Intelligent evacuation path automatic generation system and method based on artificial intelligence

Abstract

The present invention belongs to the technical field of intelligent evacuation, and discloses an artificial intelligence-based intelligent evacuation path automatic generation system and method; the method comprises: collecting building modeling data, wherein the building modeling data comprises key building parameters and equipment resource parameters; modeling is performed according to the building modeling data to obtain a building simulation model; defining personnel characteristic parameters and disaster characteristic parameters on the building simulation model to obtain an intelligent evacuation simulation model; and adaptively generating an evacuation path according to the intelligent evacuation simulation model; the solution can quickly respond to congestion conditions in a dynamic environment, and by combining diversion optimization and priority setting, it maximizes the efficiency and safety of evacuation, is applicable to a variety of complex buildings and disaster scenarios, and has significant practical value.

Description

Technical Field

[0001] The present invention relates to the field of intelligent evacuation technology, and more specifically, to an intelligent evacuation path automatic generation system and method based on artificial intelligence. Background Art

[0002] In sudden disasters such as fires, earthquakes, and floods, how to quickly and safely evacuate people from affected areas is a core issue for ensuring life safety. However, traditional evacuation planning often relies on fixed route designs and fails to fully consider real-time changes in the disaster environment and the behavioral characteristics of people. This can lead to problems such as congestion, blocked routes, and prolonged evacuation times during actual evacuations. Furthermore, traditional methods lack efficient dynamic adjustment capabilities when dealing with complex building environments, making it impossible to fully optimize evacuation routes and assign priorities, thereby reducing evacuation efficiency. Therefore, an intelligent evacuation route automatic generation system is designed to automatically simulate and evaluate evacuation routes in various disaster scenarios such as fires, earthquakes, and floods.

[0003] A Chinese patent application with publication number CN109992876A proposes a method, device, and computer equipment for processing fire evacuation paths in buildings. The method includes obtaining the structural parameters of the building's rooms, determining evacuation points and exits of the evacuation space based on the structural parameters, generating evacuation paths from each evacuation point to the exit based on the room structural parameters, and determining, based on the length of the evacuation path, whether the generated evacuation path complies with regulations. This method automatically draws evacuation route maps during the three-dimensional design process and automatically detects compliance with relevant fire regulations, enabling building information models to meet the needs of fire protection design.

[0004] Although the above methods can meet most scenarios, research and practical application of the above methods and existing technologies have revealed that the above methods and existing technologies have at least the following defects:

[0005] Traditional evacuation planning typically relies on fixed route designs, making it difficult to cope with the dynamic changes in real-time disaster environments and the complexity of human behavior, often leading to problems such as congestion, blocked routes, and evacuation delays. For example, people on some floors may have smooth passage due to priority evacuation, while people on other floors may be stranded due to blocked passages, significantly increasing the difficulty of escape. In addition, this method lacks the ability to dynamically adjust in complex building environments. It is unable to optimize evacuation routes and priorities, and it is difficult to provide targeted evacuation guidance for special groups (such as those with limited mobility, the elderly, and children), further delaying the overall evacuation time and even endangering life safety.

[0006] In view of this, the present invention proposes an intelligent evacuation path automatic generation system and method based on artificial intelligence to solve the above problems. Summary of the Invention

[0007] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: an artificial intelligence-based intelligent evacuation path automatic generation method, comprising:

[0008] Collecting building modeling data, wherein the building modeling data includes key building parameters and equipment resource parameters;

[0009] Modeling is performed according to the building modeling data to obtain a building simulation model;

[0010] Defining personnel characteristic parameters and disaster characteristic parameters on the building simulation model to obtain an intelligent evacuation simulation model;

[0011] Evacuation routes are adaptively generated based on the intelligent evacuation simulation model.

[0012] Furthermore, the method for adaptively generating an evacuation route based on the intelligent evacuation simulation model includes:

[0013] Step 1: Preset the total evacuation time and time step; take the quotient of the total evacuation time and the time step, and round up the result to obtain H time steps; the initial value of h is preset to 1, and the value range of h is 1 to H; add all evacuation network edges to the set of edges to be processed; the evacuation network edges are channels in the building simulation model;

[0014] Step 2: Starting from the highest floor of the building structure, the evacuation status is judged in sequence. If all personnel on the current floor have been evacuated, the evacuation network edge connected to the evacuation network node corresponding to the current floor is removed from the set of pending edges. The judgment is continued downward layer by layer until a floor that has not yet been evacuated is encountered, and the judgment operation is terminated to obtain the set of pending edges of the hth time step.

[0015] Step 3: Preset congestion weight threshold 1 and congestion weight threshold 2, where congestion weight threshold 1 is greater than congestion weight threshold 2; analyze and process the evacuation network edges in the edge set to be processed based on the congestion weight threshold 1 and the congestion weight threshold 2, and obtain the congested edge set, risk edge set, and smooth edge set at the hth time step;

[0016] Step 4: Plan the evacuation path for the evacuation network edge based on the congestion edge set, risk edge set, and smooth edge set at the hth time step;

[0017] Step 5: Let h=h+1. If h is less than or equal to H, continue to execute steps 2 to 4. If h is greater than H, end the current process.

[0018] Furthermore, the method for planning an evacuation path for the evacuation network edges according to the congested edge set, the risk edge set, and the smooth edge set includes:

[0019] Perform the first flow diversion on the evacuation network edge in the congested edge set according to the preset evacuation method;

[0020] The evacuation network edges in the risk edge set are analyzed and processed according to the preset processing method to obtain the secondary congestion edge set. If the number of evacuation network edges in the secondary congestion edge set is greater than zero, the evacuation network edges in the secondary congestion edge set are subjected to a second flow diversion using the same method as the first flow diversion. If the number of evacuation network edges in the secondary congestion edge set is equal to zero, the current process ends.

[0021] The same method as the second congestion edge set is used to obtain the third congestion edge set. If the number of evacuation network edges in the third congestion edge set is greater than zero, the third flow diversion is performed on the evacuation network edges in the third congestion edge set using the same method as the first flow diversion. If the number of evacuation network edges in the third congestion edge set is zero, the current process ends.

[0022] The current number of evacuees, individual attributes, maximum number of passers-by, and disaster characteristic parameters corresponding to each evacuation network edge in the congestion edge set, risk edge set, and smooth edge set are input into the channel weight model respectively to obtain the evacuation network edge weight corresponding to each evacuation network edge in the congestion edge set, risk edge set, and smooth edge set;

[0023] Evacuate personnel along the evacuation network edges according to the evacuation network edge weights.

[0024] Furthermore, the method for evacuating the evacuees in the evacuation network edge according to the evacuation network edge weight includes:

[0025] If there is an evacuation network edge with a weight greater than or equal to the congestion weight threshold 1, the evacuation network edge with a weight greater than or equal to the congestion weight threshold 1 is added to the weight setting set, and the evacuation priority is set for the individual attribute of the evacuees corresponding to the evacuation network edge in the weight setting set; the evacuees with an evacuation priority of level 3 are evacuated from the evacuation network edge in the weight setting set, and the evacuees with level 1 and level 2 enter the floor buffer zone and wait. After the evacuees with level 3 are evacuated, the evacuees with level 1 and level 2 are evacuated;

[0026] If there is no evacuation network edge whose evacuation network edge weight is greater than or equal to the congestion weight threshold 1, the current process ends.

[0027] Furthermore, the method of setting the evacuation priority for the individual attributes of the evacuees corresponding to the evacuation network edges in the weight setting set includes:

[0028] S140: Record the number of evacuation network edges in the weight setting set as QZ, preset the initial value of qz to be 1, and the value range of qz is 1 to QZ;

[0029] S141: Obtain the qzth evacuation network edge from the weight setting set, obtain the current number of evacuees in the qzth evacuation network edge, recorded as P; input the current number of evacuees, the maximum number of passers, and the individual attributes corresponding to the P evacuees into the weight setting model respectively, and obtain the evacuation priority corresponding to the P evacuees in the qzth evacuation network edge;

[0030] S142: Let qz = qz + 1. If qz is less than or equal to QZ, continue to execute S141. If qz is greater than QZ, end the current process.

[0031] Furthermore, the method for obtaining the congested edge set, the risk edge set, and the smooth edge set includes:

[0032] S110: Record the number of evacuation network edges in the edge set to be processed as TD, preset the initial value of td to be 1, and the value range of td is 1 to TD;

[0033] S111: Get the current number of evacuees and individual attributes in the tdth evacuation network edge;

[0034] S112: Input the current number of evacuees, individual attributes, maximum number of passers and disaster characteristic parameters into the channel weight model to obtain the evacuation network edge weight of the td-th evacuation network edge;

[0035] S113: If the evacuation network edge weight of the tdth evacuation network edge is greater than or equal to the congestion weight threshold 1, then the tdth evacuation network edge is added to the congested edge set; if the evacuation network edge weight of the tdth evacuation network edge is greater than or equal to the congestion weight threshold 2 and less than the congestion weight threshold 1, then the tdth evacuation network edge is added to the risk edge set; if the evacuation network edge weight of the tdth evacuation network edge is less than the congestion weight threshold 2, then the tdth evacuation network edge is added to the unobstructed edge set;

[0036] S114: Let td=td+1. If td is less than or equal to TD, continue executing S111 to S113; if td is greater than TD, end the current process.

[0037] Furthermore, the method of performing the first pedestrian flow diversion on the evacuation network edge in the congested edge set according to the preset evacuation method includes:

[0038] S120: If the number of evacuation network edges in the congested edge set is not zero, the number of evacuation network edges in the congested edge set is recorded as YD, where the initial value of yd is preset to 1 and the value range of yd is 1 to YD; if the number of evacuation network edges in the congested edge set is zero, the current process ends;

[0039] S121: Obtain the ydth evacuation network edge from the congestion edge set, and add the evacuation network edges that are adjacent to the ydth evacuation network edge and are not in the congestion edge set to the evacuation diversion set corresponding to the ydth evacuation network edge; if the number of evacuation network edges in the evacuation diversion set corresponding to the ydth evacuation network edge is not zero, calculate the evacuation flow rate shared by each evacuation network edge in the evacuation diversion set, and calculate the latest current number of evacuees for each evacuation network edge in the evacuation diversion set based on the evacuation flow rate shared by each evacuation network edge; update the current number of evacuees for each evacuation network edge in the evacuation diversion set to the corresponding latest current number of evacuees;

[0040] S122: Calculate the latest current number of evacuees on the yd-th evacuation network edge in the congestion edge set based on the evacuation flow shared by each evacuation network edge in the evacuation diversion set, and update the current number of evacuees on the yd-th evacuation network edge in the congestion edge set to the latest current number of evacuees;

[0041] S123: Let yd=yd+1. If yd is less than or equal to YD, continue executing S121 to S122. If yd is greater than YD, end the current process.

[0042] Furthermore, the calculation method of the evacuation flow shared by each evacuation network edge in the evacuation flow distribution set includes:

[0043] The evacuation flow rate shared by each evacuation network edge is calculated proportionally according to the residual capacity of each evacuation network edge; the residual capacity of each evacuation network edge is the maximum number of passers-by corresponding to each evacuation network edge minus the current number of evacuees.

[0044] Furthermore, the method of analyzing and processing the evacuation network edges in the risk edge set according to a preset processing method to obtain the secondary congestion edge set includes:

[0045] S130: If the number of evacuation network edges in the risk edge set is not zero, the number of evacuation network edges in the risk edge set is recorded as FX, where the initial value of fx is preset to 1 and the value range of fx is 1 to FX; if the number of evacuation network edges in the risk edge set is zero, the current process ends;

[0046] S131: Obtain the fxth evacuation network edge from the risk edge set, input the current number of evacuees, individual attributes, maximum number of passers-by, and disaster characteristic parameters of the fxth evacuation network edge into the channel weight model, and obtain the evacuation network edge weight of the fxth evacuation network edge;

[0047] S132: If the evacuation network edge weight of the fxth evacuation network edge is greater than or equal to the congestion weight threshold 1, then add the fxth evacuation network edge to the secondary congested edge set;

[0048] S133: Let fx=fx+1. If fx is less than or equal to FX, continue executing S131 to S132; if fx is greater than FX, end the current process.

[0049] An artificial intelligence-based intelligent evacuation path automatic generation system implements the artificial intelligence-based intelligent evacuation path automatic generation method, including:

[0050] A first acquisition module is used to acquire building modeling data, wherein the building modeling data includes key building parameters and equipment resource parameters;

[0051] A data modeling module is used to perform modeling based on building modeling data to obtain a building simulation model;

[0052] A parameter definition module is used to define personnel characteristic parameters and disaster characteristic parameters on the building simulation model to obtain an intelligent evacuation simulation model;

[0053] Intelligent evacuation module, used to adaptively generate evacuation routes based on the intelligent evacuation simulation model;

[0054] The evacuation assessment module is used to summarize and output the evacuation simulation results after the evacuation simulation is completed.

[0055] The technical effects and advantages of the artificial intelligence-based intelligent evacuation path automatic generation system and method of the present invention are as follows:

[0056] By real-time collection and analysis of the congestion status, risk status, and smooth status of the evacuation network edges, combined with a three-iteration diversion mechanism, a rapid response to the dynamic environment is achieved, effectively alleviating the local congestion problem caused by high-density crowds. The individual attributes of the evacuees (such as age, gender, special population classification, etc.) are analyzed through a weight setting model, and the evacuation priority is accurately set to provide targeted evacuation guidance for special groups, reflecting a high level of attention to safety and fairness. The introduction of a floor buffer mechanism temporarily accommodates people with lower evacuation priorities, avoiding overload and congestion while further ensuring the orderliness and continuity of the evacuation process. By real-time collection and accurate aggregation of data during the evacuation process, an intuitive and usable evacuation report is generated to support users in comprehensively evaluating the advantages and disadvantages of the evacuation plan; at the same time, the data-driven feedback mechanism promotes the continuous optimization of path planning.

[0057] This solution can quickly respond to congestion in dynamic environments. By combining diversion optimization and priority setting, it maximizes the efficiency and safety of evacuation. It is applicable to a variety of complex buildings and disaster scenarios and has significant practical value and innovation. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 This is a schematic diagram of an intelligent evacuation route automatic generation system based on artificial intelligence according to Example 1 of the present invention;

[0059] Figure 2 This is a flow chart of a method for automatically generating an intelligent evacuation path based on artificial intelligence according to embodiment 2 of the present invention;

[0060] Figure 3 A flow chart of a method for adaptively generating evacuation routes based on an intelligent evacuation simulation model;

[0061] Figure 4 The flowchart of the method for planning evacuation paths for evacuation network edges based on congested edge sets, risk edge sets, and unobstructed edge sets is provided. DETAILED DESCRIPTION

[0062] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0063] Example 1

[0064] See also Figure 1As shown, the artificial intelligence-based intelligent evacuation path automatic generation system described in this embodiment includes a first acquisition module, a data modeling module, a parameter definition module, an intelligent evacuation module and an evacuation evaluation module. Each module realizes data transmission through wired and / or wireless connections.

[0065] The first acquisition module is used to collect building modeling data, wherein the building modeling data includes key building parameters and equipment resource parameters; the key building parameters include floor height, structural layout, number of channels and exit locations; the equipment resource parameters include alarm layout, emergency lighting equipment layout and fire equipment layout.

[0066] The data modeling module is used to perform modeling based on the building modeling data to obtain a building simulation model.

[0067] The method for obtaining the building simulation model includes:

[0068] Import key building parameters into the building plan, identify walls, rooms, corridors, stairs, passages, and exits in the building plan; assign key building parameters to corresponding building areas, set the maximum number of people allowed in each passage, and obtain a floor plan with physical meaning;

[0069] Import equipment resource parameters into the floor plan and mark the locations of alarms, emergency lighting equipment, and firefighting equipment in the floor plan;

[0070] Based on the floor plan, building information modeling (BIM) technology is used to connect the floors to form a complete three-dimensional building space. The alarms, emergency lighting equipment, and fire-fighting equipment with marked positions are imported into the three-dimensional building space, and their coverage and function triggering conditions are defined to obtain a building simulation model.

[0071] The parameter definition module is used to define personnel characteristic parameters and disaster characteristic parameters on the building simulation model to obtain an intelligent evacuation simulation model.

[0072] The method for obtaining the intelligent evacuation simulation model includes:

[0073] S100: Obtain the floor height in the building simulation model, recorded as LG; the initial value of lg is preset to 1, and the value range of lg is 1 to LG;

[0074] S101: Setting personnel characteristic parameters for the lg floor on the building simulation model; the personnel characteristic parameters include the initial distribution position and individual attributes of the evacuated personnel on the lg floor; the individual attributes include gender, age, movement speed, evacuation priority, and the group category to which the individual belongs; the evacuation priority includes a first evacuation level, a second evacuation level, and a third evacuation level, and the higher the evacuation priority level, the higher the evacuation priority level;

[0075] It should be noted that the population categories are used to distinguish between ordinary people and special groups. The classification of special groups includes people with limited mobility (such as people using wheelchairs or crutches), the elderly (such as those over 65 years old), and children (such as those under 14 years old). In addition, the mobility characteristic parameters of special groups can be further refined, for example:

[0076] Mobility: such as walking normally, moving slowly, or completely relying on assistive devices;

[0077] Evacuation priority: Define the priority of special groups in evacuation based on their physiological or psychological characteristics;

[0078] Assistance needs: such as whether human assistance or professional equipment support is required.

[0079] By comprehensively considering the classification and proportion of gender, age, number of people and special groups, the needs and behavior patterns of different types of people during the evacuation process can be more accurately reflected, thereby optimizing the evacuation route design, improving the overall evacuation efficiency, and providing targeted solutions for the safe evacuation of special groups.

[0080] S102: Let lg = lg + 1. If lg is less than or equal to LG, continue to execute S101. If lg is greater than LG, execute S103.

[0081] S103: Setting disaster characteristic parameters on the building simulation model, wherein the disaster characteristic parameters include disaster type, disaster severity level, and disaster hazard; marking the exit locations in the building simulation model as evacuation network nodes, marking the channels in the building simulation model as evacuation network edges, constructing all evacuation network nodes and evacuation network edges into an evacuation network, connecting the exit nodes through edges, and obtaining an intelligent evacuation simulation model.

[0082] It should be noted that disaster types include fire, earthquake and flood, and disaster severity levels include level one, level two and level three. The higher the level of disaster severity, the more severe the disaster; disaster hazards include high temperature, suffocation and slipping.

[0083] The intelligent evacuation module is used to adaptively generate evacuation routes based on the intelligent evacuation simulation model.

[0084] like Figure 3 As shown, the method for adaptively generating an evacuation path based on the intelligent evacuation simulation model includes:

[0085] Step 1: Preset the total evacuation time and time step; take the quotient of the total evacuation time and the time step, and round up the result to obtain H time steps; the initial value of h is preset to 1, and the value range of h is 1 to H; add all evacuation network edges to the set of edges to be processed; the evacuation network edges are channels in the building simulation model;

[0086] The method for obtaining the H time steps includes:

[0087]

[0088] Among them, Time max is the total evacuation time, Time step is the time step, Indicates rounding up.

[0089] Step 2: Starting from the highest floor of the building structure, the evacuation status is judged in sequence. If all personnel on the current floor have been evacuated, the evacuation network edge connected to the evacuation network node corresponding to the current floor is removed from the set of pending edges. This process continues downward, layer by layer, until a floor is encountered where evacuation has not yet been completed, and the judgment operation is terminated, resulting in the set of pending edges for the hth time step.

[0090] Step 3: Preset congestion weight threshold 1 and congestion weight threshold 2, where congestion weight threshold 1 is greater than congestion weight threshold 2; analyze and process the evacuation network edges in the edge set to be processed based on the congestion weight threshold 1 and the congestion weight threshold 2, and obtain the congested edge set, risk edge set, and smooth edge set at the hth time step;

[0091] Step 4: Plan the evacuation path for the evacuation network edge based on the congestion edge set, risk edge set, and smooth edge set at the hth time step;

[0092] Step 5: Let h=h+1. If h is less than or equal to H, continue to execute steps 2 to 4. If h is greater than H, end the current process.

[0093] The method for obtaining the congested edge set, the risk edge set, and the smooth edge set includes:

[0094] S110: Record the number of evacuation network edges in the edge set to be processed as TD, preset the initial value of td to be 1, and the value range of td is 1 to TD;

[0095] S111: Get the current number of evacuees and individual attributes in the tdth evacuation network edge;

[0096] S112: Input the current number of evacuees, individual attributes, maximum number of passers and disaster characteristic parameters into the channel weight model to obtain the evacuation network edge weight of the td-th evacuation network edge;

[0097] S113: If the evacuation network edge weight of the tdth evacuation network edge is greater than or equal to the congestion weight threshold 1, then the tdth evacuation network edge is added to the congested edge set; if the evacuation network edge weight of the tdth evacuation network edge is greater than or equal to the congestion weight threshold 2 and less than the congestion weight threshold 1, then the tdth evacuation network edge is added to the risk edge set; if the evacuation network edge weight of the tdth evacuation network edge is less than the congestion weight threshold 2, then the tdth evacuation network edge is added to the unobstructed edge set;

[0098] S114: Let td=td+1. If td is less than or equal to TD, continue executing S111 to S113; if td is greater than TD, end the current process.

[0099] The training method of the channel weight model includes:

[0100] A weight evaluation data set is collected in advance, and the weight evaluation data set includes G groups of weight evaluation data and evacuation network edge weights corresponding to the G groups of weight evaluation data, where G is a positive integer greater than 0, and the weight evaluation data includes the current number of evacuees, individual attributes, the maximum number of passers-by, and disaster characteristic parameters; the evacuation network edge weights are evaluated by multiple technical personnel in this field based on the actual weight evaluation data, and corresponding scores are performed. After removing the maximum and minimum values, the average value is taken to obtain the evacuation network edge weights; the weight evaluation data set is divided into a training set and a test set, and the weight evaluation data in the training set is used as the input of the channel weight model, and the evacuation network edge weights in the training set are used as the output of the channel weight model, with minimizing the sum of the prediction accuracies of all predicted evacuation network edge weights as the training goal; training is stopped when the sum of the prediction accuracies converges; the channel weight model is a naive Bayes model or a support vector machine model.

[0101] like Figure 4 As shown in FIG, the method for planning evacuation paths for evacuation network edges based on the congestion edge set, risk edge set, and smooth edge set includes:

[0102] Perform the first flow diversion on the evacuation network edge in the congested edge set according to the preset evacuation method;

[0103] The evacuation network edges in the risk edge set are analyzed and processed according to the preset processing method to obtain the secondary congestion edge set. If the number of evacuation network edges in the secondary congestion edge set is greater than zero, the evacuation network edges in the secondary congestion edge set are subjected to a second flow diversion using the same method as the first flow diversion. If the number of evacuation network edges in the secondary congestion edge set is equal to zero, the current process ends.

[0104] The same method as the second congestion edge set is used to obtain the third congestion edge set. If the number of evacuation network edges in the third congestion edge set is greater than zero, the third flow diversion is performed on the evacuation network edges in the third congestion edge set using the same method as the first flow diversion. If the number of evacuation network edges in the third congestion edge set is zero, the current process ends.

[0105] The current number of evacuees, individual attributes, maximum number of passers-by, and disaster characteristic parameters corresponding to each evacuation network edge in the congestion edge set, risk edge set, and smooth edge set are input into the channel weight model respectively to obtain the evacuation network edge weight corresponding to each evacuation network edge in the congestion edge set, risk edge set, and smooth edge set;

[0106] If there is an evacuation network edge with a weight greater than or equal to the congestion weight threshold of one, it indicates that after three diversions, there is still an evacuation network edge in a congested state, then the evacuation network edge with a weight greater than or equal to the congestion weight threshold of one is added to the weight setting set, and the evacuation priority is set for the individual attributes of the evacuees corresponding to the evacuation network edge in the weight setting set; the evacuees with an evacuation priority of the third evacuation level are evacuated from the evacuation network edge in the weight setting set, and the evacuees with a non-third evacuation level enter the floor buffer zone and wait. After the evacuees with the third evacuation level are evacuated, the evacuees with the non-third evacuation level are evacuated; if there is no evacuation network edge with a weight greater than or equal to the congestion weight threshold of one, the current process is terminated.

[0107] It should be noted that by sequentially diverting congested and risky edge sets, congestion caused by high-density crowd flow is effectively alleviated, ensuring the dynamic controllability of the evacuation process. The three-iterative diversion mechanism gradually reduces congestion risks, avoids localized congestion caused by a single diversion, dynamically adjusts evacuation paths, improves overall evacuation efficiency, and makes path planning more adaptable and able to respond to dynamic environmental changes in real time. After the diversion is completed, a channel weight model is introduced to quantitatively evaluate the evacuation network edges. This model comprehensively considers the current number of evacuees, individual attributes, maximum number of passersby, and disaster characteristic parameters, providing a reliable quantitative basis for subsequent evacuation path planning and achieving precise evacuation path optimization based on real-time data.

[0108] Setting evacuation priorities based on the individual attributes of evacuees demonstrates care for special groups and helps improve evacuation safety and fairness. Prioritizing evacuees with high evacuation priorities at the edges of the evacuation network effectively reduces crowd pressure in high-density areas. Temporarily placing those with low evacuation priorities in floor buffer zones prevents corridor overload and improves the controllability of evacuation order.

[0109] Methods for performing the first flow diversion on the evacuation network edges in the congested edge set according to the preset evacuation method include:

[0110] S120: If the number of evacuation network edges in the congested edge set is not zero, the number of evacuation network edges in the congested edge set is recorded as YD, where the initial value of yd is preset to 1 and the value range of yd is 1 to YD; if the number of evacuation network edges in the congested edge set is zero, the current process ends;

[0111] S121: Obtain the ydth evacuation network edge from the congestion edge set, and add the evacuation network edges that are adjacent to the ydth evacuation network edge and are not in the congestion edge set to the evacuation diversion set corresponding to the ydth evacuation network edge; if the number of evacuation network edges in the evacuation diversion set corresponding to the ydth evacuation network edge is not zero, calculate the evacuation flow rate shared by each evacuation network edge in the evacuation diversion set, and calculate the latest current number of evacuees for each evacuation network edge in the evacuation diversion set based on the evacuation flow rate shared by each evacuation network edge; update the current number of evacuees for each evacuation network edge in the evacuation diversion set to the corresponding latest current number of evacuees;

[0112] S122: Calculate the latest current number of evacuees on the yd-th evacuation network edge in the congestion edge set based on the evacuation flow shared by each evacuation network edge in the evacuation diversion set, and update the current number of evacuees on the yd-th evacuation network edge in the congestion edge set to the latest current number of evacuees;

[0113] S123: Let yd=yd+1. If yd is less than or equal to YD, continue executing S121 to S122. If yd is greater than YD, end the current process.

[0114] The calculation method of the evacuation flow rate shared by each evacuation network edge in the evacuation flow distribution set includes:

[0115]

[0116] Among them, Split(yd) i is the evacuation flow that the ith evacuation network edge should share in the evacuation diversion set corresponding to the ydth evacuation network edge, ED yd is the number of evacuation network edges in the evacuation diversion set corresponding to the yd-th evacuation network edge, NMAX(yd)i is the maximum number of people passing through the i-th evacuation network in the evacuation diversion set corresponding to the yd-th evacuation network edge, NNOW(yd) i YDRS is the current number of evacuees in the evacuation diversion set corresponding to the yd-th evacuation network edge, yd is the current number of evacuees on the ydth evacuation network edge, YDMAX yd is the maximum number of people passing through the yd-th evacuation network edge, and δ is the weight coefficient.

[0117] The calculation method of the latest current number of evacuees for each evacuation network edge in the evacuation diversion set includes:

[0118] NNOW(yd) (i，new) =Split(yd) i +NNOW(yd) i ;

[0119] Among them, NNOW(yd) (i，new) is the latest current number of evacuees on the ith evacuation network edge in the evacuation diversion set corresponding to the ydth evacuation network edge.

[0120] The calculation method of the latest current number of evacuees on the yd-th evacuation network edge in the congested edge set includes:

[0121]

[0122] Among them, YDCU (yd，new) YDCU is the latest current number of evacuees on the ydth evacuation network edge in the congested edge set, yd is the number of evacuees on the yd-th evacuation network edge in the congested edge set before the evacuees are diverted.

[0123] The method of analyzing and processing the evacuation network edges in the risk edge set according to the preset processing method to obtain the secondary congestion edge set includes:

[0124] S130: If the number of evacuation network edges in the risk edge set is not zero, the number of evacuation network edges in the risk edge set is recorded as FX, where the initial value of fx is preset to 1 and the value range of fx is 1 to FX; if the number of evacuation network edges in the risk edge set is zero, the current process ends;

[0125] S131: Obtain the fxth evacuation network edge from the risk edge set, input the current number of evacuees, individual attributes, maximum number of passers-by, and disaster characteristic parameters of the fxth evacuation network edge into the channel weight model, and obtain the evacuation network edge weight of the fxth evacuation network edge;

[0126] S132: If the evacuation network edge weight of the fxth evacuation network edge is greater than or equal to the congestion weight threshold 1, then add the fxth evacuation network edge to the secondary congested edge set;

[0127] S133: Let fx=fx+1. If fx is less than or equal to FX, continue executing S131 to S132; if fx is greater than FX, end the current process.

[0128] Methods for setting evacuation priorities for individual attributes of evacuees corresponding to evacuation network edges in a weight setting set include:

[0129] S140: Record the number of evacuation network edges in the weight setting set as QZ, preset the initial value of qz to be 1, and the value range of qz is 1 to QZ;

[0130] S141: Obtain the qzth evacuation network edge from the weight setting set, obtain the current number of evacuees in the qzth evacuation network edge, recorded as P; input the current number of evacuees, the maximum number of passers, and the individual attributes corresponding to the P evacuees into the weight setting model respectively, and obtain the evacuation priority corresponding to the P evacuees in the qzth evacuation network edge;

[0131] S142: Let qz = qz + 1. If qz is less than or equal to QZ, continue to execute S141. If qz is greater than QZ, end the current process.

[0132] The training method of the weight setting model includes:

[0133] A weight setting data set is collected in advance, wherein the weight setting data set includes Y groups of weight setting data and evacuation priorities corresponding to the Y groups of weight setting data, where Y is a positive integer greater than 0, and the weight setting data includes the current number of evacuees, individual attributes, and the maximum number of passers-by; the weight setting data set is divided into a training set and a test set, the weight setting data in the training set is used as the input of the weight setting model, and the evacuation priorities in the training set are used as the output of the weight setting model, with minimizing the sum of the prediction accuracies of all predicted evacuation priorities as the training goal; training is stopped when the sum of the prediction accuracies reaches convergence; the weight setting model is a random forest model or a gradient boosting tree model.

[0134] The evacuation assessment module is used to summarize and output the evacuation simulation results after the evacuation simulation is completed.

[0135] Methods for evaluating evacuation simulation results include:

[0136] During the evacuation simulation, dynamic recording data is obtained for each time step. The dynamic recording data includes the current position of the personnel, the personnel's movement speed and the channel status. The personnel who have completed the evacuation and those who have not completed the evacuation are classified and marked, and their final status is recorded; the evacuation time of the personnel who have not completed the evacuation is continued to be recorded, and the evacuation time of the personnel who have completed the evacuation is stopped; the evacuation distance is recorded in the same way as the evacuation time.

[0137] Count the number of people who have completed evacuation and the number of people who have not completed evacuation; classify the people who have completed evacuation and the people who have not completed evacuation according to their individual attributes to obtain the number of people who have completed evacuation and the number of people who have not completed evacuation corresponding to different attributes; summarize the evacuation time to obtain the longest evacuation time, shortest evacuation time and average evacuation time corresponding to people with different attributes; summarize the evacuation distance to obtain the longest evacuation distance, shortest evacuation distance and average evacuation distance corresponding to people with different attributes;

[0138] The number of people who have completed evacuation, the number of people who have not completed evacuation, the longest evacuation time, the shortest evacuation time, the average evacuation time, the longest evacuation distance, the shortest evacuation distance and the average evacuation distance corresponding to people with different attributes are output in a predetermined format.

[0139] Example 2

[0140] See also Figure 2 As shown, this embodiment provides an artificial intelligence-based intelligent evacuation path automatic generation method, which also includes:

[0141] Collecting building modeling data, wherein the building modeling data includes key building parameters and equipment resource parameters;

[0142] Modeling is performed according to the building modeling data to obtain a building simulation model;

[0143] Defining personnel characteristic parameters and disaster characteristic parameters on the building simulation model to obtain an intelligent evacuation simulation model;

[0144] Evacuation routes are adaptively generated based on the intelligent evacuation simulation model.

[0145] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0146] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The method for automatically generating an intelligent evacuation path based on artificial intelligence is characterized by: include: Collecting building modeling data, wherein the building modeling data includes key building parameters and equipment resource parameters; Modeling is performed according to the building modeling data to obtain a building simulation model; Defining personnel characteristic parameters and disaster characteristic parameters on the building simulation model to obtain an intelligent evacuation simulation model; Adaptively generate evacuation routes based on intelligent evacuation simulation models; The method of adaptively generating evacuation routes based on the intelligent evacuation simulation model includes: Step 1: Preset the total evacuation time and time step; take the quotient of the total evacuation time and the time step, and round up the result to obtain H time steps; the initial value of h is preset to 1, and the value range of h is 1 to H; add all evacuation network edges to the set of edges to be processed; the evacuation network edges are channels in the building simulation model; Step 2: Starting from the highest floor of the building structure, the evacuation status is judged in sequence. If all personnel on the current floor have been evacuated, the evacuation network edge connected to the evacuation network node corresponding to the current floor is removed from the set of pending edges. The judgment is continued downward layer by layer until a floor that has not yet been evacuated is encountered, and the judgment operation is terminated to obtain the set of pending edges of the hth time step. Step 3: Preset congestion weight threshold 1 and congestion weight threshold 2, where congestion weight threshold 1 is greater than congestion weight threshold 2; analyze and process the evacuation network edges in the edge set to be processed based on the congestion weight threshold 1 and the congestion weight threshold 2, and obtain the congested edge set, risk edge set, and smooth edge set at the hth time step; Step 4: Plan the evacuation path for the evacuation network edge based on the congestion edge set, risk edge set, and smooth edge set at the hth time step; Step 5: Set h = h + 1. If h is less than or equal to H, continue with steps 2 to 4. If h is greater than H, end the current process. Methods for planning evacuation paths based on the congestion edge set, risk edge set, and smooth edge set for the evacuation network edges include: Perform the first flow diversion on the evacuation network edge in the congested edge set according to the preset evacuation method; The evacuation network edges in the risk edge set are analyzed and processed according to the preset processing method to obtain the secondary congestion edge set. If the number of evacuation network edges in the secondary congestion edge set is greater than zero, the evacuation network edges in the secondary congestion edge set are subjected to a second flow diversion using the same method as the first flow diversion. If the number of evacuation network edges in the secondary congestion edge set is equal to zero, the current process ends. The same method as the second congestion edge set is used to obtain the third congestion edge set. If the number of evacuation network edges in the third congestion edge set is greater than zero, the third flow diversion is performed on the evacuation network edges in the third congestion edge set using the same method as the first flow diversion. If the number of evacuation network edges in the third congestion edge set is zero, the current process ends. The current number of evacuees, individual attributes, maximum number of passers-by, and disaster characteristic parameters corresponding to each evacuation network edge in the congestion edge set, risk edge set, and smooth edge set are input into the channel weight model respectively to obtain the evacuation network edge weight corresponding to each evacuation network edge in the congestion edge set, risk edge set, and smooth edge set; Evacuate personnel along the evacuation network edges according to the evacuation network edge weights.

2. The method for automatically generating an intelligent evacuation path based on artificial intelligence according to claim 1, characterized in that: The methods for evacuating personnel along the evacuation network edges according to the evacuation network edge weights include: If there is an evacuation network edge with a weight greater than or equal to the congestion weight threshold 1, the evacuation network edge with a weight greater than or equal to the congestion weight threshold 1 is added to the weight setting set, and the evacuation priority is set for the individual attribute of the evacuees corresponding to the evacuation network edge in the weight setting set; the evacuees with an evacuation priority of level 3 are evacuated from the evacuation network edge in the weight setting set, and the evacuees with level 1 and level 2 enter the floor buffer zone and wait. After the evacuees with level 3 are evacuated, the evacuees with level 1 and level 2 are evacuated; If there is no evacuation network edge whose evacuation network edge weight is greater than or equal to the congestion weight threshold 1, the current process ends.

3. The method for automatically generating an intelligent evacuation path based on artificial intelligence according to claim 2, characterized in that: Methods for setting evacuation priorities for individual attributes of evacuees corresponding to evacuation network edges in a weight setting set include: S140: Record the number of evacuation network edges in the weight setting set as QZ, preset the initial value of qz to be 1, and the value range of qz is 1 to QZ; S141: Obtain the qzth evacuation network edge from the weight setting set, obtain the current number of evacuees in the qzth evacuation network edge, recorded as P; input the current number of evacuees, the maximum number of passers, and the individual attributes corresponding to the P evacuees into the weight setting model respectively, and obtain the evacuation priority corresponding to the P evacuees in the qzth evacuation network edge; S142: Let qz = qz + 1. If qz is less than or equal to QZ, continue to execute S141. If qz is greater than QZ, end the current process.

4. The method for automatically generating an intelligent evacuation path based on artificial intelligence according to claim 3, characterized in that: The method for obtaining the congested edge set, the risk edge set, and the smooth edge set includes: S110: Record the number of evacuation network edges in the edge set to be processed as TD, preset the initial value of td to be 1, and the value range of td is 1 to TD; S111: Get the current number of evacuees and individual attributes in the tdth evacuation network edge; S112: Input the current number of evacuees, individual attributes, maximum number of passers and disaster characteristic parameters into the channel weight model to obtain the evacuation network edge weight of the td-th evacuation network edge; S113: If the evacuation network edge weight of the tdth evacuation network edge is greater than or equal to the congestion weight threshold 1, then the tdth evacuation network edge is added to the congested edge set; if the evacuation network edge weight of the tdth evacuation network edge is greater than or equal to the congestion weight threshold 2 and less than the congestion weight threshold 1, then the tdth evacuation network edge is added to the risk edge set; if the evacuation network edge weight of the tdth evacuation network edge is less than the congestion weight threshold 2, then the tdth evacuation network edge is added to the unobstructed edge set; S114: Let td=td+1. If td is less than or equal to TD, continue executing S111 to S113; if td is greater than TD, end the current process.

5. The method for automatically generating an intelligent evacuation path based on artificial intelligence according to claim 4, characterized in that: Methods for performing the first flow diversion on the evacuation network edges in the congested edge set according to the preset evacuation method include: S120: If the number of evacuation network edges in the congested edge set is not zero, the number of evacuation network edges in the congested edge set is recorded as YD, where the initial value of yd is preset to 1 and the value range of yd is 1 to YD; if the number of evacuation network edges in the congested edge set is zero, the current process ends; S121: Obtain the ydth evacuation network edge from the congestion edge set, and add the evacuation network edges that are adjacent to the ydth evacuation network edge and are not in the congestion edge set to the evacuation diversion set corresponding to the ydth evacuation network edge; if the number of evacuation network edges in the evacuation diversion set corresponding to the ydth evacuation network edge is not zero, calculate the evacuation flow rate shared by each evacuation network edge in the evacuation diversion set, and calculate the latest current number of evacuees for each evacuation network edge in the evacuation diversion set based on the evacuation flow rate shared by each evacuation network edge; update the current number of evacuees for each evacuation network edge in the evacuation diversion set to the corresponding latest current number of evacuees; S122: Calculate the latest current number of evacuees on the yd-th evacuation network edge in the congestion edge set based on the evacuation flow shared by each evacuation network edge in the evacuation diversion set, and update the current number of evacuees on the yd-th evacuation network edge in the congestion edge set to the latest current number of evacuees; S123: Let yd=yd+1. If yd is less than or equal to YD, continue executing S121 to S122. If yd is greater than YD, end the current process.

6. The method for automatically generating an intelligent evacuation path based on artificial intelligence according to claim 5, characterized in that: The calculation method of the evacuation flow rate shared by each evacuation network edge in the evacuation flow distribution set includes: The evacuation flow rate shared by each evacuation network edge is calculated proportionally according to the residual capacity of each evacuation network edge; the residual capacity of each evacuation network edge is the maximum number of passers-by corresponding to each evacuation network edge minus the current number of evacuees.

7. The method for automatically generating an intelligent evacuation path based on artificial intelligence according to claim 6, characterized in that: The method of analyzing and processing the evacuation network edges in the risk edge set according to the preset processing method to obtain the secondary congestion edge set includes: S130: If the number of evacuation network edges in the risk edge set is not zero, the number of evacuation network edges in the risk edge set is recorded as FX, where the initial value of fx is preset to 1 and the value range of fx is 1 to FX; if the number of evacuation network edges in the risk edge set is zero, the current process ends; S131: Obtain the fxth evacuation network edge from the risk edge set, input the current number of evacuees, individual attributes, maximum number of passers-by, and disaster characteristic parameters of the fxth evacuation network edge into the channel weight model, and obtain the evacuation network edge weight of the fxth evacuation network edge; S132: If the evacuation network edge weight of the fxth evacuation network edge is greater than or equal to the congestion weight threshold 1, then add the fxth evacuation network edge to the secondary congested edge set; S133: Let fx=fx+1. If fx is less than or equal to FX, continue executing S131 to S132; if fx is greater than FX, end the current process.

8. An artificial intelligence-based intelligent evacuation path automatic generation system, implementing the artificial intelligence-based intelligent evacuation path automatic generation method according to any one of claims 1 to 7, characterized in that: include: A first acquisition module is used to acquire building modeling data, wherein the building modeling data includes key building parameters and equipment resource parameters; A data modeling module is used to perform modeling based on building modeling data to obtain a building simulation model; A parameter definition module is used to define personnel characteristic parameters and disaster characteristic parameters on the building simulation model to obtain an intelligent evacuation simulation model; Intelligent evacuation module, used to adaptively generate evacuation routes based on the intelligent evacuation simulation model; The evacuation assessment module is used to summarize and output the evacuation simulation results after the evacuation simulation is completed.

Images