Airport terminal passenger emergency evacuation path optimization method based on improved A* algorithm

Abstract

The invention discloses a terminal fire evacuation path optimization method based on an improved A * algorithm, and the method comprises the following steps: (1) analyzing the personnel flow of a terminal, arranging real flight data, and carrying out the preprocessing; (2) simulating an airport terminal personnel flow by using the simulation model, and arranging personnel distribution data; (3) analyzing the reasons of the airport terminal fire, simulating the airport terminal fire by using fire simulation according to the reasons of the high-incidence fire, and obtaining the data of the airport terminal fire products; (4) improving a path planning algorithm, coupling personnel distribution and fire spreading data, and re-planning a congestion path; according to the invention, a new solution thought is provided for terminal fire emergency evacuation, and scientific practical guidance is provided for emergency evacuation in complex environments such as terminals and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of civil aviation safety, and particularly relates to an optimization method for the emergency evacuation path of terminal passengers that couples a personnel flow model and a fire spread model. Background Art

[0002] With the rapid development of China's civil aviation industry, airports, as the core infrastructure of air transportation, have continuously expanded in scale and function. In particular, the construction and operation of some large airports have not only played an important role in meeting the growing demand for air transportation but also had a profound impact on regional economic development and social benefits. With the annual increase in the volume of civil aviation passenger transportation, the possibility of safety accidents is also increasing. Once effective measures cannot be taken to handle accidents, serious casualties and property losses will be caused. The airport terminal, as a gathering place for inbound, outbound, and transfer passengers, has a dense passenger population, and the difficulty of evacuation will increase significantly after an emergency occurs.

[0003] Firstly, there are significant differences in the building space and structure of the terminal, and there is a lack of a unified fire risk standard. Secondly, the combustion products of various combustibles vary greatly in different environments, making it difficult to conduct accurate quantitative analysis. In addition, most of the existing fire simulations focus on single-story buildings or single fire compartments and do not fully consider the influence of multi-story buildings and complex building structures. The research on fires in the terminal's traffic connection spaces is even scarcer. Fires are more likely to occur in the terminal, with a wide range of influence and high control difficulty. After a fire occurs, passengers need to be quickly evacuated. Due to the complex structure of the terminal, passengers are unfamiliar with the environment and cannot understand the fire situation in real time. Blind evacuation is likely to lead to congestion or entry into dangerous areas. Most current studies assume that passengers are randomly distributed in the terminal, ignoring the characteristics of passenger distribution changing over time. Summary of the Invention

[0004] Object of the Invention: In case of emergencies such as fires, it is necessary to plan scientific and effective evacuation paths in combination with the actual situation to ensure the efficiency and safety of the evacuation process. To solve the problem of inconsistent evacuation paths caused by differences in passenger distribution in the terminal and to avoid congestion as much as possible and speed up the evacuation speed, this paper proposes a path optimization method that couples the fire impact and the time-varying characteristics of passenger distribution in the terminal. This method takes into account the fire impact and the time-varying characteristics of passenger distribution in the terminal. By simulating the actual passenger behavior in the terminal and optimizing the path according to the real-time distribution data, the evacuation efficiency is improved.

[0005] Technical Solution: An optimization method for the emergency evacuation path of terminal passengers that couples a personnel flow model and a fire spread model according to the present invention includes the following steps:

[0006] (1) Combine flight data and the terminal structure to set up the topological graph of the terminal path network;

[0007] (2) Conduct an in-depth analysis of the passenger behavior process in the terminal building to determine the simulation input data; and respectively conduct simulation dynamic simulations on the departure and arrival processes of passengers to obtain the personnel distribution data;

[0008] (3) Through fire simulations of the terminal building under different scenarios, analyze the fire products during the fire spread process, set measuring point devices for the passenger evacuation paths, and obtain the fire product data;

[0009] (4) Determine the passing state of the nodes according to the impact of the fire products on the human body in the fire product data;

[0010] (5) Combine the path network topology map of the terminal building, improve the path planning algorithm, couple the fire impact and personnel distribution, and optimize the evacuation paths; dynamically evaluate the impact of the fire on different paths, consider the density of personnel at the same time, and re-plan the congested paths.

[0011] Further, in the step (2), the simulation input data includes: the terminal building structure, flight data, passenger arrival rules, and passenger body data; the simulation dynamic simulation is to simulate the passenger behavior process in the terminal building in the simulation software to obtain the personnel distribution data, and generate the terminal building personnel distribution matrix in combination with the topology map of the terminal building structure.

[0012] Further, the expression of the personnel distribution matrix D is as follows:

[0013]

[0014] Among them, d nm represents the personnel density on the path between node n and node m.

[0015] Further, in the step (3), the fire products include temperature, CO concentration, and smoke visibility.

[0016] Further, in the step (4), nodes are set at the intersections, entrances and exits, and boarding gates on the departure level of the terminal building, and the formula for judging the passing state of the nodes is as follows:

[0017] Q = min(Q temperature , Q CO , Q Visibility )

[0018] Q temerature = T temperature≥60

[0019] Q CO = T CO≥500ppm

[0020] QVisibility = T Visibility≤10

[0021] Wherein, Q is the time when the node becomes impassable, that is, the state transition time data; Q temerature is T temperature The time data when the node temperature reaches the dangerous critical value of more than 60°C; Q CO is T CO The time data when the CO concentration of the node reaches the dangerous critical value of more than 500 ppm; Q Visibility is T Visibility The time data when the visibility data of the node drops to the dangerous critical value of less than 10 m; when the node has not reached the dangerous critical value within the simulation time, the time data is recorded as infinity inf;

[0022] Combining the data obtained from the simulation of each node and the state transition time data Q of each node calculated by the above formula, by comparing the arrival time at the node with the state transition time of the node, it is judged whether the path is passable, and it is substituted into the cost function of the path planning algorithm for calculation; the state transition time data Q of each node is recorded as

[0023] Q = (Q 1 , Q 2 , Q 3 , … Q n )

[0024] Among them, Q i represents the state transition time data of the i-th node, i ∈ 1, 2... n, and n is the total number of nodes;

[0025] During the path planning process, by comparing the arrival time of the node with its state transition time, it is dynamically judged whether the path is passable:

[0026] If the arrival time of the node is earlier than its state transition time, that is, T arrival < Q i , then the node is still in a passable state; if the arrival time of the node is later than its state transition time, that is, T arrival ≥ Q i , then the node is no longer passable due to the influence of fire products.

[0027] Furthermore, in the step (5), the path planning algorithm is improved, including the improvement of the traversal method and the improvement of the path cost;

[0028] The improvement of the traversal method includes, combining the path network topology diagram of the terminal building, using the network flow model to traverse through adjacent nodes, determining the passability of each node according to the network topology diagram, and establishing an accessibility matrix A;

[0029] The adjacent nodes are represented as follows:

[0030]

[0031] Wherein, P min is the node with the minimum value of the cost function fn in this round of traversal and will be selected as the next node in this round of traversal; is the cost function value of node P i to the nth adjacent node;

[0032] The improvement of the path cost function includes improving the weight of the edge by combining the calculation formulas of density and personnel movement speed. When the personnel density is between 0.54 person / m 2 -3.8 persons / m 2 the relationship between the movement speed of pedestrians and the personnel density is as follows

[0033] v ij = v 0 - av o D

[0034] Wherein, v ij is the walking speed on edge ij during evacuation, m / s; D is the personnel density on the current edge, persons / m 2 ; v 0 is the initial evacuation speed, taken as 1.4 m / s; a is a coefficient, taken as 0.266;

[0035] Calculate the personnel passage speed matrix of each path

[0036]

[0037] Combine with the distance weight matrix of the path

[0038]

[0039] Generate a new time weight matrix for the path

[0040]

[0041] The A* algorithm tracks the search status of each node in the network topology graph through the set of nodes to be searched and the set of nodes that have been searched, and determines the next search node through the total cost function f(n). The function expression is:

[0042] f(n) = g(n) + h(n)

[0043] g(n) is the distance from the current node to the adjacent node of this node; when selecting the next node, it is necessary to judge whether the status of the next node is passable; judge the status of the node based on the measured data of fire products, so as to calculate the data Q of the exceeding standard of fire products and the state transition time of all nodes i; Determine whether the time for the fire to reach a node exceeds the time for a person to reach that node, using the following formula:

[0044] T nj = T ni + t ij

[0045]

[0046] In the formula, T nj is the total time required from node n to node j; T ni is the total time required from node n to node i; t ij is the total time required from node i to node j; Q i is the time from the start of timing at the fire source node until the state of node i changes from safe to impassable, that is, the node state transition time data; g(n) is the time from the current node to the adjacent node. If this time is greater than the remaining time for the fire to reach the adjacent node, the weight of this edge is replaced by ∞, otherwise the time required for a pedestrian to pass this path is used; h(n) is the distance from the adjacent node of this node to the target node. The setting of the heuristic function determines the efficiency and accuracy of path planning;

[0047] h(n)= dis[P]

[0048] In the formula, dis[P] is the shortest distance from the current node to the target node obtained by the Dijkstra algorithm, where the weight of the edge is time, calculated by the time weight matrix W time above;

[0049] Since the magnitudes of h(n) and g(n) will affect the search effect of the A* algorithm, in order to unify the weights and improve the algorithm efficiency, weight coefficients α and β are set for h(n) and g(n);

[0050] In summary, after improving the traditional A* algorithm, the cost function is:

[0051] f(n)= αg(n)+ βh(n)= αt ij + βdis[P]

[0052] Adjust the parameters, and confirm the values of parameters α and β according to the path optimization results.

[0053] Further, in step (5), the optimization of the evacuation path is specifically as follows: the personnel density on the path is monitored in real time and compared with a preset density threshold; when it is found that the personnel density of a certain path or node exceeds the density threshold, the path will be marked as a congested area and excluded from the original path; then, based on the remaining available paths, a new path planning is carried out to ensure that personnel can avoid the congested area and choose a safer and smoother evacuation route; during this process, the new path planning strategy will also be adjusted according to the personnel density and path length of each path segment to shorten the evacuation time and improve the evacuation efficiency.

[0054] Beneficial effects:

[0055] 1. The present invention designs fire scenarios for each floor of the terminal building in combination with different fire causes, conducts a detailed simulation of the terminal building under different fire scenarios, analyzes the dynamic characteristics of fire spread and its impact on the safety of each node in the terminal building. By simulating the spread range and the impact of fire products in different fire environments, the key data of the state transformation of nodes during the fire spread process are obtained, providing data support for subsequent path planning. Based on these simulation results, the present invention clarifies the time required for each node to transform into a dangerous node, providing an accurate basis for fire risk assessment.

[0056] 2. The present invention combines actual flight data and the actual structure of the terminal building, simulates the behavior process of personnel in the terminal building, and obtains the personnel distribution data at different time nodes and flight conditions. These data reflect the distribution and movement rules of personnel in the terminal building, providing important inputs for path planning.

[0057] 3. On this basis, the present invention proposes an improved path planning algorithm that can dynamically adjust the evacuation path according to the real-time fire spread situation and personnel distribution. The core idea of this algorithm is to optimize the safety and efficiency of the evacuation path by dynamically evaluating the impact of fire on different paths and combining the personnel density. By incorporating the time-varying characteristics of fire spread and personnel distribution into the path planning process, the path planned after comprehensively considering the fire impact and personnel distribution can effectively reduce the calculation time compared with the path planned by the traditional algorithm, improving the real-time performance and applicability of the algorithm.

[0058] 4. In summary, the present invention conducts in-depth research on the fire emergency evacuation path planning of the terminal building from multiple aspects, and the proposed path optimization method based on the dynamic changes of fire spread and personnel distribution provides a scientific solution for the fire emergency management of the terminal building. The research results not only have strong theoretical significance but also have high practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a schematic diagram of the behavior process of terminal passengers in the embodiment of the present invention.

[0060] Figure 2 This is a schematic diagram of the personnel distribution in the terminal building at different times in the embodiment of the present invention. Among them, (a) is the schematic diagram of the personnel distribution in the terminal building at 150 s, (b) is the schematic diagram of the personnel distribution in the terminal building at 245 s, and (c) is the schematic diagram of the personnel distribution in the terminal building at 450 s; the digital serial numbers are the boarding gate numbers.

[0061] Figure 3 This is a schematic diagram of the relevant data of the fire products in Scenario 1 of the embodiment of the present invention.

[0062] Figure 4 This is a schematic diagram of the corresponding numbers and positions of the measuring points in Scenario 1 of the embodiment of the present invention.

[0063] Figure 5 This is a schematic diagram of the path replanning in the embodiment of the present invention.

[0064] Figure 6 This is a comparison chart of the path planning times of the traditional algorithm and the improved A* algorithm in the fire Scenario 1 and the personnel distribution Scenario 1 in the embodiment of the present invention. Detailed implementation manners

[0065] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0066] As shown in the figure, the embodiment of the present invention provides a path optimization method that couples a personnel flow model and a fire spread model, including the following steps:

[0067] (1) Combine flight data and the specific structure of the terminal building to set up the path network topology diagram of the terminal building;

[0068] (2) In this embodiment, the passenger behavior process in the terminal building is as Figure 1 shown. Through in-depth analysis of the passenger behavior process in the terminal building, determine the simulation input data; use simulation software to dynamically simulate the departure and arrival processes of passengers respectively to obtain personnel distribution data; the relevant data includes: the terminal building structure, flight data, passenger arrival rules, passenger body data, etc.; the dynamic simulation is to set the personnel behavior process in the simulation software, input the relevant data, and obtain the personnel distribution data.

[0069] Generate a personnel distribution matrix of the terminal building in combination with the terminal building structure topology diagram, and the expression method is as follows:

[0070]

[0071] where d nm represents the personnel density on the path between node n and node m.

[0072] In this embodiment, simulations are carried out at different times within the hour with the largest number of flights. After the personnel arrive with the flights, they move within the terminal building according to the arrival passenger process, and there are obvious gatherings in the atrium and the baggage claim area. The personnel distribution at three different times in the same area within one hour is as Figure 2 shown. There are relatively obvious differences in the areas where personnel gather at different times, and the differences are most obvious in the atrium and the positions shown in the figure. There are significant differences in the personnel density on different alternative paths on the same necessary section of the road, which will thus cause differences in the personnel passage time and further affect the subsequent path selection.

[0073] (3) By conducting fire simulations on the terminal building under different scenarios, the fire products during the fire spread process are analyzed, a measuring point device is set up to obtain relevant data on fire products; the fire product data includes temperature, CO concentration, smoke visibility, etc.

[0074] In this embodiment, the fire occurrence scenarios are divided into three scenarios: a fire occurs in the machine room on the arrival level of the second floor of the terminal building due to aging of the circuit, a fire occurs to the passenger luggage on the fourth floor of the terminal building, and a fire occurs in the shops on the third floor of the terminal building. At the same time, to consider the most unfavorable principle of the fire, the fire source is set at a position relatively close to where the personnel gather. It is experimentally assumed that the doors of each room and evacuation passage are in the open state. For Scenario 1, referring to the "Technical Standard for Building Smoke Control and Exhaust Systems", the fire source power is selected as 10 MW, and the fire source type is fast fire. Considering the different internal structures and personnel compositions of the terminal building, different fires at different positions and of different types will have different impacts on the interior of the terminal building. The following settings are made for the fire-related environmental parameters: initial environmental temperature: 24 °C; initial environmental relative humidity: 50%; fire simulation running time: 600 s; combustible: PVC; fire protection conditions: both the automatic sprinkler system and the mechanical smoke exhaust facilities fail. The schematic diagram of the relevant data of the fire products in Scenario 1 is as Figure 3 shown.

[0075] (4) In this embodiment, based on the node selection method of graph theory, nodes are set at the intersections, entrances and exits, and boarding gate positions on the departure level of the airport terminal building. Measuring points are set for the passenger evacuation paths to monitor the fire products at each node and judge the node status. The measuring point settings on the arrival level in Scenario 1 are as Figure 4 shown.

[0076] Judge the node passage state according to the impact of the fire products on the human body; judge the node passage state, and the formula is as follows:

[0077] Q = min(Q temperature , Q CO , Q Visibility )

[0078] Q temerature = T temperature≥60

[0079] Q CO = T CO≥500ppm

[0080] Q Visibility = T Visibility≤10

[0081] Wherein, Q is the time data when the node becomes non - passable, that is, the state transition time data of the node; Q temerature is the time data when the node temperature reaches above the dangerous critical value of 60 °C; Q CO is the time data when the CO concentration of the node reaches above the dangerous critical value of 500 ppm; Q Visibility is the time data when the visibility data of the node drops to below the dangerous critical value of 10 m; When the node has not reached the dangerous critical value within the simulation time, the time data is recorded as infinity inf.

[0082] Combining the data obtained from the simulation of each node, and the state transition time vector Q of each node calculated by the above formula, by comparing the arrival time at the node with the state transition time of the node, it is judged whether the path is passable, and then substituted into the cost function of the path planning algorithm for calculation.

[0083] The state transition time vector Q of each node is denoted as

[0084] Q=(Q 1 , Q 2 , Q 3 , …Q n )

[0085] Wherein, Q i represents the state transition time data of the i - th node, and n is the total number of nodes.

[0086] During the path planning process, by comparing the arrival time of the node with its state transition time, it can be dynamically judged whether the path is passable:

[0087] If the arrival time of the node is earlier than its state transition time (i.e., T arrival < Q i ), then the node is still in a passable state; If the arrival time of the node is later than its state transition time (i.e., T arrival ≥ Q i ), then the node is no longer passable due to the influence of fire products.

[0088] In this embodiment, the non - passable time of the affected nodes in Scenario 1 is shown in Table 1

[0089] Table 1 Non - passable time data table of measuring points in Scenario 1

[0090] Node Time to turn into a dangerous node / s Node Time to turn into a dangerous node / s 7 175 11 282 8 201 12 376 9 179 5 inf 10 inf 6 inf 17 345 18 280 25 159 26 156 27 261

[0091] (5) Improve the path planning algorithm, including improving the traversal method and path cost.

[0092] The improvement of the traversal method includes combining the network topology graph, using the network flow model to traverse through adjacent nodes, determining the traffic conditions between nodes according to the network topology graph, and establishing the reachability matrix A.

[0093] The adjacent nodes are represented as follows:

[0094]

[0095] In the formula, P min is the node with the minimum value of the cost function fn in the current traversal, and will be selected as the next node in the current traversal process. is the node P i to the cost function value of the nth adjacent node.

[0096] The improvement of the cost function includes that the A* algorithm tracks the search status of each node in the evacuation network through the set of nodes to be searched and the set of nodes that have been searched, and determines the next search node through the total cost function f(n). The function expression is:

[0097] f(n) = g(n) + h(n)

[0098] g(n) is the distance from the current node to the adjacent node of this node. When selecting the next node, it is necessary to judge whether the status of the next node is passable. Judge the status of the node based on the measured data of three types of fire products, so as to calculate the time data Q when all nodes exceed the standard of fire products and the node becomes an impassable state. i . To judge whether the time when the fire reaches the node exceeds the time when the personnel reach the node, the following formula can be used:

[0099] T nj = T ni + t ij

[0100]

[0101] In the formula, T nj is the total time required for node n to reach node j; Q i is the time from the start of timing at the fire source node until the state of node i changes from safe to impassable, that is, the node state transition time data. g(n) is the time from the current node to the adjacent node. If this time is greater than the remaining time for the fire to reach the adjacent node, the weight of this edge is replaced by ∞, otherwise the time required for pedestrians to pass this path is used. h(n) is the distance from the adjacent node of this node to the target node. The setting of the heuristic function determines the efficiency and accuracy of path planning.

[0102] h(n) = dis[P]

[0103] where: dis[P] is the shortest distance from the current node to the target node obtained by the Dijkstra algorithm, where the weight of the edge is time, calculated from the time weight matrix W above time Calculate.

[0104] Since the magnitudes of h(n) and g(n) will affect the search effect of the A* algorithm, in order to unify the weights and improve the algorithm efficiency, weight coefficients α and β are set for h(n) and g(n).

[0105] In summary, after improving the traditional A* algorithm, the cost function is:[[]]

[0106] f(n) = αg(n) + βh(n) = αt ij + βdis[P]

[0107] Adjust the parameters to determine that α and β are taken as 1 and 0.7 respectively.

[0108] Improve the path planning algorithm, couple the fire impact and the personnel distribution situation, and optimize the evacuation path. Dynamically evaluate the impact of the fire on different paths, and at the same time consider the density of personnel to optimize the safety and efficiency of the evacuation path. Specifically: the algorithm will monitor the personnel density on the path in real time and compare it with the preset density threshold. When it is found that the personnel density of a certain path or node exceeds the critical value, the system will mark this path as a congested area and remove it from the original path. Next, based on the remaining available paths, the algorithm will re-plan the path to ensure that personnel can avoid the congested area and choose a safer and smoother evacuation route. In this embodiment, the schematic diagram of path replanning is as Figure 5 shown.

[0109] During this process, the algorithm will also adjust the new path planning strategy according to the personnel density and path length of each path segment to shorten the evacuation time as much as possible and improve the evacuation efficiency. In fire scenario one of this embodiment, the comparison chart of the path planning time of the traditional algorithm and the improved A* algorithm under the personnel distribution scenario is as Figure 6 shown.

Claims

1. A method for optimizing the emergency evacuation path of passengers in a terminal building based on an improved A* algorithm, characterized in that: The following steps are involved: (1) Combine flight data and terminal structure to set up the terminal path network topology map; (2) Conduct in-depth analysis of passenger behavior processes in the terminal to determine simulation input data; The departure and arrival processes of passengers are simulated dynamically to obtain personnel distribution data; (3) Fire simulations were conducted in the terminal under different scenarios to analyze the fire products during the fire spread process, and measurement points were set up along the passenger evacuation routes to obtain fire product data. (4) Determine the node traffic status based on the impact of fire products on the human body in the fire product data; (5) Combined with the terminal path network topology, the path planning algorithm is improved, the fire impact and personnel distribution are coupled, and the evacuation path is optimized; the impact of the fire on different paths is dynamically evaluated, while considering the density of personnel, and congested paths are replanned.

2. The terminal passenger emergency evacuation path optimization method based on the improved A* algorithm according to claim 1 is characterized in that: In the step (2), the simulation input data includes: terminal structure, flight data, passenger arrival patterns and passenger physical data; the dynamic simulation is to simulate the passenger behavior process of the terminal in the simulation software, obtain personnel distribution data, and generate a terminal personnel distribution matrix in combination with the topological diagram of the terminal structure.

3. The terminal passenger emergency evacuation path optimization method based on the improved A* algorithm according to claim 2 is characterized in that: The personnel distribution matrix D is expressed as follows: Among them, d nm Represents the density of people on the path between node n and node m.

4. The terminal passenger emergency evacuation path optimization method based on the improved A* algorithm according to claim 1 is characterized in that: In the step (3), the fire products include temperature, CO concentration and smoke visibility.

5. The terminal passenger emergency evacuation path optimization method based on the improved A* algorithm according to claim 1 is characterized in that: In the step (4), nodes are set at each intersection, entrance and exit, and boarding gate of the departure level of the terminal building, and the formula for judging the traffic status of the node is as follows: Q=min(Q temperature ,Q CO ,Q Visibility ) Q temerature =T temperature≥60 Q CO =T CO≥500ppm Q Visibility =T Visibility≤10 Where Q is the time when the node becomes inaccessible, that is, the state transition time data; Q temerature T temperature The time data when the node temperature reaches the dangerous critical value above 60℃; Q CO T CO The time data when the node CO concentration reaches the dangerous critical value of 500ppm or above; Q Visibility T Visibility The time data when the visibility data of the node drops below the critical value of 10m; if the node does not reach the critical value within the simulation time, the time data is recorded as infinity inf; The data obtained by simulating each node is combined with the state transition time data Q of each node calculated by the above formula. By comparing the arrival time of the node with the node state transition time, it is determined whether the path is passable and substituted into the cost function of the path planning algorithm for calculation; the state transition time data Q of each node is recorded as Q=(Q1,Q2,Q3…Q n ) Among them, Q i Represents the state transition time data of the i-th node, i∈1,2…n, n is the total number of nodes; During the path planning process, the path is dynamically determined to be passable by comparing the arrival time of the node with its state transition time: If the arrival time of a node is earlier than its state transition time, that is, T arrival <Q i , then the node is still in a passable state; if the arrival time of the node is later than its state transition time, that is, T arrival ≥Q i , then the node is no longer accessible due to the influence of fire products.

6. The terminal passenger emergency evacuation path optimization method based on the improved A* algorithm according to claim 1 is characterized in that: In the step (5), the path planning algorithm is improved, including the traversal method improvement and the path cost improvement; Improvements to the traversal method include: combining the terminal path network topology map, using the network flow model to traverse through adjacent nodes, determining the traffic conditions of each node according to the network topology map, and establishing a reachable matrix A; Adjacent nodes are represented as follows: Where P min The node with the smallest cost function fn value in this round of traversal will be selected as the next node in this round of traversal; For node P i The cost function value to the nth adjacent node; The improvement of the path cost function includes improving the edge weight by combining the calculation formula of density and personnel movement speed. When the personnel density is 0.54 people / m 2 -3.8 people / m 2 When the pedestrian movement speed and personnel density are v ij =v0-off o D In the formula, v ij is the walking speed on the edge ij during evacuation, m / s; D is the personnel density on the current edge, person / m 2 ; v0 is the initial evacuation speed, which is 1.4 m / s; a is the coefficient, which is 0.266; Calculate the speed matrix of people passing through each path Distance weight matrix of combined paths Generate a new time weight matrix for the path The A* algorithm tracks the search status of each node in the network topology through the set of nodes to be searched and the set of nodes that have been searched, and determines the next search node through the total cost function f(n). The function expression is: f(n)=g(n)+h(n) g(n) is the distance from the current node to the adjacent node of the node; when selecting the next node, it is necessary to determine whether the next node state is passable; the state of the node is determined based on the fire product measurement point data, so as to calculate the fire product exceeding the standard of all nodes, and the node state transition time data Q i ; To determine whether the time it takes for the fire to arrive at a node exceeds the time it takes for personnel to arrive at the node, use the following formula: T nj =T ni +t ij Where, T nj is the total time required for node n to reach node j; T ni is the total time required for node n to reach node i; t ij is the total time required for node i to reach node j; Q i is the time from the fire source node until the state of node i changes from safe to inaccessible, that is, the node state transition time data; g(n) is the time from the current node to the adjacent node. If this time is greater than the remaining time for the fire to reach the adjacent node, the weight of the edge is replaced by ∞, otherwise the time required for pedestrians to pass the path is used; h(n) is the distance from the adjacent node of the node to the target node. The setting of the heuristic function determines the efficiency and accuracy of path planning; h(n)=dis[P] Where dis[P] is the shortest distance from the current node to the target node obtained by the Dijkstra algorithm, where the edge weight is time, which is obtained from the time weight matrix W above. time calculate; Since the size of h(n) and g(n) will affect the search effect of the A* algorithm, in order to unify the weights and improve the efficiency of the algorithm, the weight coefficients α and β are set for h(n) and g(n); In summary, after improving the traditional A* algorithm, the cost function is: f(n)=αg(n)+βh(n)=αt ij +βdis[P] Adjust the parameters and confirm the values ​​of parameters α and β according to the path optimization results.

7. The terminal passenger emergency evacuation path optimization method based on the improved A* algorithm according to claim 1 is characterized in that: In the step (5), the optimization of the evacuation path is specifically as follows: real-time monitoring of the density of people on the path and comparing it with a preset density threshold; when it is found that the density of people on a certain path or node exceeds the density threshold, the path is marked as a congested area and removed from the original path; then, based on the remaining available paths, the path is replanned to ensure that people can avoid the congested area and choose a safer and smoother evacuation channel; During this process, new path planning strategies will also be adjusted according to the population density and path length of each path segment to shorten the evacuation time and improve the evacuation efficiency.

Images