and outputting a correction instruction of the rail transit marshalling adjustment instruction through a distribution model; and inputting the correction instruction into a hybrid verification unit for edge correction, optimizing the correction instruction to optimize the rail transit marshalling adjustment instruction, thereby achieving rail transit and greatly improving the accuracy of the scheduling rail transit result through double-layer correction.

Description

Technical Field

[0001] This invention relates to the technical field of rail transit scheduling, and discloses a rapid rail transit scheduling method and system. Background Technology

[0002] Current rail transit train formation scheduling mainly relies on physical coupling and fixed formation modes, which have significant drawbacks. Traditional formation requires decision-making through ground interlocking equipment and manual operation, which takes several minutes and cannot meet the surge in capacity demand during peak hours. Existing methods are based on fixed timetables and static priority strategies, which lack the ability to dynamically adapt to sudden changes in passenger flow and equipment health status, resulting in high empty load rates during off-peak hours. Physical formation relies on fixed traction and braking parameters, making it difficult to balance safety intervals and energy consumption constraints in emergency scenarios. Summary of the Invention

[0003] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions have been made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions are not intended to limit the scope of the invention.

[0004] To address the aforementioned technical problems, the main objective of this invention is to provide a rapid train scheduling method, comprising:

[0005] By collecting multi-dimensional data in real time through on-board sensors and ground dispatch center, a passenger flow change model is established to predict passenger flow density data.

[0006] Based on multidimensional data, the train formation adjustment instructions are generated through the formation unit. The power of the preceding vehicle and the road conditions ahead are retrieved from the traffic database, and the correction instructions for the train formation adjustment instructions are output through the distributed model.

[0007] The correction command is input into the hybrid verification unit for edge correction, and the correction command is optimized to optimize the rail transit train formation adjustment command.

[0008] As a preferred embodiment of the rapid scheduling method for rail transit according to the present invention, wherein:

[0009] The multidimensional data includes passenger flow density, rail transit energy consumption, and equipment health status.

[0010] The energy consumption of the rail transit and the health status of the equipment are used to set a priority strategy, which prioritizes the power demand and equipment health.

[0011] The priority strategy includes peak hours, off-peak hours, and valley hours.

[0012] As a preferred embodiment of the rapid scheduling method for rail transit according to the present invention, wherein:

[0013] The passenger flow change model includes a data processing layer, an input layer, a feature extraction layer, a fusion prediction layer, and a hybrid loss layer;

[0014] The feature extraction layer includes spatiotemporal feature fusion, which obtains passenger flow and passenger flow transfer features between stations through graph convolutional networks, and enhances feature identification for key time periods;

[0015] The feature extraction layer also includes multimodal feature fusion, which includes a multi-channel input structure to encode passenger flow and time features and sort the encoding results;

[0016] The fusion prediction layer is configured with a bottom layer to capture short-term passenger flow fluctuations and a top layer to learn long-term passenger flow trends and periodicity models. The fusion prediction layer also includes an enhanced attention mechanism to focus on the impact of historical periods on the next moment and to enhance the contribution of feature dimensions.

[0017] The passenger flow change prediction model is set up with a three-terminal parallel output layer, which includes a peak prediction terminal, an off-peak prediction terminal, and a valley-peak prediction terminal.

[0018] As a preferred embodiment of the rapid scheduling method for rail transit according to the present invention, wherein:

[0019] The orchestration unit includes a multi-source data interface, dynamic decision-making, and instruction verification.

[0020] The multi-source data interface receives peak, off-peak, and valley peak prediction values ​​and real-time carriage passenger volume data from the three-terminal parallel output layer of the passenger flow change model, as well as the vehicle power parameters and road condition information ahead retrieved from the traffic database.

[0021] Dynamic decision-making is based on multi-dimensional data to build an elastic allocation rule base, which includes priority strategies and verification of the matching between group length and platform capacity.

[0022] Command verification verifies the feasibility of commands by simulating the grouping and adjusting command execution effects, whether the traction system load exceeds the limit after coupling, and generates visual alarms.

[0023] As a preferred embodiment of the rapid scheduling method for rail transit according to the present invention, wherein:

[0024] The dynamic decision-making implementation method includes,

[0025] The input feature weights are dynamically adjusted according to the time period type. If the grouping adjustment command conflicts with the road conditions ahead, an online game model is triggered to generate a compromise speed curve based on the power performance of the preceding vehicle and the braking capacity of the following vehicle.

[0026] If the station capacity is insufficient, train sets will be split up and stopping times will be dynamically allocated.

[0027] The rules for generating rail transit train formation adjustment instructions include generating mandatory association instructions during peak hours, ignoring the rail transit maintenance status, and calling up spare train formations; and generating downgraded operation instructions during off-peak hours.

[0028] As a preferred embodiment of the rapid scheduling method for rail transit according to the present invention, wherein:

[0029] The method for outputting correction instructions by the distribution model includes:

[0030] Based on the real-time grouping status data collected by the vehicle-mounted edge computing node, the grouping status command is dynamically adjusted by a correction factor.

[0031] The rail transit systems communicate through a distributed communication protocol to adjust operating parameters based on local decisions.

[0032] The distributed model also includes instruction isolation, which includes physical isolation, logical isolation, and temporal isolation.

[0033] The physical isolation allocates independent communication channels for instructions of different priorities;

[0034] The logical isolation establishes a sandbox environment in the vehicle controller, restricting hardware access permissions for non-safety-related commands and allowing only safety commands to directly control the traction and braking system.

[0035] The timing isolation adopts a time-triggered architecture, polling the instruction queue at fixed intervals to ensure that high-priority instructions are executed first, and that only instructions from the same control domain are processed within a single cycle.

[0036] As a preferred embodiment of the rapid scheduling method for rail transit according to the present invention, wherein:

[0037] The hybrid verification unit includes an input layer, a verification and optimization layer, an edge optimization layer, and an output layer;

[0038] The input layer receives correction instructions from the distributed model, real-time rail transit formation status data, and a safety threshold library.

[0039] The verification and optimization layer verifies whether the formation operation status meets the power matching, braking safety and energy consumption constraints by simulating the formation operation state after the execution of the correction command.

[0040] The edge optimization layer uses a constrained genetic algorithm with the objective function of minimizing risk score and energy consumption increment to generate an optimized correction instruction set, which includes group length adjustment priority and dynamic speed curve.

[0041] The output layer sends the optimized correction instructions to the vehicle controller and simultaneously updates the instruction execution records in the traffic database.

[0042] As a preferred embodiment of the rapid scheduling method for rail transit according to the present invention, wherein:

[0043] The edge optimization layer also collects historical optimization instruction execution result data to construct a feedback training set and dynamically adjusts risk weights and energy consumption weights.

[0044] If safety requirements conflict with energy-saving requirements, a tiered optimization strategy is adopted, prioritizing forced speed reduction when braking distance is insufficient, regardless of energy consumption limits.

[0045] If there are conflicts in the coordination of multiple groups, resources should be allocated evenly.

[0046] The edge optimization layer compares the metrics before and after the optimization command is executed in real time, generates a verification report, and if the metric deviation exceeds the tolerance threshold, it triggers the command rollback mechanism, automatically restores the grouping configuration to the previous stable state, sends a fault code to the cloud scheduling center, and initiates the manual intervention process.

[0047] As a preferred embodiment of the rapid scheduling system for rail transit according to the present invention, wherein:

[0048] The data acquisition module integrates the communication interface between the vehicle-mounted sensors and the ground dispatch center to collect multi-dimensional data in real time;

[0049] The passenger flow prediction module outputs the predicted probabilities of peak, off-peak, and valley periods, as well as passenger flow fluctuation characteristics.

[0050] The orchestration unit includes a multi-source data interface, dynamic decision-making, and instruction verification. The multi-source data interface receives passenger flow prediction results, the power parameters of the preceding vehicle, and the road conditions ahead. The dynamic decision-making generates grouping adjustment instructions and executes conflict resolution logic. The instruction verification generates visual alarms by simulating the execution effect of instructions.

[0051] The distribution model obtains the correction factor and corrects the command via communication.

[0052] The hybrid verification unit verifies the train's power matching, braking safety, and energy consumption constraints.

[0053] The edge optimization submodule uses a constrained algorithm to generate an optimization instruction set and dynamically adjusts the optimization weights through an adaptive learning module.

[0054] Feedback closed units compare the deviation of indicators before and after instruction execution, triggering a rollback mechanism or manual intervention process.

[0055] As a preferred embodiment of the rapid scheduling system for rail transit according to the present invention, wherein:

[0056] The edge optimization submodule integrates a learning model and dynamically updates the fitness function weights based on historical instruction execution data.

[0057] To resolve conflicts, when multiple train sets compete for the same coupling section, priority is ranked according to the remaining capacity of the onboard battery.

[0058] When safety requirements conflict with energy-saving requirements, forced speed reduction is used to meet braking distance constraints, ignoring energy consumption limits.

[0059] Communication isolation allocates high-priority channels in the sliced ​​network for emergency braking commands, sets up a sandbox environment in the onboard controller, restricts access to the traction braking system for non-safe commands, and only allows safe commands to directly control the hardware.

[0060] The beneficial effects of this invention are:

[0061] A compromise speed curve is generated by a distributed game model to resolve the conflict between the power of the preceding vehicle and the braking of the following vehicle. A priority strategy is flexibly combined with peak-hour forced coupling and valley-peak degraded operation to reduce the no-load rate. The hybrid verification unit achieves automatic speed reduction when the braking safety distance is insufficient by constraining energy consumption and risk.

[0062] In the event of a sudden surge in passenger flow, the train formation adjustment instructions generated by the formation unit are used to prioritize the allocation of passage rights for large train formations, shorten the response time for coupling of spare train formations, and achieve dynamic matching between the power performance of the preceding train and the braking capacity of the following train based on the distribution model.

[0063] A hybrid verification unit is set up to perform edge correction on the correction instructions. The accuracy of the track traffic scheduling results is greatly improved through the two-layer correction. Attached Figure Description

[0064] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used 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 those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0065] Figure 1 This is a flowchart of a rapid scheduling method for rail transit according to the present invention;

[0066] Figure 2 This is a diagram illustrating the application of a passenger flow change model in a rapid scheduling method for rail transit according to the present invention.

[0067] Figure 3 This is a diagram illustrating the working method of a game theory model for a rapid scheduling method for rail transit according to the present invention.

[0068] Figure 4This is a monitoring diagram of the marshalling system for a rapid marshalling method for rail transit according to the present invention;

[0069] Figure 5 This is a curve simulation diagram of real-time monitoring and prediction of passenger flow at Lujiazui Station among multiple stations in a rapid scheduling method for rail transit according to the present invention.

[0070] Figure 6 This is a curve simulation diagram of the real-time monitoring and prediction of passenger flow at the Medieval Avenue Station, a multi-station rapid scheduling method for rail transit according to the present invention. Detailed Implementation

[0071] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0072] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0073] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0074] Example 1

[0075] like Figure 1 As shown, a rapid scheduling method for rail transit includes:

[0076] By interacting in real time with the ground dispatch center through onboard sensors, multi-dimensional data is collected, and a passenger flow change model is established to predict passenger flow density data.

[0077] Furthermore, the multidimensional data includes passenger flow density, rail transit energy consumption, and equipment health status;

[0078] Passenger flow density is obtained through platform cameras and on-board infrared counters; rail transit energy consumption is obtained through traction system current monitoring; and equipment health is obtained through bearing temperature sensors and vibration monitoring.

[0079] Platform cameras use target detection algorithms to count the number of people entering and exiting the station in real time, while onboard infrared counters monitor the distribution of passengers in the carriages. Passenger flow density is cross-verified by combining data from turnstile card swipes.

[0080] False detections caused by changes in lighting or occlusion are eliminated, and the deviation between infrared counting and video statistics results is dynamically calibrated.

[0081] The rail transit system's own traction system current monitoring module collects the current values ​​of the train during acceleration, constant speed, and braking in real time. It combines the voltage data to calculate the instantaneous energy consumption. When the high current lasts for more than a threshold, it is marked as abnormal energy consumption, triggering a health check of the equipment.

[0082] The bearing temperature sensor and vibration monitoring module collect the status of mechanical components in real time, and generate an early warning signal when the temperature is abnormal or the vibration frequency changes abruptly.

[0083] The multidimensional data includes passenger flow density, rail transit energy consumption, and equipment health status.

[0084] The energy consumption of the rail transit and the health status of the equipment are used to set a priority strategy, which prioritizes the power demand and equipment health.

[0085] The priority strategy includes peak hours, off-peak hours, and valley hours.

[0086] The passenger flow change model includes a data processing layer, an input layer, a feature extraction layer, a fusion prediction layer, and a hybrid loss layer.

[0087] The data processing layer is used to fill in missing data with the average of nearby time periods, align passenger flow, energy consumption, and equipment data by timestamp, and construct a unified spatiotemporal data matrix.

[0088] The feature extraction layer includes spatiotemporal feature fusion, which obtains passenger flow and passenger flow transfer features between stations through graph convolutional networks, and enhances feature identification for key time periods;

[0089] Furthermore, the spatiotemporal feature fusion constructs a site topology graph using a graph convolutional network. With sites as nodes and edge weights representing historical passenger flow correlation, it extracts cross-site passenger flow transfer features. For example, during the morning rush hour, node A represents concentrated travel from the residential area to the commercial area at node B, which represents the passenger flow transfer features from node A to node B.

[0090] like Figure 2 As shown, the application methods of the passenger flow prediction model include:

[0091] The site's association weight is dynamically increased during holidays and concert periods to amplify the impact of special events.

[0092] The feature extraction layer also includes multimodal feature fusion, which includes a multi-channel input structure to encode passenger flow and time features and sort the encoding results;

[0093] The fusion prediction layer is configured with a bottom layer to capture short-term passenger flow fluctuations and a top layer to learn long-term passenger flow trends and periodicity models. The fusion prediction layer also includes an enhanced attention mechanism to focus on the impact of historical periods on the next moment and to enhance the contribution of feature dimensions.

[0094] The passenger flow change prediction model is set up with a three-terminal parallel output layer, which includes a peak prediction terminal, an off-peak prediction terminal, and a valley-peak prediction terminal.

[0095] Furthermore, a preferred technical solution, including the passenger flow change model, also includes:

[0096] The fusion prediction layer includes short-term fluctuation capture at the bottom layer, long-term trend learning, and parallel output from three ends.

[0097] Furthermore, short-term fluctuations are captured by analyzing the underlying network to track real-time passenger flow changes, including sudden surges in passenger flow during special holidays and smaller passenger flow on weekdays. Transient fluctuation patterns are identified, such as sudden drops in passenger flow at subway entrances due to temporary flow restrictions and sudden increases in passenger flow at tourist attractions during statutory holidays.

[0098] Long-term trend learning models the periodic patterns of high-level networks and combines attention mechanisms to focus on similar historical periods, thereby enhancing prediction robustness.

[0099] Periodic patterns include a delayed peak hour on Friday evenings and a relatively even distribution of passenger flow at stations from Tuesday to Thursday evenings.

[0100] Historically similar time periods include a month of similar site traffic distribution excluding holidays, concerts, etc., and a week of similar sites.

[0101] The three terminals output peak forecasts in parallel: the peak forecast terminal outputs the probability of future passenger flow overload; the off-peak forecast terminal predicts the range of normal passenger flow fluctuations; and the valley peak forecast terminal identifies low passenger flow periods and triggers energy-saving operation.

[0102] Furthermore, the hybrid loss layer sets a loss function, which iterates the mean square error between the predicted and actual passenger flow for each station at multiple time periods for one week, and outputs the loss value of the predicted value. This corrects and compensates for short-term fluctuations and long-term trend learning. For example, if the long-term trend learning predicts a sudden increase in passenger flow at station y during time z in the second week of month x, less passenger flow at station y during time z in the third week of month x than at station y during time z in the second week of month x, and less passenger flow at station y during time z in the fourth week of month x than at station y during time z in the second week of month x, then the loss of the predicted value at station y during time z in the second week of month x is used to compensate for the sudden situation at station y during time z in the second week of month x.

[0103] If the rail transit scheduling system receives a special event, such as a concert being held at station y during time z in the second week of month x, the weight of stations associated with station y will be dynamically increased. The attention weight of station y and its associated stations will be increased through a passenger flow change model, thereby raising the threshold of the passenger flow prediction range.

[0104] The orchestration unit includes a multi-source data interface, dynamic decision-making, and instruction verification.

[0105] The multi-source data interface receives peak, off-peak, and valley peak prediction values ​​and real-time carriage passenger volume data from the three-terminal parallel output layer of the passenger flow change model, as well as the vehicle power parameters and road condition information ahead retrieved from the traffic database.

[0106] The multi-source data interface module constructs a standardized channel for heterogeneous data and uses multi-source information fusion technology to solve the problem of data format differences.

[0107] The data acquisition layer deploys a lightweight data agent program to receive peak, off-peak, valley, and peak time prediction values ​​output by the passenger flow prediction model and real-time passenger volume data reported by vehicle sensors in real time. The protocol conversion engine converts heterogeneous data such as bus signals and consultation messages into the same format.

[0108] A two-tiered structure is established, consisting of a real-time data cache and a historical data warehouse. The power parameters of the preceding vehicle are stored in a time-series database, while road condition information is stored in a spatial topology database.

[0109] Dynamic decision-making is based on a flexible allocation rule base constructed from multi-dimensional data. The rule base includes priority strategies and verification of the matching between group length and platform capacity.

[0110] Command verification verifies the feasibility of commands by simulating the grouping and adjusting command execution effects, whether the traction system load exceeds the limit after coupling, and generates visual alarms.

[0111] The dynamic decision-making implementation method includes,

[0112] The input feature weights are dynamically adjusted according to the time period type. If the grouping adjustment command conflicts with the road conditions ahead, an online game model is triggered to generate a compromise speed curve based on the power performance of the preceding vehicle and the braking capacity of the following vehicle.

[0113] If the station capacity is insufficient, train sets will be split up and stopping times will be dynamically allocated.

[0114] The rules for generating rail transit train formation adjustment instructions include generating mandatory association instructions during peak hours, ignoring the rail transit maintenance status, and calling up spare train formations; and generating downgraded operation instructions during off-peak hours.

[0115] A specific implementation method for a dynamic strategy includes:

[0116] The system receives peak, off-peak, and valley peak predictions and real-time carriage passenger volume data from the three-terminal parallel output layer via a multi-source data interface. During peak hours and at stations, the system aims to maximize capacity by forcibly coupling spare train sets, shortening departure intervals to the limit, and ignoring some equipment maintenance alarms, including air conditioning filter cleaning and passenger window cleaning. During off-peak hours and at stations, the system prioritizes energy conservation by splitting train sets to reduce empty load rates, extending departure intervals, and shutting down unnecessary onboard equipment. During valley peak hours, some trains are suspended for maintenance, while core services are retained, maintaining only basic services.

[0117] like Figure 5 The figures shown are the actual passenger flow curve and the predicted passenger flow curve at Lujiazui Station at 4 PM.

[0118] like Figure 6 The figures shown are the actual passenger flow detection curve and the predicted passenger flow curve at 4 PM on Century Avenue. Figure 5 and Figure 6 The passenger flow monitoring and prediction charts for different stations at the same time show that the actual passenger flow at Lujiazui was 1,000, while the predicted passenger flow was 900, with an error of less than 10%. The actual passenger flow at Century Avenue was 900, while the predicted passenger flow was 880, with an error of less than 10%.

[0119] Furthermore, dynamic decision-making is based on a hybrid decision-making mechanism that combines rule-based reasoning and case-based reasoning to achieve multi-objective optimization.

[0120] The system automatically matches peak, off-peak, and valley time periods by parsing timestamps, sets up a three-level weighting system with safety weighting greater than efficiency weighting greater than energy consumption weighting, establishes a three-dimensional platform model to calculate effective parking time, and dynamically matches train length with platform dwell time.

[0121] Input features are dynamically weighted. During peak hours, the weight of passenger flow features at major stations is increased. During off-peak hours, the weight of features at each station is balanced. During valley and peak hours, the rail transit scheduling focuses on energy consumption features.

[0122] Furthermore, dynamic decision-making also includes a conflict detection mechanism. Conflict monitoring uses topology analysis algorithms to detect logical conflicts between the train adjustment path and the status of the forward signal and track occupancy.

[0123] A specific implementation method for topology analysis algorithm detection:

[0124] The dynamic path topology graph is constructed based on the electronic map of the rail transit network. A dynamic path topology graph is established, where nodes represent the endpoints of rail segments, and edge weights include attributes such as segment length, gradient, and speed limit. Real-time access to road condition information ahead is used, and the node status is stored and dynamically updated through a graph database.

[0125] Based on the traction power and acceleration of the preceding vehicle and the braking distance and deceleration rate of the following vehicle, the time the following vehicle occupies the preceding section at the current speed is obtained. If the predicted occupancy time window of the following vehicle overlaps with the red light period of the preceding signal or the track occupancy period, a conflict marker is triggered. The conflict marker divides the conflict into hard conflict and soft conflict. Hard conflict requires immediate replanning, while soft conflict can be adjusted with a lag.

[0126] The conflict resolution and replanning process utilizes an online game model to generate a compromise speed curve based on the power curve of the preceding train and the braking curve of the following train. This ensures that the following train decelerates to a safe interval before entering the conflict zone. Through onboard edge nodes, the train communicates with adjacent trains to negotiate the release of turnout control or the adjustment of stopping positions, prioritizing the needs of high-priority trains.

[0127] When insufficient power is detected in the vehicle ahead, a braking traction power balance equation is established, and a compromise speed curve is obtained through iterative calculation. A specific implementation method for the braking traction power balance equation is as follows:

[0128] Data is collected on the braking of the front and rear trains in the rail transit system. The braking parameters of the front train include the traction motor current, voltage, and speed. The real-time traction power is calculated as P1 = I × V × η, where η is the motor efficiency, I is the current, and v is the voltage.

[0129] The following parameters are collected: regenerative braking current, braking resistor temperature, and wheel speed. The real-time braking power P2 = k × I × ω is calculated, where k is the energy conversion coefficient and ω is the braking efficiency.

[0130] Power weights are adjusted according to time period type, with priority given to traction power during peak hours and regenerative braking energy recovery during off-peak hours.

[0131] The maximum permissible speed that satisfies power balance is calculated iteratively using the Newton-Raphson method. A compromise speed curve is then generated to ensure dynamic matching between the traction power of the leading vehicle and the braking power of the trailing vehicle. If the current speed is greater than the maximum permissible speed, the priority of regenerative braking is increased, and more braking power is allocated to the trailing vehicle. If the current speed is less than the maximum permissible speed, the leading vehicle is allowed to moderately increase its traction.

[0132] Specifically, the Newton-Raphson method is a numerical method for solving the roots of equations. The principle is to approximate the roots of the equation using the tangent line of the function. An initial point is selected, and the intersection of the tangent lines is calculated by the derivative of the function, gradually approximating the root. It is used to solve simple roots, repeated roots, and complex roots. In this application, a compromise speed curve is used.

[0133] The method for outputting correction instructions by the distribution model includes:

[0134] Grouping status data collected in real time from vehicle-mounted edge computing nodes;

[0135] The rail transit systems communicate through a distributed communication protocol to adjust operating parameters based on local decisions.

[0136] The distributed model also includes instruction isolation, which includes physical isolation, logical isolation, and temporal isolation.

[0137] The physical isolation allocates independent communication channels for instructions of different priorities;

[0138] The logical isolation establishes a sandbox environment in the vehicle controller, restricting hardware access permissions for non-safety-related commands and allowing only safety commands to directly control the traction and braking system.

[0139] The timing isolation adopts a time-triggered architecture, polling the instruction queue at fixed intervals to ensure that high-priority instructions are executed first, and that only instructions from the same control domain are processed within a single cycle.

[0140] The correction factor calculation is based on an edge computing framework to achieve localized real-time decision-making. Through real-time dynamic allocation of rail transit scheduling instructions, it enables adaptive adjustment of rail transit scheduling instructions.

[0141] Lightweight data acquisition agents are deployed on vehicle-mounted edge nodes to obtain real-time data on rail transit train power output, braking pressure, and passenger distribution.

[0142] Data noise is eliminated by sensor fusion technology, and the extreme values ​​and trends of group power output, braking pressure, and passenger distribution are statistically analyzed using a sliding time window.

[0143] By comparing the current status with the safe operating traction power threshold and braking response time standard, parameters exceeding the deviation limit are identified, and the correction weights of each parameter are automatically adjusted according to the fault level and operating environment.

[0144] The weights of power output, braking pressure, and passenger distribution parameters are dynamically adjusted, and a set of correction instructions is generated based on the adjustment results. This includes traction compensation coefficient, braking timing offset, and passenger flow, which leads to an increase in the number of rail transit train formations.

[0145] In a preferred embodiment, the rapid scheduling of rail transit can also be configured with a distributed communication protocol architecture, constructing a hierarchical communication system to achieve multi-node collaborative decision-making and ensure command consistency. High-priority commands are allocated low-latency channels with dedicated bandwidth, while regular commands use multicast communication channels. When a high-priority command is detected, the buffered data of the low-priority channel is automatically cleared.

[0146] Furthermore, the traction braking control commands of the material isolation layer are transmitted through an independent bus and are physically isolated from the Ethernet channel of the vehicle information system. The safety-critical controller adopts dual independent power supply, and automatically switches to the supercapacitor backup power supply when the main power supply fails.

[0147] Furthermore, in the sandbox environment of the logical isolation layer, insecure instructions run within the virtualized container and can only access the mapped I / O ports; secure instructions directly access the control registers through hardware-level memory protection; when the vehicle controller starts up, it loads a whitelist policy, authorizing only instruction sets that have passed security authentication to access the underlying hardware; during operation, it monitors abnormal access attempts through a behavior analysis engine and triggers instruction execution interruption.

[0148] Furthermore, the timing isolation layer divides the control cycle into fixed time slots, assigns high-priority instructions to the first time slot, and prohibits concurrent execution of instructions across control domains within the same cycle.

[0149] The instruction queue adopts a double-buffered queue structure. The foreground queue executes the instructions of the current cycle, while the background queue preloads the instructions of the next cycle. When an instruction execution timeout is detected, the watchdog is immediately triggered to reset and roll back to a safe state.

[0150] The hybrid verification unit includes an input layer, a verification and optimization layer, an edge optimization layer, and an output layer;

[0151] The input layer receives correction instructions from the distributed model, real-time rail transit formation status data, and a safety threshold library.

[0152] The verification and optimization layer verifies whether the formation operation status meets the power matching, braking safety and energy consumption constraints by simulating the formation operation state after the execution of the correction command.

[0153] In this application, a preferred embodiment includes:

[0154] Among them, in terms of power matching, the total traction force of the train group is greater than or equal to the traction force required for the current gradient multiplied by the safety factor;

[0155] In braking safety, emergency braking distance ≤ distance to the obstacle ahead - redundancy distance;

[0156] In energy consumption constraints, the energy consumption increment of a single adjustment command is less than or equal to the remaining capacity of the current energy storage module.

[0157] The edge optimization layer uses a constrained genetic algorithm with the objective function of minimizing risk score and energy consumption increment to generate an optimized set of correction instructions, including priority adjustment of group length and dynamic speed curve.

[0158] The output layer sends the optimized correction instructions to the vehicle controller and simultaneously updates the instruction execution records in the traffic database.

[0159] In this application, a specific implementation method of a hybrid verification unit includes:

[0160] The input layer constructs a multi-source data fusion channel to achieve precise alignment between the instruction stream and the state stream.

[0161] Data access and cleaning converts the correction instructions output by the distributed model, real-time data from rail transit vehicle sensors, and the rail transit equipment safety threshold library into standardized data frames.

[0162] Hardware timestamps are used to mark each data source, and a sliding window compensation mechanism is used to eliminate timing misalignment caused by network transmission delay.

[0163] A three-level threshold tree is established based on the rail transit power system, braking system, and energy consumption system. When equipment aging or maintenance is detected, the emergency threshold branch is automatically loaded.

[0164] Further verification and optimization layers are implemented based on a digital twin virtual execution environment to perform instruction pre-verification.

[0165] The power matching verification involves inputting the traction force parameters from the correction command, calculating the theoretical acceleration curve, comparing it with the rail transit power system parameters, and outputting the power matching verification result.

[0166] The load matching detection compares the dynamic relationship between theoretical acceleration and actual train weight and track gradient. If the limit is exceeded, an alarm code is triggered.

[0167] Braking safety verification is based on the current speed, braking pressure correction value, and track adhesion coefficient. The complete stopping distance is calculated and the position information of obstacles ahead is superimposed. When the braking endpoint is predicted to intrude into the danger zone, a braking compensation command is generated.

[0168] Energy consumption constraint analysis simulates the energy flow of the traction and braking system based on the speed curve in the correction instruction, predicts the energy consumption increment per unit mileage, and marks the instruction as "optimizable" and transfers it to the edge optimization layer if the predicted energy consumption exceeds the threshold and there is no safety risk.

[0169] Furthermore, the edge optimization layer integrates multi-objective optimization evolutionary algorithms to achieve instruction set re-optimization.

[0170] The modified instruction set is converted into a gene sequence, with group length = gene segment 1, speed curve = gene segment 2, and passenger flow prediction = gene segment 3. A death penalty strategy is adopted, and gene segments that violate safety thresholds and have high risk scores are directly eliminated and replaced with other instructions.

[0171] The evolutionary optimization process uses the original correction instructions as seed individuals and generates an initial population through random perturbation.

[0172] The grouping length adjustment adopts single-point crossover, and each gene segment is individually verified to ensure that the initial gene sequence is correct. Legal grouping combinations are retained, and Gaussian noise is introduced to modify the gene sequence transformed by the instruction set. The results are repeatedly verified, and finally, the individual with the highest fitness in each generation is retained and directly enters the next generation.

[0173] Establish a risk-energy consumption two-dimensional coordinate system, eliminate dominated solutions, and select the optimal solution based on the current rail transit operation mode:

[0174] If the passenger flow forecast is high, the safety priority mode is triggered, and the solution with the lowest risk score is selected.

[0175] If passenger flow is predicted to be low, the economic mode is triggered, and the solution with the minimum energy consumption increment is selected.

[0176] Furthermore, the output layer constructs an instruction lifecycle management system, digitally signs and modifies the instruction set ID of the optimized instruction set, establishes an instruction traceability chain, synchronously issues key control instructions through a redundant CAN bus, and transmits non-real-time instructions such as group length adjustment plans in batches over the network.

[0177] Furthermore, the instruction correction is stored in the instruction database, and a time-series database is used to record the instruction issuance time, execution result, and device response curve. The actual execution data is compared with the prediction model to automatically calibrate the simulation parameters of the verification layer.

[0178] In this application, a preferred implementation of the edge optimization layer includes, for example, a learning unit, a conflict optimization unit, and a feedback unit.

[0179] The learning unit collects execution result data of historical optimization instructions, constructs a feedback training set, and dynamically adjusts risk weights and energy consumption weights.

[0180] The learning unit constructs a hierarchical online incremental learning framework. The bottom layer establishes a real-time data access interface to support bidirectional communication with trackside sensors, on-board controllers, and cloud databases. The middle layer builds an incremental learning engine to extract features from newly input execution result data and automatically identify changes in data patterns. The top layer designs a strategy output interface to establish a bidirectional interactive channel with the conflict optimization unit and feedback unit to collect and store historical optimization instruction execution data in real time.

[0181] The established historical data storage module uses a time-series database to store optimized instruction sets and execution results.

[0182] By using a weight distribution statistical database, the historical risk weight and energy consumption weight values ​​are analyzed at fixed intervals. When continuous safety alarms are detected, the weight adjustment process is automatically initiated: the risk weight is gradually increased based on the current risk weight, while the downward adjustment of the energy consumption weight is frozen, until no new safety alarms are detected for N consecutive cycles, after which the normal adjustment mechanism is restored.

[0183] If safety requirements conflict with energy-saving requirements, a hierarchical optimization strategy is adopted through the conflict optimization unit to prioritize the forced speed reduction when the braking distance is insufficient, ignoring energy consumption limitations.

[0184] If there are conflicts in the coordination of multiple groups, resources should be allocated evenly.

[0185] The feedback unit is used to compare the indicators before and after the execution of the optimization instruction in real time, generate a verification report, and if the indicator deviation exceeds the tolerance threshold, trigger the instruction rollback mechanism, automatically restore the group configuration to the previous stable state, send a fault code to the cloud scheduling center, and initiate the manual intervention process.

[0186] A specific implementation of an edge optimization layer:

[0187] The learning unit constructs an online incremental learning framework to realize the dynamic evolution of optimization strategies and balance the weight allocation of safety and energy efficiency.

[0188] Storage optimization instruction set and execution results, including site passenger flow forecast, actual risk value, energy consumption value, equipment response latency, etc.

[0189] The command parameters are discretized and encoded, and the adjustment range of the rail transit train length is mapped to three levels: small, medium, and large, and then associated with the passenger flow density prediction.

[0190] Successful instructions with actual risk scores lower than predicted values ​​and energy consumption in compliance with standards are selected to strengthen the current strategy. Failed instructions that cause equipment overload or emergency braking are captured and marked as high-risk scenario features.

[0191] The system analyzes the historical distribution of risk weights and energy consumption weights, automatically increases the risk value when continuous safety alarms are detected, identifies the current operating mode, and prohibits the learning unit from adjusting weights in emergency situations.

[0192] The conflict optimization unit uses a hierarchical decision tree to resolve multi-objective conflicts, and also includes multi-group resource conflict resolution.

[0193] The hierarchical decision tree calculates the difference ΔS between the theoretical braking distance and the actual available braking range in real time: when ΔS<0, a forced speed reduction command is immediately generated to cover the original speed curve; the speed reduction magnitude is dynamically calculated based on the absolute value of ΔS, and the speed is reduced to the minimum safe speed threshold, and the energy consumption limit release protocol is triggered, allowing the traction system to operate at over-rated power.

[0194] To resolve resource conflicts among multiple train sets, a train set resource competition matrix is ​​established to quantify the demand intensity of each train set for track, passenger flow density, and platform resources. The maximum-minimum fair allocation principle is adopted: the minimum resource guarantee for all train sets is determined; the remaining resources are dynamically allocated according to demand priority. During peak hours, train sets have doubled permissions, and train sets that are delayed for more than 2 minutes are automatically given higher priority and forcibly inserted into the scheduling queue.

[0195] Furthermore, the feedback unit constructs a two-way verification channel for the execution effect of instructions, enabling rapid isolation and recovery from abnormal states.

[0196] The indicator deviation calculation collects data in real time after execution and compares it item by item with the predicted value of the optimization instruction. The moving average algorithm is used to eliminate instantaneous indicator deviation.

[0197] Furthermore, a three-level alarm system is established to output the three-level deviation comparison results and determine whether to trigger command rollback.

[0198] Level 1 deviation is logged without intervention; Level 2 deviation triggers local parameter fine-tuning; Level 3 deviation immediately freezes control output.

[0199] The instruction rollback includes calling the historical configuration library to load the parameters of the previous stable state; issuing a recovery instruction through the safe channel to simultaneously reset the vehicle controller; locking the automatic optimization function and switching to the preset safe operation mode.

[0200] The fault handling process automatically generates a fault code package containing timestamps, deviation data, and environmental parameters; it is then uploaded to the dispatch center via a dedicated channel, triggering a manual console wake-up protocol.

[0201] In the event of an on-site emergency, the vehicle system will activate a degradation control strategy, shutting down non-core loads; a delay notice will be issued through the passenger information system, and emergency evacuation guidelines will be activated.

[0202] By exchanging real-time data through a shared memory pool and employing an event-driven architecture to ensure millisecond-level response, the rail transit train formation optimization system can ultimately achieve autonomous evolution and safe operation in complex environments.

[0203] like Figure 3 As shown, the specific implementation methods of the game theory model further include:

[0204] The inputs to the game model include the preceding train formation information and the waiting train formation information. The preceding train formation information has the core objective of maintaining traction efficiency and must avoid speed decay caused by insufficient power. The following train formation information prioritizes ensuring braking safety and must reserve sufficient braking distance to prevent the risk of rear-end collisions. The dispatch center is responsible for the global priority strategy and dynamically allocates formations that adapt to the passenger flow resources at each station.

[0205] The game rules include conflict detection, dynamic game theory, and distributed communication.

[0206] Conflict monitoring includes both hard and soft conflicts;

[0207] Hard conflicts include overlapping rail transit train formations and completely conflicting time windows, which require immediate replanning.

[0208] Soft conflicts include partial resource competition and insufficient platform capacity, which allow for delayed adjustments or dynamic priority allocation.

[0209] Dynamic game theory includes velocity curve negotiation and resource competition arbitration.

[0210] The speed curve negotiation is a compromise speed curve generated by iterative calculation between the preceding and following vehicles based on real-time power-braking parameters.

[0211] The compromise speed curve includes speed curves for peak hours, off-peak hours, and valley hours.

[0212] During peak hours, priority is given to ensuring the traction power of the vehicle in front, while allowing the vehicle behind to increase its braking power to shorten the safe distance.

[0213] During off-peak, valley, and peak periods, emphasis should be placed on regenerative braking energy recovery from following vehicles to reduce overall energy consumption.

[0214] Furthermore, resource competition arbitration is used to split and group trains and dynamically allocate dwell time and departure intervals, prioritizing the needs of stations with high passenger flow.

[0215] Furthermore, distributed communication includes physical channel isolation and logical sandbox isolation, and polls the instruction queue at fixed intervals to ensure that high-priority instructions are executed first.

[0216] Physical channel isolation includes emergency braking commands, which are transmitted through an independent high-priority slice network with dedicated bandwidth.

[0217] Logical sandbox isolation includes non-safe instructions running in virtualized containers, while only safe instructions can directly control the traction and braking system.

[0218] like Figure 4 As shown,

[0219] The energy consumption monitoring line graph on the left panel displays two energy consumption indicators: the horizontal axis represents time, and the vertical axis represents energy consumption value.

[0220] The left panel displays the health status of the equipment: traction system: green indicates normal status; braking system: yellow indicates maintenance is required; door system: green indicates normal status.

[0221] The central area displays a map showing the train's operating area, marking key locations as virtual simulation locations, and showing the distribution of track lines and stations. All stations are identified by serial numbers 1 to 15, including A1 corresponding to station 1, A2 corresponding to station 7, and A3 corresponding to station 4.

[0222] The right panel includes a list of train dispatch information, Train 1 (T2024011801), Adjustment type: train formation split; Execution time, location, and status.

[0223] Train 2 (T2024011802) has completed its dynamic formation adjustment; execution time; location: Station A2;

[0224] Train 3 (T2024011803) Adjustment type: Emergency formation cancelled; Execution time; Location; Status.

[0225] The bottom status bar displays the system status: Normal operation; Normal data communication; 2 alarms pending processing.

[0226] Performance metrics show: CPU utilization: 45%; memory utilization: 60%; network latency: 25ms; data processing rate: 2.5MBs.

[0227] Energy consumption and equipment health are related:

[0228] Energy consumption monitoring data is used for dynamic grouping decisions, and grouping reduces energy consumption;

[0229] The braking system being in a "pending inspection" state affects the safety of emergency formation operations, and alarms must be handled with priority.

[0230] Scheduling operation logic:

[0231] The train split T2024011801 is pending execution at station A1, possibly as an energy-saving strategy during off-peak hours;

[0232] Dynamic train formation T2024011802 has been completed, and the train length has been adjusted to adapt to passenger flow.

[0233] Emergency formation T2024011803 has been cancelled due to conflict or equipment status.

[0234] Train locations A1 and A2 correspond to the areas marked on the map: Chapel Acres and South County.

[0235] The distance between locations is represented by scale values ​​of 500, 200, and 900, which are used to calculate the departure interval.

[0236] The system status and alarms show that CPU and memory usage are moderate and network latency is low at 25ms, indicating that the system load is normal and data processing efficiency meets the standards.

[0237] The pending alarms are related to brake system inspection, train group conflicts, or other abnormal events, and require further investigation.

[0238] Example 2

[0239] The data acquisition module integrates the communication interface between the vehicle-mounted sensors and the ground dispatch center to collect multi-dimensional data in real time;

[0240] The passenger flow prediction module outputs the predicted probabilities of peak, off-peak, and valley periods, as well as passenger flow fluctuation characteristics.

[0241] The orchestration unit includes a multi-source data interface, dynamic decision-making, and instruction verification. The multi-source data interface receives passenger flow prediction results, the power parameters of the preceding vehicle, and the road conditions ahead. The dynamic decision-making generates grouping adjustment instructions and executes conflict resolution logic. The instruction verification generates visual alarms by simulating the execution effect of instructions.

[0242] The distribution model obtains the correction factor and corrects the command via communication.

[0243] The hybrid verification unit verifies the train's power matching, braking safety, and energy consumption constraints.

[0244] The edge optimization submodule uses a constrained algorithm to generate an optimization instruction set and dynamically adjusts the optimization weights through an adaptive learning module.

[0245] The edge optimization submodule integrates a learning model and dynamically updates the fitness function weights based on historical instruction execution data.

[0246] The conflict resolution rules include prioritizing trains based on the remaining battery capacity when multiple train sets compete for the same coupling section.

[0247] When safety requirements conflict with energy-saving requirements, forced speed reduction is used to meet braking distance constraints, ignoring energy consumption limits.

[0248] Communication isolation allocates high-priority channels of the sliced ​​network to emergency braking commands, sets up a sandbox environment in the on-board controller, restricts access to the traction braking system by non-safe commands, and only allows safe commands to directly control the hardware.

[0249] The feedback closed unit compares the deviation of indicators before and after instruction execution, triggering a rollback mechanism or manual intervention process.

[0250] It is important to note that the constructions and arrangements of this application shown in several different exemplary embodiments are merely illustrative. Although only two embodiments are described in detail in this disclosure, those who consult this disclosure will readily understand that variations in the size, dimensions, structure, shape and proportion of various elements, as well as parameter values ​​(e.g., temperature, pressure, etc.), mounting arrangements, use of materials, color, orientation, etc., are possible without substantially departing from the subject matter described in this application. For example, an element shown as integrally formed may be composed of multiple parts or elements, the position of elements may be inverted or otherwise altered, and the nature or number or position of discrete elements may be changed or altered. Therefore, all such modifications are intended to be included within the scope of the invention. The order or sequence of any process or method steps may be changed or rearranged according to alternative embodiments. Any "device plus function" clause is intended to cover the structure performing the function described herein, and not only structurally equivalent but also equivalent in structure. Other substitutions, modifications, alterations, and omissions may be made in the design, operation, and arrangement of the exemplary embodiments without departing from the scope of the invention. Therefore, the present invention is not limited to the specific embodiments, but extends to various modifications that still fall within the scope of the appended claims.

[0251] Furthermore, in order to provide a concise description of exemplary embodiments, not all features of actual embodiments (i.e., those features that are not relevant to the currently considered best mode for carrying out the invention, or those features that are not relevant to implementing the invention) may be omitted.

[0252] It should be understood that numerous specific implementation decisions can be made during the development of any practical implementation, such as in any engineering or design project. Such development efforts are complex and time-consuming, but for those of ordinary skill in the art who benefit from this disclosure, the development effort will be a routine task in design, manufacturing, and production without requiring extensive experimentation.

[0253] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A rapid scheduling method for rail transit, characterized in that, include: By collecting multi-dimensional data in real time through on-board sensors and ground dispatch center, a passenger flow change model is established to predict passenger flow density data. Based on multidimensional data, the train formation adjustment instructions are generated through the formation unit. The power of the preceding vehicle and the road conditions ahead are retrieved from the traffic database, and the correction instructions for the train formation adjustment instructions are output through the distributed model. The correction command is input into the hybrid verification unit for edge correction, and the correction command is optimized to optimize the rail transit train formation adjustment command. The hybrid verification unit includes an input layer, a verification and optimization layer, an edge optimization layer, and an output layer; The input layer receives correction instructions from the distributed model, real-time rail transit formation status data, and a safety threshold library. The verification and optimization layer verifies whether the formation operation status meets the power matching, braking safety and energy consumption constraints by simulating the formation operation state after the execution of the correction command. Energy consumption constraint analysis simulates the energy flow of the traction and braking system based on the speed curve in the correction instruction, and predicts the energy consumption increment per unit mileage. If the predicted energy consumption exceeds the threshold and there is no safety risk, the instruction is marked as an optimizable state handover edge optimization layer. The edge optimization layer uses a constrained genetic algorithm with the objective function of minimizing risk score and energy consumption increment to generate an optimized correction instruction set, which includes group length adjustment priority and dynamic speed curve. The output layer sends the optimized correction instructions to the vehicle controller and simultaneously updates the instruction execution records in the traffic database.

2. The rapid scheduling method for rail transit according to claim 1, characterized in that: The multidimensional data includes passenger flow density, rail transit energy consumption, and equipment health status. The energy consumption of the rail transit and the health status of the equipment are used to set a priority strategy, which prioritizes the power demand and equipment health. The priority strategy includes peak hours, off-peak hours, and valley hours.

3. The rapid scheduling method for rail transit according to claim 1, characterized in that: The passenger flow change model includes a data processing layer, an input layer, a feature extraction layer, a fusion prediction layer, and a hybrid loss layer; The feature extraction layer includes spatiotemporal feature fusion, which obtains passenger flow and passenger flow transfer features between stations through graph convolutional networks, and enhances feature identification for key time periods; The feature extraction layer also includes multimodal feature fusion, which includes a multi-channel input structure to encode passenger flow and time features and sort the encoding results; The fusion prediction layer is configured with a bottom layer to capture short-term passenger flow fluctuations and a top layer to learn long-term passenger flow trends and periodicity models. The fusion prediction layer also includes an enhanced attention mechanism to focus on the impact of historical periods on the next moment and to enhance the contribution of feature dimensions. The passenger flow change prediction model is set up with a three-terminal parallel output layer, which includes a peak prediction terminal, an off-peak prediction terminal, and a valley-peak prediction terminal.

4. The rapid scheduling method for rail transit according to claim 1, characterized in that: The orchestration unit includes a multi-source data interface, dynamic decision-making, and instruction verification. The multi-source data interface receives peak, off-peak, and valley peak prediction values ​​and real-time carriage passenger volume data from the three-terminal parallel output layer of the passenger flow change model, as well as the vehicle power parameters and road condition information ahead retrieved from the traffic database. Dynamic decision-making is based on a flexible allocation rule base constructed from multi-dimensional data. The rule base includes priority strategies and verification of the matching between group length and platform capacity. Command verification verifies the feasibility of commands by simulating the grouping and adjusting command execution effects, whether the traction system load exceeds the limit after coupling, and generates visual alarms.

5. The rapid scheduling method for rail transit according to claim 4, characterized in that: The dynamic decision-making implementation method includes, The input feature weights are dynamically adjusted according to the time period type. If the grouping adjustment command conflicts with the road conditions ahead, an online game model is triggered to generate a compromise speed curve based on the power performance of the preceding vehicle and the braking capacity of the following vehicle. If the station capacity is insufficient, train sets will be split up and stopping times will be dynamically allocated. The rules for generating rail transit train formation adjustment instructions include generating mandatory association instructions during peak hours, ignoring the rail transit maintenance status, and calling up spare train formations; and generating downgraded operation instructions during off-peak hours.

6. The rapid scheduling method for rail transit according to claim 1, characterized in that: The method for outputting correction instructions by the distribution model includes: Based on the real-time grouping status data collected by the vehicle-mounted edge computing node, the grouping status command is dynamically adjusted by a correction factor. The rail transit systems communicate through a distributed communication protocol to adjust operating parameters based on local decisions. The distributed model also includes instruction isolation, which includes physical isolation, logical isolation, and temporal isolation. The physical isolation allocates independent communication channels for instructions of different priorities; The logical isolation establishes a sandbox environment in the vehicle controller, restricting hardware access permissions for non-safety-related commands and allowing only safety commands to directly control the traction and braking system. The timing isolation adopts a time-triggered architecture, polling the instruction queue at fixed intervals to ensure that high-priority instructions are executed first, and that only instructions from the same control domain are processed within a single cycle.

7. The rapid scheduling method for rail transit according to claim 1, characterized in that: The edge optimization layer also collects historical optimization instruction execution result data to construct a feedback training set and dynamically adjusts risk weights and energy consumption weights. If safety requirements conflict with energy-saving requirements, a tiered optimization strategy is adopted, prioritizing forced speed reduction when braking distance is insufficient, regardless of energy consumption limits. If there are conflicts in the coordination of multiple groups, resources should be allocated evenly. The edge optimization layer compares the metrics before and after the optimization command is executed in real time, generates a verification report, and if the metric deviation exceeds the tolerance threshold, it triggers the command rollback mechanism, automatically restores the grouping configuration to the previous stable state, sends a fault code to the cloud scheduling center, and initiates the manual intervention process.

8. A rapid train formation system, implemented based on any one of claims 1-7, characterized in that, The method includes the following specific steps: The data acquisition module integrates the communication interface between the vehicle-mounted sensors and the ground dispatch center to collect multi-dimensional data in real time; The passenger flow prediction module outputs the predicted probabilities of peak, off-peak, and valley periods, as well as passenger flow fluctuation characteristics. The orchestration unit includes a multi-source data interface, dynamic decision-making, and instruction verification. The multi-source data interface receives passenger flow prediction results, the power parameters of the preceding vehicle, and the road conditions ahead. The dynamic decision-making generates grouping adjustment instructions and executes conflict resolution logic. The instruction verification generates visual alarms by simulating the execution effect of instructions. The distribution model obtains the correction factor and corrects the command via communication. The hybrid verification unit verifies the train's power matching, braking safety, and energy consumption constraints. The edge optimization submodule uses a constrained algorithm to generate an optimization instruction set and dynamically adjusts the optimization weights through an adaptive learning module. Feedback closed units compare the deviation of indicators before and after instruction execution, triggering a rollback mechanism or manual intervention process.

9. A rapid train formation system for rail transit according to claim 8, characterized in that: The edge optimization submodule integrates a learning model and dynamically updates the fitness function weights based on historical instruction execution data. The conflict resolution rules include prioritizing trains based on the remaining battery capacity when multiple train sets compete for the same coupling section. When safety requirements conflict with energy-saving requirements, forced speed reduction is used to meet braking distance constraints, ignoring energy consumption limits. Communication isolation allocates high-priority channels in the sliced ​​network for emergency braking commands, sets up a sandbox environment in the onboard controller, restricts access to the traction braking system for non-safe commands, and only allows safe commands to directly control the hardware.

Images