A rail transit sudden passenger flow collaborative evacuation control method and control system based on a three-layer evacuation network

Abstract

This invention provides a collaborative evacuation control method and system for sudden passenger flow in rail transit based on a three-layer evacuation network, comprising the following steps: acquiring multi-source status data from underground, ground, and low-altitude sources; constructing a three-layer evacuation network including underground node-type evacuation units, ground section-type evacuation units, and low-altitude auxiliary guidance nodes, wherein low-altitude nodes only affect the status of ground units through coverage mapping and do not participate in path sequencing; performing risk normalization on multi-source parameters and calculating the comprehensive passage cost, and determining the risk level by integrating the cumulative path cost and local peak values; generating underground guidance, ground signal, and low-altitude task control parameters, and uniformly distributing them in a transactional manner after consistency verification; and writing back the low-altitude execution results as evacuation unit status parameters to trigger closed-loop reconstruction. This invention solves the problems of cross-system command conflicts, chaotic low-altitude resource modeling, and the masking of local bottlenecks by averaging, thereby improving the collaborative efficiency and safety of sudden passenger flow evacuation.

Description

Technical Field

[0001] This invention relates to the fields of urban rail transit emergency management, intelligent traffic control, and low-altitude unmanned aerial vehicle (UAV) assisted guidance technology, specifically to a collaborative evacuation control method and control system for sudden passenger flow in rail transit based on a three-layer evacuation network. Background Technology

[0002] Urban rail transit stations are characterized by relatively enclosed spaces, complex passageways, high passenger density, and close coupling between entrances / exits and ground roads. When large events end, trains are delayed, equipment malfunctions occur, fire alarms sound, severe weather occurs, or public safety emergencies occur, passenger flow within the station may surge towards the entrances / exits in a short period of time, forming secondary gatherings on roads outside the station, pedestrian crossings, or transfer nodes. If not handled promptly, this could lead to congestion, stampedes, or traffic disorder.

[0003] Existing rail transit emergency evacuation technologies primarily focus on station guidance, passenger flow monitoring, turnstile control, and public address systems; ground traffic control systems mainly focus on vehicle throughput efficiency and intersection signal timing. These two systems differ in data format, update frequency, controlled objects, and management entities, making synchronized responses difficult. Current solutions typically handle rail transit-side evacuation organization and ground signal control separately, lacking unified quantitative evaluation criteria and consistent control timing. This frequently leads to discrepancies between exit directions indicated by station guidance systems and ground traffic signal directions, causing passenger congestion at exits or even backflow, increasing safety risks.

[0004] Meanwhile, low-altitude drones and other equipment are increasingly being used for urban emergency monitoring, lighting, communication relay, and on-site inspections. Some low-altitude assisted evacuation schemes directly incorporate drone waypoints, hovering points, or temporary guidance points into the evacuation route map, leading to confusion between route nodes, controlled objects, and sensing resources, affecting the consistency of route sequencing. Other schemes only record the execution results of low-altitude equipment as completed or failed, failing to convert feedback parameters such as illumination improvement, communication restoration, changes in directional consistency, density changes, and speed changes into evacuation route status, local remaining traffic capacity, or on-site support status parameters. Therefore, these parameters cannot participate in the next cycle of evacuation route reconstruction and risk assessment, rendering the actual effectiveness of low-altitude guidance unusable for subsequent decision-making.

[0005] Furthermore, existing route assessments often focus on overall route length, average congestion, or total travel time, which can easily average out localized high-risk bottleneck areas and fail to simultaneously identify both overall route burden and local peak risks in key evacuation units. When severe congestion occurs at an exit or on a road segment, the overall average route cost may still be within an acceptable range, but the local peak risk may already threaten pedestrian safety. In multi-subsystem coordinated control, command conflicts may also occur between underground guidance systems, ground traffic signal systems, and low-altitude mission systems. For example, in-station guidance may point to an exit, but ground signals may not allow the corresponding crossing direction, or the low-altitude guidance coverage area may overlap with impassable areas, causing people to be guided to dangerous areas.

[0006] Existing low-altitude guidance schemes often treat drones as purely positive resource benefits, ignoring the potential negative impacts of guidance failures, incorrect directions, or equipment malfunctions. When the drone's guidance direction is inconsistent with the actual expected direction of the crowd, or when equipment failure leads to incorrect guidance, existing models lack a penalty mechanism for such anomalies. This makes the path cost model overly optimistic and unable to detect and correct incorrect guidance in a timely manner. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a collaborative evacuation control method and system for sudden passenger flow in rail transit based on a three-layer evacuation network. By constructing an underground, ground-level, and low-altitude evacuation network, it clarifies that low-altitude nodes serve only as state correction resources, not as path passage units, thus resolving the problem of inconsistent modeling standards. A unified risk-based normalization and weighted fusion method is employed to convert multi-source heterogeneous parameters into a comprehensive passage cost, providing a quantitative basis for cross-system collaboration. A two-way feedback mechanism is introduced, incorporating a low-altitude guidance correction factor and anomaly penalty term, enabling dynamic adjustment of passage costs based on the actual effectiveness (including success and failure) of low-altitude equipment. A regional assessment method, fusing path cumulative cost with local peak values ​​of key evacuation units, avoids local bottlenecks being masked by average values. Cross-subsystem consistency verification and transactional unified issuance are performed before command issuance, eliminating conflicts between station guidance, ground signals, and low-altitude tasks. The execution result state write-back drives the path reconstruction for the next cycle, forming a closed-loop control. This invention significantly improves the collaborative efficiency, safety, and system robustness of sudden passenger flow evacuation.

[0008] The present invention achieves the above-mentioned technical objectives through the following technical means.

[0009] A collaborative evacuation control method for sudden passenger flow in rail transit based on a three-layer evacuation network includes the following steps:

[0010] S01: Acquire multi-source status data of rail transit station entrances and exits, station hall passages, road sections outside the station, ground traffic signals and low-altitude auxiliary equipment, and determine whether a sudden passenger flow event has occurred based on the traffic risk status index; when the traffic risk status index meets the preset abnormal triggering conditions or receives an emergency triggering signal, enter the dynamic collaborative evacuation control mode;

[0011] S02: A three-layer evacuation network is constructed under the dynamic collaborative evacuation control mode. The three-layer evacuation network includes an underground evacuation layer, a ground evacuation layer, and a low-altitude auxiliary guidance layer. Among them, the key passages of rail transit entrances or station halls in the underground evacuation layer are mapped as node-type underground evacuation units, and the road segments outside the station in the ground evacuation layer are mapped as road segment-type ground evacuation units. The node-type underground evacuation units and road segment-type ground evacuation units serve as the calculation subjects for passage cost calculation, path sequencing, and risk assessment. The low-altitude auxiliary guidance nodes in the low-altitude auxiliary guidance layer do not serve as path passage units. They only affect the state correction of the road segment-type ground evacuation units within their coverage area through the coverage mapping relationship, and do not participate in passage cost calculation and path sequencing.

[0012] S03: Perform risk-based normalization on the multi-source state parameters of each evacuation unit and calculate the basic passage cost; based on the coverage mapping relationship between the low-altitude auxiliary guidance node and the road segment ground evacuation unit, calculate the low-altitude guidance correction factor and the low-altitude guidance anomaly penalty term, and obtain the comprehensive passage cost at the evacuation unit level accordingly.

[0013] S04: Based on the comprehensive travel cost of the nodal underground evacuation units and road segment ground evacuation units contained in the candidate evacuation routes, calculate the cumulative travel cost of the candidate evacuation routes, and combine the peak local travel cost of key evacuation units to obtain the overall travel cost index of the monitoring area, and determine the evacuation risk level and control level.

[0014] S05: Generate a control strategy parameter set based on the evacuation risk level, control level, candidate path priority sequence, path passability status, and execution resource status. The control strategy parameter set includes underground guidance parameters, ground traffic signal parameters, and low-altitude mission parameters.

[0015] S06: Map the control strategy parameter set to control commands for underground guidance systems, ground traffic signal systems, and low-altitude equipment platforms, and perform cross-subsystem consistency verification before issuing them; if the verification passes, the three types of control commands will carry the same control cycle identifier, strategy version number, and receipt identifier and be issued uniformly in the same effective cycle; if the verification fails, it will be handled according to the preset conflict handling rules.

[0016] S07: Receive control command execution receipts and field feedback results, write back the execution results of low-altitude guidance or support missions as evacuation unit status parameters, and recalculate passage costs and reconstruct candidate evacuation paths based on the written-back status parameters; when the closed-loop reconstruction triggering conditions, including time period triggering conditions or status change triggering conditions, are met, S02 to S06 are re-executed.

[0017] Furthermore, the traffic risk status index is obtained by aggregating the unit risk index of at least one evacuation unit within the monitoring area. The unit risk index is obtained by risk-normalizing and weighting at least two of the following: crowd density, queue length, average traffic speed, number of stranded people, and traffic saturation of ground road sections.

[0018] The preset abnormal triggering conditions include the passage risk status indicator exceeding the triggering threshold, or any one of the following indicators exceeding the corresponding threshold: crowd density at key entrances / exits, average passage speed decrease rate, queue length, or number of stranded people.

[0019] Furthermore, the multi-source status data includes data from the automatic fare collection system, video passenger flow recognition data, ground traffic signal control data, road traffic flow detection data, low-altitude equipment sensing data, and low-altitude equipment task execution feedback data. When there is a time discrepancy between different data sources, time synchronization is performed according to a unified timestamp, and fusion is performed based on equipment confidence, data freshness, and spatial coverage. When any multi-source status data is missing, delayed, or abnormal, alternative status parameters are generated based on the data source confidence, the most recent valid sample value, and the status estimate of adjacent evacuation units, and the status confidence of the corresponding evacuation unit is reduced.

[0020] Furthermore, the ground connection points, pedestrian crossings, intersections, exit directions, target areas, guide screen locations, and low-altitude equipment hovering points in the three-layer evacuation network are only used as topological connection objects, control objects, display objects, or location identifiers, and are not used as independent evacuation units to participate in the calculation of basic passage cost, path cumulative passage cost, overall passage cost index of the monitoring area, and risk level determination. For node-type underground evacuation units, at least two of the following should be selected as passage status characterization parameters: crowd density, queue length, average passage speed, and number of stranded people. For road segment-type ground evacuation units, at least two of the following should be selected as passage status characterization parameters: traffic flow saturation, remaining capacity ratio, degree of intersection conflict, road segment length, expected passage time, and path passability status.

[0021] Furthermore, the low-altitude guidance correction factor η q According to the set of low-altitude equipment D covering evacuation unit q q Spatial influence coefficient α q,d Low-altitude equipment efficiency index γ d Effective guided response coefficient ρq,d and available state coefficients s d Calculate, and adjust by the upper limit of the correction factor η max Amplitude limiting is applied:

[0022] ,

[0023] In the formula:

[0024] D q A collection of low-altitude equipment covering evacuation unit q;

[0025] s d This is the availability coefficient for low-altitude equipment, where 0 indicates unavailable and 1 indicates available.

[0026] η max To adjust the upper limit of the factor;

[0027] When multiple low-altitude auxiliary guidance nodes cover the same road segment ground evacuation unit, the optimal mode, amplitude-limiting overlay mode, or conflict suppression mode is selected based on the consistency of guidance direction, overlap of coverage areas, direction of ground traffic signal release, and equipment availability to determine the summation term. The specific calculation method.

[0028] Furthermore, the efficiency indicators of the low-altitude equipment It is obtained by weighting three factors: brightness, coverage area, and height effect.

[0029] ,

[0030] in:

[0031] This represents the comprehensive performance index of the low-altitude equipment d, with a value range of 0 to 1;

[0032] , , The weighting coefficients of the three indicators satisfy the following conditions: ;

[0033] This is the normalized value for the projected brightness or directional sign contrast of the low-altitude device d;

[0034] This is the normalized value of the area of ​​the effective guidance zone formed by the low-altitude equipment d on the ground;

[0035] Let d be the normalized value of the hovering height effect of the low-altitude equipment. The formula for calculating the value is:

[0036] ,

[0037] Where: H d H is the current hovering altitude of the low-altitude equipment. max and H min These are the upper and lower limits of the effective height range, respectively;

[0038] The spatial influence coefficient α of low-altitude equipment d on evacuation unit q q,d The calculation formula depends on the ratio of the ground projection distance to the effective guiding radius, as follows:

[0039] ,

[0040] In the formula: dist q,d The distance from the geometric center of the evacuation unit to the ground projection center of the low-altitude equipment;

[0041] R d The effective guiding radius of the low-altitude equipment d;

[0042] Effective guided response coefficient ρ q,d It is determined by a combination of factors including successful execution status, changes in directional consistency rate, density changes, and velocity changes.

[0043] ,

[0044] In the formula:

[0045] ρ q,d This represents the effective guidance response coefficient of the low-altitude equipment d to the evacuation unit q;

[0046] ACK d This indicates the successful execution status of the low-altitude device d. The value is 1 when the execution is successful, and 0 when the execution fails, times out, or the receipt is missing.

[0047] ΔDir q is the periodic change in the consistency rate between the actual movement direction of the crowd and the low-altitude guidance indication direction within the coverage area of ​​evacuation unit q.

[0048] ΔDen q Let q be the rate of change in population density within the coverage area of ​​evacuation unit q;

[0049] Let q be the rate of change in the average passage speed of the population within the coverage area of ​​evacuation unit q.

[0050] Furthermore, the low-altitude guidance anomaly penalty item This low-altitude guidance anomaly penalty term is generated when at least one of the following occurs: low-altitude assisted guidance node execution failure, timeout, missing feedback, coverage deviation, waypoint deviation, decreased direction consistency rate, increased local density, decreased travel speed, or deviation of the path selection ratio from the target path. It is used as a proportional penalty factor in the comprehensive travel cost calculation of the corresponding road segment-type ground evacuation unit. The calculation is as follows:

[0051] ,

[0052] In the formula:

[0053] This represents the abnormal penalty term generated by the low-altitude auxiliary guidance node for evacuation unit q;

[0054] F d This indicates that the low-altitude device d has failed to execute its task, timed out, or is missing feedback.

[0055] O q This indicates a coverage deviation or waypoint deviation status;

[0056] b1, b2, b3, b4, and b5 are coefficient weights.

[0057] Furthermore, the cumulative travel cost A of the candidate evacuation routes p According to the comprehensive traffic value C at the evacuation unit level q Length L of evacuation unit q Expected passage time, passage impedance or remaining passage capacity K q We obtain the following by weighted summation:

[0058] ,

[0059] Among them, P p L is the set of all evacuation units contained in path p; q K represents the length of the evacuation unit or the estimated travel time. q L represents the remaining capacity or capacity factor; in the same path calculation, L q and K q They are converted into dimensionless scale coefficients or uniform pass-through impedance coefficients of the same scale.

[0060] Furthermore, the overall traffic cost index G of the monitoring area is accumulated through candidate evacuation routes to determine the traffic cost A. p Peak local access costs for key evacuation units The result of fusion:

[0061] ,

[0062] In the formula:

[0063] β1 and β2 are the fusion weights, and β1 + β2 = 1;

[0064] A ref For reference, the cumulative cost value;

[0065] K represents the set of key evacuation units;

[0066] r is the total number of feasible paths in the candidate path set;

[0067] π p Let p be the fusion weight of the p-th candidate path in the overall regional cost;

[0068] G represents the overall traffic cost index for the monitored area, with a value ranging from 0 to 1.

[0069] The key evacuation units include node-type underground evacuation units associated with the exit direction, road segment-type ground evacuation units corresponding to cross-street connection bottlenecks, high-load evacuation units in the current main evacuation route, and evacuation units that rank in the top m in terms of comprehensive traffic value or exceed the p' quantile of historical safety samples in the current assessment period, where m and p' are preset parameters.

[0070] A system for a collaborative evacuation control method for sudden passenger flow in rail transit based on a three-layer evacuation network includes:

[0071] The multi-source data access unit is used to access data from rail transit automatic fare collection systems, video surveillance systems, ground traffic signal control systems, road traffic detection equipment, and low-altitude equipment platforms, and to perform time synchronization, spatial coordinate transformation, and data confidence assessment, and output multi-source status data.

[0072] The event determination and mode switching unit is used to receive the multi-source status data, determine whether a sudden passenger flow event has occurred based on the traffic risk status index, and output a switching signal to enter the dynamic collaborative evacuation control mode when the preset abnormal triggering conditions are met or an emergency triggering signal is received.

[0073] The three-layer network modeling unit is used to construct a three-layer evacuation network including an underground evacuation layer, a ground evacuation layer, and a low-altitude auxiliary guidance layer under the dynamic collaborative evacuation control mode. In the underground evacuation layer, the rail transit entrances or key passages in the station hall are mapped as node-type underground evacuation units, the road segments outside the station in the ground evacuation layer are mapped as road segment-type ground evacuation units, and the low-altitude auxiliary guidance nodes in the low-altitude auxiliary guidance layer are associated with the road segment-type ground evacuation units within their coverage area through the coverage mapping relationship, and the network topology is output.

[0074] The passage cost calculation unit is used to receive the network topology and multi-source state data, perform risk normalization processing on the state parameters of each evacuation unit and calculate the basic passage cost value, calculate the low-altitude guidance correction factor and low-altitude guidance anomaly penalty item based on the coverage mapping relationship between the low-altitude auxiliary guidance node and the road segment ground evacuation unit, and then obtain the comprehensive passage cost value at the evacuation unit level, and output it to the risk assessment unit.

[0075] The risk assessment unit is used to calculate the cumulative passage cost of candidate evacuation routes based on the comprehensive passage cost of nodal underground evacuation units and road segment ground evacuation units contained in the candidate evacuation routes, and to obtain the overall passage cost index of the monitoring area by combining the local passage cost peak of key evacuation units, determine the evacuation risk level and control level, and output it to the control strategy generation unit.

[0076] The control strategy generation unit is used to generate a control strategy parameter set based on the evacuation risk level, control level, candidate path priority sequence, path passability status and execution resource status. The control strategy parameter set includes underground guidance parameters, ground traffic signal parameters and low-altitude mission parameters, and is output to the collaborative control interface unit.

[0077] The collaborative control interface unit is used to map the control strategy parameter set into control commands for the underground guidance system, the ground traffic signal system, and the low-altitude equipment platform, and to perform cross-subsystem consistency verification before issuing the commands. If the verification passes, the three types of control commands carry the same control cycle identifier, strategy version number, and acknowledgment identifier and are issued uniformly in the same effective cycle. If the verification fails, the command is handled according to the preset conflict handling rules. At the same time, the issuance result and execution acknowledgment are fed back to the closed-loop control unit.

[0078] The closed-loop control unit is used to receive control command execution acknowledgments and field feedback results, write back the execution results of low-altitude guidance or support tasks as evacuation unit status parameters, and trigger the passage cost calculation unit and risk assessment unit to recalculate passage costs and reconstruct candidate evacuation paths; when the closed-loop reconstruction triggering conditions are met, the three-layer network modeling unit is re-triggered to the collaborative control interface unit to execute the next cycle of collaborative control.

[0079] The beneficial effects of this invention are as follows:

[0080] 1. The rail transit emergency passenger flow collaborative evacuation control method based on a three-layer evacuation network described in this invention constructs a three-layer evacuation network and clarifies that low-altitude nodes are not used as path passage units. Underground entrances and passages are mapped as node-type evacuation units, and external roads are mapped as road segment-type evacuation units. Low-altitude auxiliary guidance nodes only affect the state of ground units through coverage mapping relationships. This solves the problem of chaotic modeling standards caused by directly writing UAV waypoints and hovering points into the path map in the prior art, and maintains the consistency and comparability of path sorting.

[0081] 2. The rail transit emergency passenger flow collaborative evacuation control method based on a three-layer evacuation network described in this invention, by adopting a unified passage cost quantification system of risk normalization and weighted fusion, transforms heterogeneous parameters such as crowd density, queue length, passage speed, number of stranded people, traffic saturation, remaining capacity ratio, and degree of cross-conflict into a comprehensive passage cost value on the same scale. This solves the problems of inconsistent evaluation standards between underground and surface systems and difficulties in cross-scenario comparison, and provides a quantitative basis for multi-system collaborative decision-making.

[0082] 3. The rail transit emergency passenger flow collaborative evacuation control method based on a three-layer evacuation network described in this invention introduces a two-way feedback mechanism of low-altitude guidance correction factor and low-altitude guidance anomaly penalty term. This mechanism transforms parameters such as brightness, coverage area, height, execution success status, directional consistency rate change, density change, and speed change of low-altitude equipment into positive correction and negative penalty for passage costs. This solves the problem in existing schemes where low-altitude guidance is simply regarded as a purely positive benefit, and misleading or failure cannot be reflected in the path cost model in a timely manner. This improves the robustness and authenticity of evacuation decisions.

[0083] 4. The rail transit emergency passenger flow collaborative evacuation control method based on a three-layer evacuation network described in this invention uses a regional overall cost assessment method that integrates the cumulative cost of the path with the local peak value of key evacuation units. When calculating the overall passage cost, it simultaneously considers the cumulative burden of candidate paths and the local peak value of key nodes (such as exits, bottleneck sections, and high-load units), which solves the problem in the prior art that local high-risk bottlenecks are masked by the average value of the path and cannot simultaneously identify overall and local risks, thus improving the accuracy of risk assessment.

[0084] 5. The rail transit emergency passenger flow collaborative evacuation control method based on a three-layer evacuation network described in this invention performs direction consistency, content consistency, spatial conflict detection, and object type matching verification before issuing underground guidance instructions, ground traffic signal instructions, and low-altitude mission instructions through cross-subsystem consistency verification and transactional unified issuance mechanism. It also binds the three types of instructions to the same transaction, carries the same control cycle identifier and strategy version number, solves the cross-system instruction inconsistency problems such as conflicts between station guidance and ground signal directions, and overlap between low-altitude guidance and impassable areas, and reduces the conflict risk in multi-system collaborative control.

[0085] 6. The rail transit emergency passenger flow collaborative evacuation control method based on a three-layer evacuation network described in this invention, through the execution result state write-back and closed-loop reconstruction mechanism, writes back feedback parameters such as illumination improvement value, communication packet loss rate improvement, direction consistency rate change, density change, and speed change in the low-altitude support task as the state parameters of the evacuation unit, and uses them to drive the recalculation of passage cost and path reconstruction in the next cycle. This solves the problem in the prior art that the low-altitude execution result is only recorded as "success / failure" and cannot participate in the subsequent path evaluation, and realizes the deep coupling and continuous optimization of the low-altitude guidance effect and the evacuation network state.

[0086] 7. The rail transit emergency passenger flow collaborative evacuation control method based on a three-layer evacuation network described in this invention, through low-altitude equipment safety constraints and edge-side anonymization processing mechanisms, reads restricted flight zone, no-fly zone, and electronic fence data before task generation, prohibits equipment that does not meet safety conditions from participating in guidance, and converts video streams into state parameters such as density, direction, and speed at the equipment end or edge node before uploading, thus solving the flight safety and data privacy issues when low-altitude equipment participates in evacuation control and improving the engineering usability of the system.

[0087] 8. The rail transit emergency passenger flow collaborative evacuation control method based on a three-layer evacuation network described in this invention, through a multi-device coverage conflict handling mechanism, solves the problem of passenger flow confusion and false correction that may occur when multiple low-altitude auxiliary guidance nodes act simultaneously on the same section-type ground evacuation unit. This is achieved by selecting the optimal mode based on the consistency of guidance direction, the overlap of coverage area, and the direction of ground signal release, as well as by using amplitude limiting superposition or conflict suppression mode. This ensures the effectiveness of the guidance effect.

[0088] 9. The rail transit emergency passenger flow collaborative evacuation control method based on a three-layer evacuation network described in this invention, through transactional rollback and resource weight degradation mechanism, when any subsystem fails, times out, or has inconsistent versions, the current cycle strategy is rolled back as a whole while maintaining the safety parameters of the previous cycle, and at the same time the available weight of the failed device is reduced. This solves the problem of inconsistent state and system unreliability caused by some instructions being effective and some instructions being invalid, and enhances the stability and recoverability of cross-system collaborative control. Attached Figure Description

[0089] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. The drawings described below are some embodiments of the present invention. For those skilled in the art, it is obvious that other drawings can be obtained from these drawings without creative effort.

[0090] Figure 1 This is a schematic diagram of the collaborative evacuation control system for sudden passenger flow in rail transit based on a three-layer evacuation network, as described in this invention.

[0091] Figure 2 This is a flowchart of the collaborative evacuation control method for sudden passenger flow in rail transit based on a three-layer evacuation network, as described in this invention.

[0092] Figure 3 This is a schematic diagram of the object mapping relationship in a three-layer evacuation network.

[0093] Figure 4 The diagram shows the simulation results of coordinated evacuation based on the method of this patent in the embodiments of the present invention. It includes the cumulative evacuation ratio curve under different control schemes and the curve of the change of the overall traffic cost index G value of the monitoring area. The first part of the curve reflects the congestion spread after the formation of sudden passenger flow, the middle part reflects the path reconstruction process after the coordinated effect of underground guidance, ground signal and low-altitude guidance, and the last part reflects the trend of the overall cost gradually stabilizing after the execution result status is written back.

[0094] In the picture:

[0095] 100 - Multi-source data access unit; 200 - Event judgment and mode switching unit; 300 - Three-layer network modeling unit; 301 - Underground evacuation layer; 302 - Ground evacuation layer; 303 - Low-altitude auxiliary guidance layer; 400 - Traffic cost calculation unit; 500 - Risk assessment unit; 600 - Control strategy generation unit; 700 - Cooperative control interface unit; 800 - Closed-loop control unit; 900 - Execution object module; 910 - Underground guidance system; 920 - Ground traffic signal system; 930 - Low-altitude equipment platform. Detailed Implementation

[0096] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0097] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "axial," "radial," "vertical," "horizontal," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0098] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0099] like Figure 1 As shown, the rail transit emergency passenger flow collaborative evacuation control system based on a three-layer evacuation network described in this invention includes: a multi-source data access unit 100, an event determination and mode switching unit 200, a three-layer network modeling unit 300, a passage cost calculation unit 400, a risk assessment unit 500, a control strategy generation unit 600, a collaborative control interface unit 700, a closed-loop control unit 800, and an execution object module 900.

[0100] The multi-source data access unit 100 is used to access underground, ground, and low-altitude multi-source data, and outputs it to the event determination and mode switching unit 200 and the three-layer network modeling unit 300. The event determination and mode switching unit 200 is used to determine whether to enter the dynamic coordinated evacuation mode based on the traffic risk status index, and outputs the mode switching signal to the three-layer network modeling unit 300. The three-layer network modeling unit 300 is used to construct the topology relationship of the underground evacuation layer, the ground evacuation layer, and the low-altitude auxiliary guidance layer, and outputs the network model to the traffic cost calculation unit 400. The traffic cost calculation unit 400 is used to calculate the basic traffic cost value, low-altitude guidance correction factor, and abnormal penalty item of each evacuation unit, and outputs the comprehensive traffic cost value to the risk assessment unit 500. The risk assessment unit 500 is used to calculate the cumulative traffic cost value of candidate paths and the overall regional traffic cost index, and outputs the risk level and... The control level is sent to the control strategy generation unit 600; the control strategy generation unit 600 is used to generate underground guidance parameters, ground traffic signal parameters, and low-altitude mission parameters, and outputs the control strategy parameter set to the collaborative control interface unit 700; the collaborative control interface unit 700 is used to map the control parameters into control commands and perform cross-subsystem consistency checks, and output commands to the execution object 900; the closed-loop control unit 800 is used to receive execution receipts and on-site feedback results, write back the status parameters to the passage cost calculation unit 400 and the risk assessment unit 500, and trigger the next cycle reconstruction; the execution object module 900 includes an underground guidance system 910, a ground traffic signal system 920, and a low-altitude equipment platform 930, which respectively receive and execute the guidance commands, signal commands, and mission commands issued by the collaborative control interface unit 700, and return the execution receipts to the closed-loop control unit 800.

[0101] like Figure 2 As shown, the collaborative evacuation control method for sudden passenger flow in rail transit based on a three-layer evacuation network described in this invention specifically includes the following steps:

[0102] S01: Multi-source data acquisition and judgment of sudden passenger flow events, specifically:

[0103] The system collects real-time status data from rail transit station entrances and exits, station hall passages, external road sections, ground traffic signals, and low-altitude auxiliary equipment via a multi-source data access unit. Specifically, underground data comes from gate passage records of the automatic fare collection system and the station's video passenger flow recognition system, providing data on crowd density, queue length, average passage speed, and number of stranded passengers at each exit. Ground data comes from traffic signal controllers and road traffic flow detection equipment, providing real-time traffic saturation, remaining capacity ratio, and intersection conflict levels on external road sections. Low-altitude data comes from ground crowd distribution heatmaps, movement direction vectors, and the locations of local congestion points collected by sensing drones or tethered drones. The low-altitude equipment also transmits mission execution feedback, coverage deviation, and waypoint confirmation status.

[0104] Since the time base and spatial coordinate system of different data sources may differ, spatiotemporal alignment is performed on all data: BeiDou or GPS timestamps are added uniformly in time, and missing data is replaced with the most recent valid sampled value or the estimated value of adjacent units, and the state reliability of the corresponding evacuation unit is reduced; in space, affine transformation and geographic correction are used to map the station plan, road GIS coordinates and low-altitude image coordinates to a unified geographic coordinate system (such as CGCS2000).

[0105] After data fusion is completed, the event determination and mode switching unit 200 determines whether a sudden passenger flow event has occurred based on the traffic risk status indicators, specifically:

[0106] Calculate the unit risk index e of a single evacuation unit q q (k):

[0107]

[0108] In the formula, k is the control cycle number. Let a be the value of the i-th state parameter after risk normalization; i The corresponding weights are non-negative and normalized, and s is the total number of state parameters.

[0109] For risk-related indicators such as crowd density, queue length, number of stranded people, and traffic saturation, positive normalization is used:

[0110] ,

[0111] For efficiency-related indicators such as traffic speed and remaining capacity ratio, inverse normalization is used:

[0112] ,

[0113] In the formula: This means limiting the value to the range of 0 to 1; This represents the historical maximum value of the i-th state parameter; This represents the historical minimum value of the i-th state parameter;

[0114] The traffic risk status index E(k) of the monitoring area is obtained by aggregating the risk indicators of evacuation units within the area:

[0115] ,

[0116] In the formula, This is the weight of the maximum value item, and its value generally ranges from 0 to 1. For the set of event-triggered monitoring units; ω q ω is the aggregation weight of the evacuation unit q. q Non-negative and in Ω E The summation of the inner components is 1.

[0117] When E(k) exceeds the preset trigger threshold, or when any of the following indicators at key entrances / exits—crowd density, average passage speed decrease rate, queue length, or number of stranded people—exceeds the corresponding threshold, a sudden passenger flow event is determined to have occurred, and the dynamic collaborative evacuation control mode is entered.

[0118] For example: At a train station during evening rush hour, exit A corresponds to evacuation unit N1, and a crowd density of 4.5 people / m² was detected. 2 (Normalized value 0.667), queue length 28m (normalized value 0.700), average speed 0.55m / s (normalized value 0.708), number of stranded people 120 (normalized value 0.800). Taking weights a1=0.60, a2=0.25, a3=0.05, a4=0.10, the calculated value is e. N1 (k)=0.690, exceeding the trigger threshold of 0.65, the system immediately enters the coordinated evacuation control mode.

[0119] S02: Construct a three-layer evacuation network, as follows:

[0120] Under the dynamic collaborative evacuation control mode, the three-layer network modeling unit 300 is constructed as follows: Figure 3 The three-layer evacuation network shown includes an underground evacuation layer 301, a ground evacuation layer 302, and a low-altitude auxiliary guidance layer 303.

[0121] The underground evacuation layer 301 maps the rail transit entrances and exits and key passageways in the station hall into nodal underground evacuation units 310. Each nodal unit has state parameters such as crowd density, queue length, average passage speed, and number of people stranded.

[0122] The ground evacuation layer 302 maps the road segments outside the station into road segment-type ground evacuation units 320. Each road segment-type unit has state parameters such as traffic flow saturation, remaining capacity ratio, degree of intersection conflict, road segment length, and estimated travel time.

[0123] The low-altitude auxiliary guidance layer 303 maps the effective coverage area of ​​each guiding or supporting UAV to a low-altitude auxiliary guidance node 330. It is important to note that the low-altitude auxiliary guidance node is not considered a path passage unit and does not participate in the calculation of cumulative path passage costs or path ranking. It only affects the state correction of road segment-type ground evacuation units within its coverage area through the coverage mapping relationship 340. Intersections, pedestrian crossings, exit directions, and target areas are only considered as topological connection objects 350 and do not participate in passage cost calculation as independent evacuation units.

[0124] The method for determining the mapping relationship is as follows:

[0125] When the Euclidean distance between the ground projection center of a low-altitude auxiliary guidance node and a road segment-type ground evacuation unit is less than the effective guidance radius of the low-altitude auxiliary guidance node, the low-altitude auxiliary guidance node is considered to cover the road segment-type ground evacuation unit. One low-altitude node can cover multiple ground evacuation units, and a ground evacuation unit may also be covered by multiple low-altitude nodes simultaneously.

[0126] S03: Calculate the comprehensive passage cost at the evacuation unit level, as follows:

[0127] The passage cost calculation unit 400 performs risk-based normalization on the multi-source state parameters of each evacuation unit, and then calculates the basic passage cost value, as follows:

[0128] S3.1 The basic passage cost B for evacuation unit q q The weighted sum of the normalized state parameters:

[0129]

[0130] In the formula: w i The weight is the weight of the i-th state parameter. The initial value of the weight is set according to the station type, passenger flow direction and road network topology, and can be dynamically adjusted according to the control cycle. The correction amount is determined based on the evacuation efficiency error, state reliability and current control level correction factor of the most recent M cycles, and then normalized again.

[0131] S3.2: For road segment-type ground evacuation units covered by low-altitude auxiliary guidance nodes, calculate the low-altitude guidance correction factor. and low-altitude guidance abnormality penalty item The specific calculations are as follows:

[0132] Low-altitude equipment performance indicators It is obtained by weighting three factors: brightness, coverage area, and height effect.

[0133]

[0134] in:

[0135] This represents the comprehensive performance index of the low-altitude equipment d, with a value ranging from 0 to 1. The larger the value, the stronger the guidance or protection capability of the equipment.

[0136] , , The weighting coefficients of the three indicators satisfy the following conditions: ;

[0137] The normalized value of the projection brightness or directional sign contrast of the low-altitude device d is obtained by forward normalizing the measured brightness; the higher the brightness, the closer the normalized value is to 1.

[0138] This is the normalized value of the area of ​​the effective guidance zone formed by the low-altitude equipment d on the ground; the larger the area and the wider the coverage, the closer the normalized value is to 1.

[0139] This is the normalized value of the hovering height effect of the low-altitude device d. The lower the altitude, the better. The higher the value, the stronger the visual guidance effect. The formula for calculating the value is:

[0140] ,

[0141] Where: H d H is the current hovering altitude of the low-altitude equipment. max and H min These are the upper and lower limits of the effective height range, respectively.

[0142] The spatial influence coefficient α of low-altitude equipment d on evacuation unit q q,d The calculation formula depends on the ratio of the ground projection distance to the effective guiding radius, as follows:

[0143] ,

[0144] In the formula: dist q,d R is the distance from the geometric center of the evacuation unit to the ground projection center of the low-altitude equipment. d Let d be the effective guiding radius of the low-altitude equipment.

[0145] Effective guided response coefficient ρ q,dIt is used to reflect the effect of low-altitude guidance in actual evacuation, and is determined by a combination of execution success status, changes in directional consistency rate, density changes, and velocity changes:

[0146]

[0147] In the formula:

[0148] ρ q,d The effective guidance response coefficient of the low-altitude equipment d to the evacuation unit q is represented by the value range of 0 to 1, reflecting the positive effect of the guidance measures in actual evacuation.

[0149] ACK d This indicates the successful execution status of the low-altitude equipment d. The value is 1 when the execution is successful, and 0 when the execution fails, times out, or the receipt is missing (or a value between 0 and 1 according to the confidence level).

[0150] ΔDir q This refers to the periodic variation in the consistency rate between the actual movement direction of the population and the low-altitude guidance direction within the coverage area of ​​evacuation unit q. Specifically, it is defined as: the consistency rate of the current period minus the consistency rate of the previous period. The consistency rate is calculated as the proportion of the number of people conforming to the guidance direction to the total number of monitored people in the area. When ΔDir... q >0 indicates that more people are starting to follow the guidance (positive improvement); when ΔDir q <0 indicates an increase in the number of people deviating from the guiding direction (negative deterioration);

[0151] ΔDen q Let ΔDen be the rate of change in population density within the coverage area of ​​evacuation unit q, defined as: (current density − previous cycle density) / previous cycle density. q >0 indicates increased density (intensified congestion), when ΔDen q A value less than 0 indicates a decrease in density (congestion relief). In the formula... This term is positive when density decreases, indicating a positive effect; conversely, it is 0 when density increases, meaning it does not contribute a positive response.

[0152] Let ΔV be the rate of change of the average passage speed of the population within the coverage area of ​​evacuation unit q, defined as (current speed - previous cycle speed) / previous cycle speed. q >0 indicates an increase in speed (improved traffic efficiency), when ΔV q <0 indicates a decrease in speed. Only contributes a positive response when speed increases;

[0153] For each item's weight, satisfying ;

[0154] This means limiting the value to the range of 0 to 1;

[0155] Low-altitude guidance correction factor η q The calculation method is as follows:

[0156] ,

[0157] In the formula:

[0158] D q A collection of low-altitude equipment covering evacuation unit q; s d η represents the availability status coefficient of low-altitude equipment, where 0 indicates unavailable and 1 indicates available; max This is the upper limit of the correction factor.

[0159] Furthermore, when multiple low-altitude auxiliary guidance nodes cover the same road segment ground evacuation unit, the low-altitude guidance correction factor η q The summation term in the formula needs to be corrected based on the directional consistency and spatial conflict relationship between the devices: if the guiding directions are consistent and there is no conflict, the optimal mode (only the largest contribution term is taken) or the amplitude limiting superposition mode (truncation after summation) can be used; if there is inconsistency in direction or spatial conflict, the conflict suppression mode (setting the contribution of conflicting devices to zero) is used.

[0160] Low-altitude guidance anomaly penalty item This occurs when low-altitude equipment fails to execute, directional consistency decreases, density increases, speed decreases, or coverage deviates.

[0161] ,

[0162] In the formula:

[0163] This represents the abnormal penalty term generated by the low-altitude auxiliary guidance node for evacuation unit q, with a value range of 0~ , Set a maximum limit for preset penalty items; The larger the value, the more severe the guidance anomaly, which will lead to an increase in the comprehensive passage cost of evacuation unit q (equivalent to a deterioration in passage conditions).

[0164] F d This indicates that the low-altitude device d's task execution failed, timed out, or lacked feedback. The value is 1 when the low-altitude device d experiences any of the following: task execution failed, no response was received after the preset response time, or the response content is missing or incorrectly formatted; otherwise, the value is 0.

[0165] O qThis indicates a coverage deviation or waypoint deviation. When the deviation between the actual hovering position of the low-altitude equipment and the target waypoint exceeds a preset threshold, or when the equipment's coverage area deviates significantly from the geometric range of evacuation unit q, O... q =1; otherwise O q =0.

[0166] b1, b2, b3, b4, and b5 are coefficient weights.

[0167] S3.3: The comprehensive passage cost C of evacuation unit q q for:

[0168] .

[0169] For evacuation units not effectively covered by low-altitude auxiliary guidance nodes, η q =0; when the low-altitude guidance anomaly is not triggered, ψ q =0.

[0170] S04: Calculate the cumulative cost of the route and the overall regional travel cost index to determine the risk level and control level. This includes the following steps:

[0171] S4.1: Risk assessment unit 500 generates a set of candidate evacuation routes based on the initial underground evacuation unit, target area, topological connectivity, route accessibility, maximum detour constraint, and ground traffic signal release constraint. Candidate routes can be generated using the K-shortest path algorithm, A* algorithm, Dijkstra's algorithm, or a graph search algorithm based on access cost.

[0172] S4.2: Calculate the cumulative travel cost of each candidate path, specifically:

[0173] The cumulative passage cost A for candidate path p p Defined as the sum of the scale-corrected comprehensive access costs of the evacuation units contained in the path, the formula is:

[0174] ,

[0175] In the formula: P p L is the set of all evacuation units (nodal-type underground evacuation units and road segment-type surface evacuation units) contained in path p; q K represents the length of the evacuation unit or the estimated travel time. q This represents the remaining capacity or capacity factor. In the same path calculation, L... q and K q A matching combination of parameters should be used and the units should be unified: when L q When determining the length of a road segment, K qTake the corresponding normalized traffic capacity or design capacity, and convert the ratio into a dimensionless traffic impedance; when L q When determining the estimated travel time, K q It can be 1 or a normalized capacity coefficient.

[0176] S4.3: Further determine the priority order of the primary / secondary paths, specifically as follows:

[0177] The priority ranking of candidate paths is directly based on unnormalized A p Value execution, A p A smaller value indicates a better path. The main evacuation path is selected from feasible candidate paths A. p The shortest path is determined, and the secondary evacuation path is determined by the next shortest path. To prevent frequent path switching, when the A value of the primary path and the secondary path is... p When the difference is less than the preset switching hysteresis threshold, the system maintains the current main path; however, when the comprehensive passage cost of any key evacuation unit in the main path exceeds the local forced switching threshold, the main path is downgraded, removed, or forced switched.

[0178] S4.4: Integrate the overall regional traffic cost indicators and determine the risk level and control level, specifically:

[0179] The overall traffic cost index G of the monitoring area is obtained by fusing the cumulative path cost and the peak local traffic cost of key evacuation units:

[0180] ,

[0181] In the formula:

[0182] β1 and β2 are the fusion weights, and β1 + β2 = 1;

[0183] A ref To reference the cumulative cost value, empirical values ​​or quantile values ​​of historical safety samples are usually taken;

[0184] K represents the set of key evacuation units, including nodal underground evacuation units associated with the exit direction, road segment ground evacuation units corresponding to cross-street connection bottlenecks, high-load units in the current main evacuation route, and evacuation units with a comprehensive passage cost ranking in the top m or exceeding the 85th percentile of historical safety samples.

[0185] r is the total number of feasible paths in the candidate path set;

[0186] π p Let π be the fusion weight of the p-th candidate path in the overall cost of the region. By default, all path weights are equal, i.e., π. p =1 / r; If the system can predict the passenger flow allocated to each path, then weight the passenger flow according to the predicted passenger flow ratio: ;

[0187] D p To predict the passenger flow (in people) that will be assigned to the p-th candidate path. p It can be obtained through historical passenger flow distribution ratios, real-time route selection probability models, or simulation predictions.

[0188] G represents the overall traffic cost index for the monitored area, with a value ranging from 0 to 1.

[0189] Evacuation risk levels are divided into multiple levels based on configurable thresholds. A preferred classification method is shown in Table 1:

[0190] Table 1

[0191] Risk level Overall traffic cost index G in the monitoring area control meaning Level 1 0≤G＜0.45 Regular adjustment Level 2 0.45≤G＜0.70 Enhanced traffic diversion Level 3 0.70≤G≤1.00 High-risk intervention

[0192] The control level specifies the intensity of control actions. It can be consistent with the evacuation risk level, or it can be increased due to excessively high local peak values ​​in critical evacuation units, insufficient availability of execution resources, or low reliability of the status. When the comprehensive passage cost of any critical evacuation unit exceeds the local alarm threshold, the control level will be raised by one level or enter an enhanced control state, even if the overall area G has not reached a high-risk level.

[0193] Step S05: Generate the control strategy parameter set, as follows:

[0194] S5.1: Generate basic control strategy parameters

[0195] The control strategy generation unit 600 generates a control strategy parameter set based on the evacuation risk level, control level, candidate path priority sequence, path accessibility status, and execution resource status. This parameter set includes: primary evacuation path identifier, secondary evacuation path identifier, guidance display template number, broadcast voice template number, ground traffic signal phase template number, low-altitude equipment waypoint template number, effective period, duration, control period identifier, strategy version number, and receipt identifier.

[0196] S5.2: Generate low-altitude emergency response parameters

[0197] When the control level reaches a high level and at least one of the following conditions is met, the system additionally generates low-altitude emergency response parameters: illuminance below a threshold, communication packet loss rate above a threshold, designated evacuation unit status unobservable, physical blockage, or local remaining passage capacity below a capacity threshold. These emergency response parameters are used to dispatch low-altitude equipment to perform temporary lighting support, communication relay deployment, aerial visual guidance, or emergency evacuation auxiliary equipment delivery.

[0198] S06: Map control commands and perform cross-subsystem consistency checks and unified distribution, as detailed below:

[0199] S6.1: Mapping to generate control commands

[0200] The collaborative control interface unit 700 maps the control strategy parameter set into control commands for underground guidance systems, ground traffic signal systems, and low-altitude equipment platforms.

[0201] S6.2: Perform cross-subsystem consistency checks

[0202] Before unified distribution, cross-subsystem consistency verification must be performed, including:

[0203] Consistency verification between the exit direction of the main evacuation route and the direction of ground traffic signal release;

[0204] Consistency verification between the displayed directional information and the broadcast announcements within the station;

[0205] Spatial conflict detection between the coverage area of ​​low-altitude auxiliary guidance nodes and impassable areas;

[0206] Object type matching validation for computed object fields, control object fields, and display object fields.

[0207] When the consistency check fails, the current control command group is prohibited from being issued, and the conflict is handled according to the preset conflict handling rules, such as rematching the guide template and the signal phase template, or suppressing the guidance contribution of conflicting low-altitude equipment.

[0208] S6.3: Transaction-based unified distribution

[0209] To ensure the synchronized effectiveness of the three types of instructions, the system employs a transactional unified issuance mechanism: An instruction group transaction number is generated, binding underground guidance instructions, ground traffic signal instructions, and low-altitude mission instructions to the same transaction. All three types of instructions are required to carry the same control cycle identifier, policy version number, and receipt identifier. The transaction enters the commit state only when all three types of instructions pass consistency checks, version consistency checks, and effectiveness cycle constraints, ensuring that all three types of instructions are issued uniformly within the same effectiveness cycle. If any subsystem returns a failure, timeout, version inconsistency, or missing receipt status, the transaction enters the rollback state, stopping the current cycle's policy from taking effect while maintaining the safety control parameters of the previous cycle.

[0210] S07: Receive execution receipt, write back status, and reconstruct closed loop, as detailed below.

[0211] The closed-loop control unit 800 receives control command execution acknowledgments and field feedback results, and writes back the execution results of low-altitude guidance or support tasks as the state parameters of the corresponding evacuation units according to preset mapping rules (for example, writing back the illumination improvement value as the reduction coefficient of the illumination deficiency penalty, and directly writing back the changes in direction consistency rate, density, and speed as state update quantities). Based on the written-back state parameters, it recalculates the passage cost and reconstructs candidate evacuation paths. When the closed-loop reconstruction trigger conditions are met (including but not limited to the arrival of the control cycle timer, the change in the overall passage cost index of the area exceeding the threshold, the sudden change in the state of key evacuation units, and the subsystem execution failure acknowledgment, etc.), the system re-executes steps S02 to S06, forming a continuous closed-loop control process until the evacuation is completed.

[0212] Example 1: Coordinated evacuation control in the event of a sudden surge in passenger flow

[0213] The following is a specific numerical example. A sudden surge in passenger flow occurred at a rail transit station during the evening rush hour. Exit A corresponds to a node-type underground evacuation unit N1. Road R1 outside the station is the key road segment after exiting the station, R2 is the road segment leading to area C, and R3 is the road segment leading to area D. Drone D1 covers R2, and drone D2 covers R3.

[0214] Key parameters are set as follows: normalized population density range of 1.5–6.0 people / m² 2 The normalized ranges for queue length (0–40 m), average traffic speed (0.2–1.4 m / s), number of stranded people (0–150), traffic saturation (0.30–1.00), remaining capacity ratio (0–1), and intersection conflict level (0–1) are as follows: The road segment weights are calculated as 0.396 for traffic saturation, 0.359 for remaining capacity risk, and 0.245 for intersection conflict level (Example 1 emphasizes balanced weights). The reference upper limit for low-altitude equipment brightness is 1000 lux, and the reference value for coverage area is 600 m². 2 Effective height range: 20–60 m. Upper limit of correction factor η. max =0.35, upper limit of penalty term ψ max =0.20. The overall index fusion weights for the monitoring area are β1=0.7 and β2=0.3.

[0215] The original state and normalized values ​​of N1 are shown in Table 2:

[0216] Table 2

[0217] parameter Original value Normalized value Crowd density 4.5 people / m² 0.667 Queue length 28m 0.700 Average traffic speed 0.55m / s 0.708 Number of stranded people 120 people 0.800

[0218] Taking the basic passage cost weights of 0.35, 0.20, 0.25, and 0.20, we obtain B according to formula (5). N1 =0.710.

[0219] For road segment-type ground evacuation units, in Example 1, the weights for traffic flow saturation risk, remaining capacity risk, and intersection conflict degree are 0.396, 0.359, and 0.245, respectively; the states of R1, R2, and R3 and their basic traffic costs are shown in Table 3.

[0220] Table 3

[0221] evacuation unit Traffic saturation Remaining capacity ratio Cross-conflict level Basic travel value R1 0.92 0.25 0.80 0.817 R2 0.78 0.40 0.65 0.647 R3 0.60 0.58 0.40 0.419

[0222] Drone D1 parameters: brightness 850 lux, coverage area 480m 2 Hovering height: 30m; D2 parameters: brightness 760 lux, coverage area 420m² 2 Hovering height 35m.

[0223] Take the low-altitude equipment performance weight μ B =0.40, μ A =0.40, μ H =0.20, γ is calculated. D1 =0.810, γ D2 =0.709. The distance from D1 to the center of the ground projection of R2 is 20m, and the effective radius is 50m. According to the formula, α is obtained. R2,D1 =0.600; The distance from D2 to R3 is 15m, and the effective radius is 45m, so α R3,D2 =0.667.

[0224] Under initial feedback, the effective guided response coefficient ρ is taken. R2,D1 =0.72, ρ R3,D2 =0.76, s D1 =s D2 =1. Calculate the low-altitude guidance correction factor using the formula:

[0225] ,

[0226] There was no deviation in direction, increase in density, or decrease in velocity, therefore: ψ R2 =ψ R3 =0. Calculate the comprehensive passage cost using the formula:

[0227] ,

[0228] ,

[0229] N1 and R1 were not covered by low-altitude airspace, therefore ;

[0230] Let candidate paths P1 be N1→R1→R2 and P2 be N1→R1→R3. Take the scale factor L for each evacuation unit. q / Kq =1, then the cumulative cost of the path is:

[0231]

[0232] Take reference value A ref =3.0, π P1 =π P2 =0.5, the local peak value of the critical evacuation unit is 0.817 of R1, calculate G:

[0233]

[0234] Therefore, the current risk level is Level 3. The system selects P2 as the primary evacuation route and P1 as the secondary evacuation route.

[0235] The generated control strategy is as follows: The underground guidance system displays "Prioritize evacuation via area D" on the guidance screen at exit A, and the broadcast frequency is increased from once every 30 seconds to once every 10 seconds; the ground traffic signal system extends the green light for the P2 direction at the intersection corresponding to R1 from 25 seconds to 40 seconds, and compresses the conflict direction from 35 seconds to 20 seconds; the low-altitude equipment platform schedules UAV D2 to hover above area D, increasing the guidance intensity to 0.85, and UAV D1 to cover the entrance of area C to perform auxiliary diversion. After the consistency verification is passed, the system issues instructions uniformly in the same effective cycle.

[0236] Feedback after one control period showed that the N1 population density dropped to 3.6 people / m². 2 Queue length decreased to 18m, average traffic speed increased to 0.78m / s, and number of stranded people decreased to 78; R1 traffic saturation decreased to 0.72, remaining capacity ratio increased to 0.45, and intersection conflict level decreased to 0.55; R3 directional consistency rate increased from 0.61 to 0.83, and average traffic speed increased from 0.72m / s to 0.91m / s.

[0237] The system writes the above feedback back to the corresponding evacuation unit state parameters and recalculates A. P1 =1.503, A P2 =1.330, G=0.501, the risk level has decreased from level three to level two. The system maintains P2 as the primary path, but lowers the control level to level two, reducing the broadcast frequency and low-altitude guidance intensity.

[0238] Example 2:

[0239] This Example 2 illustrates how low-altitude support missions participate in the closed loop through state write-back, rather than simply recording "support completed." Station exit E corresponds to a node-type underground evacuation unit N2, and road R4 outside the exit is a section-type ground evacuation unit. Due to the collapse of temporary fencing, the effective passage width of R4 decreased from 6.0m to 2.5m, the on-site illuminance decreased from 150 lux to 35 lux, and the communication packet loss rate increased to 18%. The initial state of R4 is shown in Table 4:

[0240] Table 4

[0241] parameter numerical values Traffic saturation 0.84 Remaining capacity ratio 0.18 Cross-conflict level 0.76 Illuminance 35 lux Packet loss rate 18% Path passability status Restricted access

[0242] Example 2 pertains to a scenario of partial exit blockage. The weights of road segment-type ground evacuation units are prioritized based on remaining capacity. The weights for traffic flow saturation risk, remaining capacity risk, and intersection conflict severity are set to 0.143, 0.757, and 0.100, respectively. From the traffic flow saturation of 0.84 corresponding to a normalized risk of 0.771, a remaining capacity ratio of 0.18 corresponding to a risk of 0.820, and an intersection conflict severity of 0.760, we obtain B. R4 =0.807. The overall regional traffic cost index reached 0.812, which the system judged as a level three risk.

[0243] Candidate route P3 includes R4, with P4 detouring to the adjacent road R5. Because R4 is marked as restricted access and its local capacity is below the threshold, the system temporarily does not use P3 as the primary route, but instead selects P4 as the primary evacuation route. Consistency verification reveals that the original ground signal control template still allows passage in the R4 direction, and the original guidance screen content still points to exit E. The system determines that the direction consistency verification has failed and prohibits the issuance of this set of instructions. A new control strategy is generated: the underground guidance system switches the guidance from exit E to the adjacent exit F, broadcasting instructions for passengers to detour to exit F; the ground signal system compresses the signal in the R4 direction and enhances the passage in the R5 direction; the low-altitude equipment platform dispatches and supports UAV D3 to perform lighting and communication relay tasks above R4. The support mission message includes equipment identification, target evacuation unit, mission type (temporary lighting + communication relay), hovering altitude 32m, target illuminance ≥180lux, target packet loss rate ≤5%, duration 120s, and acknowledgment identifier ACK-UAV-D3.

[0244] Feedback after D3 arrived: Illuminance increased by 155 lux (current illuminance 190 lux), and communication packet loss rate decreased to 4%. ACK obtained. D3 =1, coefficients a1=a2=a3=a4=0.25, according to the formula, ρ R4,D3 =0.405. The low-altitude equipment efficiency index is 0.780, and the spatial impact coefficient is 0.750. According to the formula, η is... R4=0.35×0.750×0.780×0.405=0.083. Since the feedback is a positive improvement, ψ R4 =0.

[0245] The system does not directly classify R4 as fully recovered, but instead writes back the state according to the mapping rules: illuminance increases from 35 lux to 190 lux, and the insufficient illuminance penalty decreases from 0.26 to 0.05; the communication packet loss rate decreases from 18% to 4%, and the communication availability status changes from "unavailable" to "available"; local density decreases by 18%, and the local remaining capacity ratio increases from 0.18 to 0.38; the direction consistency rate increases by 0.21, and the speed increases by 23%. After this write-back, the basic passage cost of R4 decreases from 0.807 to 0.560, and the comprehensive passage cost C... R4 =0.560×(1−0.083)=0.514. The next cycle calculation found that path P3, which includes R4, was restored as a feasible candidate path, but its cumulative cost value was still higher than that of the detour path P4. Therefore, the system continued to keep P4 as the primary path and used P3 as a backup path. Only when the comprehensive toll cost value of R4 is lower than the preset threshold for two consecutive cycles will the system allow P3 to be reinstated into the primary path competition set.

[0246] Example 3:

[0247] Example 3 illustrates the handling methods when multiple low-altitude devices cover the same road segment and when low-altitude guidance malfunctions. Assume that a road segment-type ground evacuation unit R6 is simultaneously covered by low-altitude auxiliary guidance nodes U4 and U5. U4's guidance direction points to target area M, while U5's guidance direction points to target area N due to a waypoint template error. The system verifies the tasks of U4 and U5 based on the same control cycle identifier and strategy version number. It finds that U5's guidance direction is inconsistent with the direction displayed by the underground guidance system and the direction released by the ground signal, thus determining a direction conflict. The system performs conflict suppression on U5's contribution item, setting its available state coefficient to 0, i.e., s. U5 =0. U4 continues to participate in the calculation of the low-altitude guidance correction factor.

[0248] If both U4 and U5 pass the direction consistency check, the system can use the amplitude limiting and superposition method:

[0249] ,

[0250] If, after the low-altitude mission is executed, feedback shows that the crowd direction consistency rate of R6 decreases by 0.12, the local density increases by 0.10, and the passage speed decreases by 0.08, then the system generates a low-altitude guidance anomaly penalty term ψ according to the formula. R6The value is greater than 0, and this value is added to the overall passage cost of R6. Even if low-altitude coverage exists, the system will not simply reduce the passage cost of R6, but will increase the risk based on anomaly feedback. If R6 is simultaneously marked as restricted passable, it will only participate in the candidate path set as a backup path; if the state confidence level is lower than a preset threshold for two consecutive cycles, it will be marked as unobservable, triggering manual confirmation or conservative weight reduction.

[0251] Example 4:

[0252] Example 4 illustrates the specific processing procedure when a cross-subsystem instruction transaction enters a rollback state. In the k-th control cycle, the system generates the same policy version number V. k and the same effective period T k The instructions included underground guidance, ground traffic signal instructions, and low-altitude mission instructions. The underground guidance instructions required the station's guidance screens to switch to Exit F; the ground traffic signal instructions required the green light for direction R5 to be extended from 25 seconds to 40 seconds; and the low-altitude mission instructions required ensuring that UAV D4 performed lighting and visual guidance above R5.

[0253] Before unified distribution, all three types of instructions pass directional consistency checks and space conflict detection, and the system generates instruction group transaction number TX. k The system then entered a pending submission state. After the command was issued, both the underground guidance system and the ground traffic signal system returned successful execution receipts, but the low-altitude equipment platform returned inconsistent version receipts, and D4 failed to complete waypoint confirmation within the preset receipt time window. Based on this, the system determined that the transaction did not meet the conditions for the simultaneous effectiveness of the three types of instructions, and the transaction entered a rollback state. At this time, the system stopped the current cycle strategy V. k The change takes effect, but the exit directions F and R5 are not written into the path status as completed coordinated handover results. Instead, the safety control parameters from the previous cycle are maintained (the previous cycle's guidance display template and signal phase template are preserved), and the low-altitude auxiliary guidance node corresponding to D4 is marked as restricted availability. At the same time, the failure reason, failed subsystem, transaction number, and policy version number are written into the receipt record to avoid misusing the results of the ineffective low-altitude mission in the next cycle.

[0254] During the generation of the control strategy in the (k+1)th control cycle, the system reduces the resource availability weight of D4 from 1.00 to 0.30; if two consecutive timeouts occur, D4 is removed from the available resource set. The calculation of the low-altitude guidance correction factor no longer incorporates D4's contribution. The control strategy generation unit prioritizes selecting the backup low-altitude equipment D5 or reduces the low-altitude guidance intensity, and re-matches the underground guidance template and the ground signal phase template. Only after D4 returns consistent versions and a successful mission execution confirmation for two consecutive cycles does the system gradually restore its resource availability weight.

[0255] Example 5: Simulation Verification

[0256] In an exemplary simulation scenario, Example 5 constructs a collaborative evacuation scenario for sudden passenger flow in rail transit, comprising an underground evacuation layer, a ground evacuation layer, and a low-altitude auxiliary guidance layer. This scenario is used to illustrate the impact trends of the present invention's method on traffic costs, route reconstruction, and cross-system command consistency, and does not represent completion of engineering-level field verification. The simulation area includes one rail transit station, three entrances / exits, five road sections outside the station, and two low-altitude auxiliary guidance nodes.

[0257] The underground evacuation layer includes nodal-type underground evacuation units N1, N2, and N3, corresponding to entrance / exit A, the station hall passage, and entrance / exit B, respectively. The ground evacuation layer includes road segment-type ground evacuation units R1, R2, R3, R4, and R5. The low-altitude auxiliary guidance layer includes low-altitude auxiliary guidance nodes U1 and U2, where U1 covers R2 and R4, and U2 covers R3. The low-altitude auxiliary guidance nodes only have a corrective effect on the ground evacuation units through coverage mapping relationships and do not participate in path sequencing as path passage units.

[0258] The simulation sets up three candidate evacuation routes, as shown in Table 5:

[0259] Table 5

[0260] Candidate Path Forming evacuation units Evacuation direction P1 N1→R1→R2 Leading to Area C P2 N1→R1→R3 Leading to Area D P3 N2→R4→R5 Leading to Area E

[0261] The true cycle is 30s, the total simulation time is 20min, and the initial sudden passenger flow is 1200 people. Pedestrian movement adopts a capacity-constrained passenger flow allocation model based on discrete time steps, and evacuation route selection is updated according to the comprehensive traffic cost of candidate routes; ground traffic adopts a fixed-cycle signal model with a reference cycle of 90s, and the green light duration in the target direction is adjusted according to the control strategy in the coordinated control scheme; the low-altitude guidance coverage radius is taken as 45-50m, and the task response delay is taken as 1 control cycle. The four schemes use the same initial passenger flow, road network topology, road capacity, signal cycle, and low-altitude coverage parameters. Simulation comparison of the four control schemes:

[0262] The simulation cycle is 30 seconds, with a total duration of 20 minutes, and an initial surge of 1200 passengers. Four schemes are compared: Scheme A involves no coordinated control (only existing station guidance); Scheme B utilizes coordinated underground guidance and ground signals; Scheme C utilizes underground guidance, ground signals, and low-altitude guidance, but the low-altitude execution results are not written back to the evacuation unit status; Scheme D employs the complete method of this invention (three-layer network modeling, low-altitude guidance correction, consistency verification, and closed-loop status write-back), as detailed in Table 6.

[0263] Table 6

[0264] Scheme Number Control method A There is no coordinated control; only existing station guidance is used. B Activation of coordinated control of underground guidance and surface traffic signals C Underground guidance, ground traffic signals, and low-altitude guidance were activated, but the results of the low-altitude guidance were not updated to reflect the status of the evacuation units. D The method described in this invention enables three-layer network modeling, low-altitude guidance correction, consistency verification, and state write-back closed loop.

[0265] Figure 4The trends of cumulative evacuation ratio and overall traffic cost index G in the monitoring area under different control schemes are presented. From 0 to 300 seconds, Scheme A, lacking cross-system coordination, saw a slower increase in the cumulative evacuation ratio and maintained a high G value. From 300 to 600 seconds, Schemes B and C, due to the successive activation of underground guidance, ground signals, and low-altitude guidance, saw their cumulative evacuation ratios exceed those of Scheme A. After 600 seconds, Scheme D, by writing back the low-altitude execution results as evacuation unit state parameters and triggering path reconstruction, saw its cumulative evacuation ratio continue to rise rapidly, and its G value decreased more significantly. The summary table of indicators corresponds to completion time, maximum queue length, maximum local traffic cost, stable G value, and number of command conflicts.

[0266] The simulation results are shown in Table 7:

[0267] Table 7

[0268] Control scheme Completion time / s Maximum queue length / m Maximum local traffic cost G value after stabilization Command conflict count A 1110 52.5 0.935 0.502 3 B 900 46.6 0.888 0.473 1 C 870 42.9 0.886 0.455 1 D 720 39.3 0.883 0.437 0

[0269] Compared to the non-cooperative scheme A, scheme D using the method of this invention reduces the completion time from 1110s to 720s, the maximum queue length from 52.5m to 39.3m, and the overall traffic cost index of the monitored area after stabilization from 0.502 to 0.437, without any command conflicts. Compared to scheme C, scheme D further introduces execution result status write-back, transforming the execution results of low-altitude guidance or support tasks into evacuation unit status parameters such as on-site illumination status, communication availability status, directional consistency rate change, local density change rate, and traffic speed change rate, which participate in the recalculation of traffic cost in the next cycle. Therefore, the cumulative evacuation ratio increases faster and the final G value is lower. The above results illustrate the effectiveness of this invention.

[0270] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0271] The detailed descriptions listed above are merely specific illustrations of feasible embodiments of the present invention and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for coordinated evacuation control of sudden passenger flow in rail transit based on a three-layer evacuation network, characterized in that, Includes the following steps: S01: Acquire multi-source status data of rail transit station entrances and exits, station hall passages, road sections outside the station, ground traffic signals and low-altitude auxiliary equipment, and determine whether a sudden passenger flow event has occurred based on traffic risk status indicators; When the traffic risk status indicator meets the preset abnormal triggering conditions or receives an emergency triggering signal, it enters the dynamic collaborative evacuation control mode. S02: A three-layer evacuation network is constructed under the dynamic collaborative evacuation control mode. The three-layer evacuation network includes an underground evacuation layer, a ground evacuation layer, and a low-altitude auxiliary guidance layer. Among them, the key passages of rail transit entrances or station halls in the underground evacuation layer are mapped as node-type underground evacuation units, and the road segments outside the station in the ground evacuation layer are mapped as road segment-type ground evacuation units. The node-type underground evacuation units and road segment-type ground evacuation units serve as the calculation subjects for passage cost calculation, path sequencing, and risk assessment. The low-altitude auxiliary guidance nodes in the low-altitude auxiliary guidance layer do not serve as path passage units. They only affect the state correction of the road segment-type ground evacuation units within their coverage area through the coverage mapping relationship, and do not participate in passage cost calculation and path sequencing. S03: Perform risk-based normalization on the multi-source state parameters of each evacuation unit and calculate the basic passage cost; based on the coverage mapping relationship between the low-altitude auxiliary guidance node and the road segment ground evacuation unit, calculate the low-altitude guidance correction factor and the low-altitude guidance anomaly penalty term, and obtain the comprehensive passage cost at the evacuation unit level accordingly. S04: Based on the comprehensive travel cost of the nodal underground evacuation units and road segment ground evacuation units contained in the candidate evacuation routes, calculate the cumulative travel cost of the candidate evacuation routes, and combine the peak local travel cost of key evacuation units to obtain the overall travel cost index of the monitoring area, and determine the evacuation risk level and control level. S05: Generate a control strategy parameter set based on the evacuation risk level, control level, candidate path priority sequence, path passability status, and execution resource status. The control strategy parameter set includes underground guidance parameters, ground traffic signal parameters, and low-altitude mission parameters. S06: Map the control strategy parameter set to control commands for underground guidance systems, ground traffic signal systems, and low-altitude equipment platforms, and perform cross-subsystem consistency verification before issuing them; if the verification passes, the three types of control commands will carry the same control cycle identifier, strategy version number, and receipt identifier and be issued uniformly in the same effective cycle; if the verification fails, it will be handled according to the preset conflict handling rules. S07: Receive control command execution receipts and field feedback results, write back the execution results of low-altitude guidance or support missions as evacuation unit status parameters, and recalculate passage costs and reconstruct candidate evacuation paths based on the written-back status parameters; when the closed-loop reconstruction triggering conditions, including time period triggering conditions or status change triggering conditions, are met, S02 to S06 are re-executed.

2. The method according to claim 1, characterized in that, The traffic risk status index is obtained by aggregating the unit risk index of at least one evacuation unit within the monitoring area. The unit risk index is obtained by risk-normalizing and weighting at least two of the following: crowd density, queue length, average traffic speed, number of stranded people, and traffic saturation of ground road sections. The preset abnormal triggering conditions include the passage risk status indicator exceeding the triggering threshold, or any one of the following indicators exceeding the corresponding threshold: crowd density at key entrances / exits, average passage speed decrease rate, queue length, or number of stranded people.

3. The method according to claim 1, characterized in that, The multi-source status data includes data from the automatic fare collection system, video passenger flow recognition data, ground traffic signal control data, road traffic flow detection data, low-altitude equipment sensing data, and low-altitude equipment task execution feedback data. When there is a time discrepancy between different data sources, time synchronization is performed according to a unified timestamp, and fusion is performed based on equipment confidence, data freshness, and spatial coverage. When any multi-source status data is missing, delayed, or abnormal, alternative status parameters are generated based on the data source confidence, the most recent valid sample value, and the status estimate of adjacent evacuation units, and the status confidence of the corresponding evacuation unit is reduced.

4. The method according to claim 1, characterized in that, The ground connection points, street crossings, intersections, exit directions, target areas, guide screen locations, and low-altitude equipment hovering points in the three-layer evacuation network are only used as topology connection objects, control objects, display objects, or location identifiers, and are not used as independent evacuation units to participate in the calculation of basic passage cost, the calculation of cumulative passage cost of the path, the calculation of the overall passage cost index of the monitoring area, and the determination of risk level. For nodal-type underground evacuation units, at least two of the following parameters should be selected as traffic status characterization parameters: crowd density, queue length, average passage speed, and number of stranded people. For road segment-type ground evacuation units, at least two of the following parameters should be selected as traffic status characterization parameters: traffic flow saturation, remaining capacity ratio, degree of intersection conflict, road segment length, expected passage time, and route passability.

5. The method according to claim 1, characterized in that, The low-altitude guidance correction factor η q According to the set of low-altitude equipment D covering evacuation unit q q Spatial influence coefficient α q,d Low-altitude equipment efficiency index γ d Effective guided response coefficient ρ q,d and available state coefficients s d Calculate, and adjust by the upper limit of the correction factor η max Amplitude limiting is applied: ， In the formula: D q A collection of low-altitude equipment covering evacuation unit q; s d This is the availability coefficient for low-altitude equipment, where 0 indicates unavailable and 1 indicates available. η max To adjust the upper limit of the factor; When multiple low-altitude auxiliary guidance nodes cover the same road segment ground evacuation unit, the optimal mode, amplitude-limiting overlay mode, or conflict suppression mode is selected based on the consistency of guidance direction, overlap of coverage areas, direction of ground traffic signal release, and equipment availability to determine the summation term. The specific calculation method.

6. The method according to claim 5, characterized in that, The efficiency index of low-altitude equipment It is obtained by weighting three factors: brightness, coverage area, and height effect. ， in: This represents the comprehensive performance index of the low-altitude equipment d, with a value range of 0 to 1; , , The weighting coefficients of the three indicators satisfy the following conditions: ; This is the normalized value for the projected brightness or directional sign contrast of the low-altitude device d; This is the normalized value of the area of ​​the effective guidance zone formed by the low-altitude equipment d on the ground; Let d be the normalized value of the hovering height effect of the low-altitude equipment. The formula for calculating the value is: ， Where: H d H is the current hovering altitude of the low-altitude equipment. max and H min These are the upper and lower limits of the effective height range, respectively; The spatial influence coefficient α of low-altitude equipment d on evacuation unit q q,d The calculation formula depends on the ratio of the ground projection distance to the effective guiding radius, as follows: ， In the formula: dist q,d The distance from the geometric center of the evacuation unit to the ground projection center of the low-altitude equipment; R d The effective guiding radius of the low-altitude equipment d; Effective guided response coefficient ρ q,d It is determined by a combination of factors including successful execution status, changes in directional consistency rate, density changes, and velocity changes. ， In the formula: ρ q,d This represents the effective guidance response coefficient of the low-altitude equipment d to the evacuation unit q; ACK d This indicates the successful execution status of the low-altitude device d. The value is 1 when the execution is successful, and 0 when the execution fails, times out, or the receipt is missing. ΔDir q is the periodic change in the consistency rate between the actual movement direction of the crowd and the low-altitude guidance indication direction within the coverage area of ​​evacuation unit q. ΔDen q Let q be the rate of change in population density within the coverage area of ​​evacuation unit q; Let q be the rate of change in the average passage speed of the population within the coverage area of ​​evacuation unit q.

7. The method according to claim 5, characterized in that, The low-altitude guidance anomaly penalty item This low-altitude guidance anomaly penalty term is generated when at least one of the following occurs: low-altitude assisted guidance node execution failure, timeout, missing feedback, coverage deviation, waypoint deviation, decreased direction consistency rate, increased local density, decreased travel speed, or deviation of the path selection ratio from the target path. It is used as a proportional penalty factor in the comprehensive travel cost calculation of the corresponding road segment-type ground evacuation unit. The calculation is as follows: ， In the formula: This represents the abnormal penalty term generated by the low-altitude auxiliary guidance node for evacuation unit q; F d This indicates that the low-altitude device d has failed to execute its task, timed out, or is missing feedback. O q This indicates a coverage deviation or waypoint deviation status; b1, b2, b3, b4, and b5 are coefficient weights.

8. The method according to claim 1, characterized in that, The cumulative travel cost A of the candidate evacuation routes p According to the comprehensive traffic value C at the evacuation unit level q Length L of evacuation unit q Expected passage time, passage impedance or remaining passage capacity K q We obtain the following by weighted summation: ， Among them, P p L is the set of all evacuation units contained in path p; q K represents the length of the evacuation unit or the estimated travel time. q L represents the remaining capacity or capacity factor; in the same path calculation, L q and K q They are converted into dimensionless scale coefficients or uniform pass-through impedance coefficients of the same scale.

9. The method according to claim 1, characterized in that, The overall traffic cost index G of the monitored area is accumulated through the traffic cost A of candidate evacuation routes. p Peak local access costs for key evacuation units The result of fusion: ， In the formula: β1 and β2 are the fusion weights, and β1 + β2 = 1; A ref For reference, the cumulative cost value; K represents the set of key evacuation units; r is the total number of feasible paths in the candidate path set; π p Let p be the fusion weight of the p-th candidate path in the overall regional cost; G represents the overall traffic cost index for the monitored area, with a value ranging from 0 to 1. The key evacuation units include node-type underground evacuation units associated with the exit direction, road segment-type ground evacuation units corresponding to cross-street connection bottlenecks, high-load evacuation units in the current main evacuation route, and evacuation units that rank in the top m in terms of comprehensive traffic value or exceed the p' quantile of historical safety samples in the current assessment period, where m and p' are preset parameters.

10. A system for a collaborative evacuation control method for sudden passenger flow in rail transit based on a three-layer evacuation network according to any one of claims 1-9, characterized in that, include: The multi-source data access unit is used to access data from rail transit automatic fare collection systems, video surveillance systems, ground traffic signal control systems, road traffic detection equipment, and low-altitude equipment platforms, and to perform time synchronization, spatial coordinate transformation, and data confidence assessment, and output multi-source status data. The event determination and mode switching unit is used to receive the multi-source status data, determine whether a sudden passenger flow event has occurred based on the traffic risk status index, and output a switching signal to enter the dynamic collaborative evacuation control mode when the preset abnormal triggering conditions are met or an emergency triggering signal is received. The three-layer network modeling unit is used to construct a three-layer evacuation network including an underground evacuation layer, a ground evacuation layer, and a low-altitude auxiliary guidance layer under the dynamic collaborative evacuation control mode. In the underground evacuation layer, the rail transit entrances or key passages in the station hall are mapped as node-type underground evacuation units, the road segments outside the station in the ground evacuation layer are mapped as road segment-type ground evacuation units, and the low-altitude auxiliary guidance nodes in the low-altitude auxiliary guidance layer are associated with the road segment-type ground evacuation units within their coverage area through the coverage mapping relationship, and the network topology is output. The passage cost calculation unit is used to receive the network topology and multi-source state data, perform risk normalization processing on the state parameters of each evacuation unit and calculate the basic passage cost value, calculate the low-altitude guidance correction factor and low-altitude guidance anomaly penalty item based on the coverage mapping relationship between the low-altitude auxiliary guidance node and the road segment ground evacuation unit, and then obtain the comprehensive passage cost value at the evacuation unit level, and output it to the risk assessment unit. The risk assessment unit is used to calculate the cumulative passage cost of candidate evacuation routes based on the comprehensive passage cost of nodal underground evacuation units and road segment ground evacuation units contained in the candidate evacuation routes, and to obtain the overall passage cost index of the monitoring area by combining the local passage cost peak of key evacuation units, determine the evacuation risk level and control level, and output it to the control strategy generation unit. The control strategy generation unit is used to generate a control strategy parameter set based on the evacuation risk level, control level, candidate path priority sequence, path passability status and execution resource status. The control strategy parameter set includes underground guidance parameters, ground traffic signal parameters and low-altitude mission parameters, and is output to the collaborative control interface unit. The collaborative control interface unit is used to map the control strategy parameter set into control commands for underground guidance systems, ground traffic signal systems, and low-altitude equipment platforms, and to perform cross-subsystem consistency verification before issuing commands. When the verification passes, the three types of control commands carry the same control cycle identifier, strategy version number, and receipt identifier and are issued uniformly in the same effective cycle. If the verification fails, it will be handled according to the preset conflict handling rules. Simultaneously, the issued results and execution receipts will be fed back to the closed-loop control unit; The closed-loop control unit is used to receive control command execution acknowledgments and field feedback results, write back the execution results of low-altitude guidance or support tasks as evacuation unit status parameters, and trigger the passage cost calculation unit and risk assessment unit to recalculate passage costs and reconstruct candidate evacuation paths; when the closed-loop reconstruction triggering conditions are met, the three-layer network modeling unit is re-triggered to the collaborative control interface unit to execute the next cycle of collaborative control.

Images