Real-time path planning method and system for fire scene evacuation

Abstract

The invention discloses a real-time path planning method and system for fire scene evacuation. The method comprises the following steps: initializing a to-be-explored list FS, an explored list ES and a set P; judging whether the to-be-explored list FS is empty or not, and if yes, backtracking to find a father node and outputting a path; if not, setting the grid with the minimum cost as the current grid Gi, and adding the neighbor grid Gj into the to-be-explored list FS; judging whether the current grid Gj is equal to the target network Gt or not, and if so, backtracking to find a father node and outputting a path; otherwise, checking a neighbor grid Gj of the grid Gi; whether the neighbor grid Gj is not in the explored list ES or not is judged, if yes, the neighbor grid Gj is added into the to-be-searched list FS, and returning is continued to judge whether the to-be-searched list FS is empty or not; otherwise, neglecting the grid Gj, and continuously returning to judge whether the to-be-explored list FS is empty or not. According to the invention, more accurate, real-time and globally considered evacuation path planning can be provided so as to ensure safe and rapid evacuation of personnel in a fire scene.

Description

Technical Field

[0001] The present invention belongs to the technical field of path planning, and in particular relates to a real-time path planning method and system for fire scene evacuation. Background Art

[0002] In recent years, with the rapid development of urban modernization, the importance of fire prevention has gradually attracted people's attention. Unfortunately, when a fire occurs, people often miss the golden time for escape because they do not know the escape route of the entire environment. In addition, the collapse or congestion of the escape intersection may cause casualties of escapees or the inability of rescuers to provide real-time rescue work. However, in order to respond to the occurrence of fires, various sensors are arranged in buildings, public places and residential areas to monitor the temperature, smoke, gas concentration and other physical information in the environment in real time. Then, through the technology of the Internet of Things, sensors and mobile devices are connected to the Internet to establish an intelligent fire monitoring and early warning system. Once signs of fire are detected, the system will immediately issue an alarm and transmit relevant information to relevant departments and residents so that they can quickly take measures to escape. People can get real-time information and guidance about fires through mobile phone applications, such as fire escape guides, fire emergency contacts, etc. These mobile devices can also be used as emergency communication tools. People can keep in touch with rescuers and relatives and friends through voice calls or social media platforms to seek help or provide their location information.

[0003] In the prior art, a dynamic path optimization method based on an improved A* algorithm is proposed for the sudden fire of a cruise ship to obtain the optimal evacuation path in real time. According to the dynamic diffusion data of the fire, the evacuation path is planned using the improved A* algorithm and the breadth-first search strategy to determine the optimal path from the current position to the safe exit and reduce the possibility of casualties during the evacuation process. A new fire detection and path planning algorithm is disclosed in the prior art, which uses a target detection algorithm based on a convolutional neural network and is trained using real fire images to detect the fire area in the image. A path planning method for an improved A* algorithm is proposed, which sets weights for different areas and uses a box fuzzy method to make the escape path trajectory stay away from obstacles. The performance analysis of the A* algorithm in the Indonesian Fire Robot Competition (KRPAI) is discussed. The author uses time complexity and space complexity as the evaluation criteria. According to the experimental results, the A* algorithm requires an average of 4,270.72 time units and 14,192 bytes of memory to find a solution. In addition, there are path planning algorithms designed based on RRT and greedy algorithms. RRT does not consider the development trend of the fire, so it will have a certain probability of exploring the target point, but it will also have a certain probability of exploring in any direction. The advantage is that it is simple to implement, but the disadvantage is that it takes too much time to plan the path, which is not enough for application environments that require high real-time performance. The greedy method is close to the planning thinking of the escapee, and the selected path is planned according to the principle of local optimality. In an environment with many escapees, it may cause congestion at the escape exit, because the escapee only focuses on the path that is most beneficial to him at the moment to plan. Summary of the invention

[0004] To solve the above technical problems, the present invention proposes a real-time path planning method and system for fire evacuation, which can provide more accurate, real-time and global evacuation path planning to ensure the safe and rapid evacuation of personnel in fire scenes.

[0005] On the one hand, to achieve the above-mentioned purpose, the present invention provides a real-time path planning method for fire evacuation, comprising the following steps:

[0006] Initialize the to-be-explored list FS, the explored list ES and the set P;

[0007] Determine whether the list to be explored FS is empty. If it is empty, backtrack to find the parent node and output the path; if it is not empty, set the grid with the minimum cost as the current grid G i , the neighbor grid G j Add to the list to be explored FS;

[0008] Determine the current grid G j Is it equal to the target network G? t, then backtrack to find the parent node and output the path; otherwise check the grid G i The neighbor grid G j ;

[0009] Determine the neighbor grid G j Is it not in the explored list ES? If yes, add it to the list to be searched FS, and continue to return to determine whether the list to be explored FS is empty; otherwise, ignore the grid G j , continue to return to determine whether the list to be explored FS is empty.

[0010] Optionally, the escape rate is calculated as follows:

[0011]

[0012] Among them, R E is the escape rate, NP total is the total number of survivors, NP suc is the total number of people who successfully escaped.

[0013] Optionally, set D is a set of dangerous grids, and set D is calculated as follows:

[0014]

[0015] Where D is the set of dangerous grids, DG k is the dangerous grid k.

[0016] Optionally, assume an escape path sequence G = {G1, G2, .., G m}, when all grids in the sequence belong to the safe grid SG, it means that this is a safe escape path. The calculation is as follows:

[0017]

[0018] Among them, SG is the set of safe grids, m is the total number of grids, G k is the grid k.

[0019] Optional, the time T that a particular escapee spends in the escape process q , calculated as follows:

[0020]

[0021] Among them, T q is the time spent in the escape process, ms is the escapee's moving speed, c ij is a binary variable, and N is the total number of grids.

[0022] Optionally, the average time required to complete escape in the event of a fire, calculated as follows:

[0023]

[0024] Among them, T E is the average escape time, NP total is the total number of survivors.

[0025] Optional,

[0026]

[0027]

[0028]

[0029] G i =min(cost(FS))

[0030] Among them, FS is the list to be explored, ES is the explored list, and G is the set of all grids.

[0031] On the other hand, to achieve the above-mentioned purpose, the present invention also provides a real-time path planning system for fire scene evacuation, comprising: a data acquisition module, a data broadcast module, a position positioning module and a path selection module;

[0032] The data acquisition module is used to collect abnormal temperature or smoke concentration and dangerous grid location, and send them to the data broadcast module;

[0033] The data broadcast module is used to send data to users in the form of broadcast according to abnormal temperature or smoke concentration and dangerous grid location, and inform users that there are dangerous grids in the current area;

[0034] The location positioning module is used for the user to send its own location information to the path selection module after receiving the notification;

[0035] The path selection module is used to calculate an escape path based on the collected dangerous grid information, the number of exits and the degree of exit congestion, and complete real-time path planning.

[0036] Technical effects of the present invention:

[0037] (1) The present invention is based on the A* algorithm and makes full use of information ERP such as real-time fire scene status and indoor building environment status to dynamically adjust the evacuation path; compared with the traditional static path-finding method, ERP can more accurately adapt to the rapid evolution of the fire situation and improve the accuracy and timeliness of evacuation decisions.

[0038] (2) The present invention adopts ERP which not only considers the real-time perception of the fire situation, but also considers the congestion of escapees, and improves the escape efficiency and ensures the safety of personnel by dispersing the escape crowd. In addition, ERP is also suitable for multiple escapes and multiple fire source scenarios, which verifies its adaptability and scalability.

[0039] (3) The ERP used in the present invention shows superiority in terms of escape rate and dynamic flexibility. By comparing with the greedy method, the original A* algorithm, and the RRT algorithm, ERP can provide more accurate and efficient evacuation path planning during emergency evacuation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0041] Figure 1 An indoor bird's-eye view of the Jinrong Teaching Building in an embodiment of the present invention;

[0042] Figure 2 A schematic diagram of the system architecture and information transmission process of an embodiment of the present invention;

[0043] Figure 3 This is a diagram of the format of an emergency package according to an embodiment of the present invention;

[0044] Figure 4 The format diagram of the escape instruction packet according to an embodiment of the present invention;

[0045] Figure 5 The state of the first round of the embodiment of the present invention and the state of the escapee reaching the exit; wherein (a) is the state of the first round, and (b) is the state of the escapee reaching the exit;

[0046] Figure 6 This is a schematic diagram showing the effect of the number of fire points on the escape rate according to an embodiment of the present invention;

[0047] Figure 7 This is a schematic diagram of the effect of the number of ignition points on the path length according to an embodiment of the present invention;

[0048] Figure 8 This is a schematic diagram showing the effect of the number of ignition points on time efficiency according to an embodiment of the present invention;

[0049] Fig. 9 Schematic diagram of the effect of the number of escape exits on the escape rate according to an embodiment of the present invention;

[0050] Fig.10 Schematic diagram of the effect of the number of escape exits on the path length according to an embodiment of the present invention;

[0051] Fig.11Schematic diagram of the effect of the number of escape exits on time efficiency according to an embodiment of the present invention;

[0052] Fig.12 This is a flow chart of ERP path planning according to an embodiment of the present invention;

[0053] Fig.13 This is a flow chart of a method for balancing queue waiting time according to an embodiment of the present invention;

[0054] Fig.14 The present invention is a schematic diagram of the structure of a real-time path planning system for fire evacuation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0055] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0056] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0057] like Fig.12 As shown, this embodiment provides a real-time path planning method for fire evacuation, including:

[0058] Initialize the to-be-explored list FS, the explored list ES and the set P;

[0059] Determine whether the list to be explored FS is empty. If it is empty, backtrack to find the parent node and output the path; if it is not empty, set the grid with the minimum cost as the current grid G i , the neighbor grid G j Add to the list to be explored FS;

[0060] Determine the current grid G j Is it equal to the target network G? t , then backtrack to find the parent node and output the path; otherwise check the grid G i The neighbor grid G j ;

[0061] Determine the neighbor grid G j Is it not in the explored list ES? If yes, add it to the list to be searched FS, and continue to return to determine whether the list to be explored FS is empty; otherwise, ignore the grid G j , continue to return to determine whether the list to be explored FS is empty.

[0062] Figure 1This is a bird's-eye view of the teaching building "Jinrong Building" of Putian University. The area is divided into N grids. The movement process will be based on grids. The number of escape exits is M = 5. Assume that G is the grid set of the entire area, then Et is the set of escape exits, then

[0063] Assuming that temperature and smoke sensors are deployed in the building, in the present invention, the danger level is classified according to the fire danger level in the NFPA 704 standard, and the danger threshold is set when the temperature is greater than 80 degrees or the smoke concentration reaches 12000ppm. When the smoke concentration reaches or exceeds the threshold, the sensor transmits the temperature or smoke concentration information to the back-end server. The server uses the information collected by the sensor (including the location of the danger), the total number of exits and the congestion level of each exit to perform an ERP path planning program, and finally sends the path planning results to the user, who then escapes from the fire scene according to the instructions of the mobile device.

[0064] In the present invention, a path planning algorithm is designed to improve the escape rate of the escapee. E It can be defined as: When a fire occurs, the number of people who successfully escape NP suc NP total The first objective of the present invention is to improve the escape rate by planning a suitable path to ensure the safe evacuation of as many people as possible. The calculation method of the escape rate is shown in formula (1).

[0065]

[0066] Among them, R E is the escape rate, NP total is the total number of survivors, NP suc is the total number of people who successfully escaped.

[0067] When the escapee's moving speed is fixed, a shorter path length means a faster escape time and higher efficiency, that is, it takes less time for the escapee to escape from the disaster site. The path length is defined as the distance required to reach the safe exit from the starting point. According to the fire hazard level in the NFPA 704 standard, when the temperature is greater than 80 degrees or the smoke concentration reaches 12000ppm, it is set as a dangerous area. Therefore, the grids in the entire area can be divided into two categories: safe grids SG and dangerous grids DG. Assume that set D is a set composed of dangerous grids, and set D is expressed as formula (2),

[0068]

[0069] Where D is the set of dangerous grids, DG kis the dangerous grid k.

[0070] The constraint condition is that the path does not pass through the dangerous area, that is, all the grids in the path do not belong to DG. Assume an escape path sequence G = {G1, G2, .., G m}, when all grids in the sequence belong to the safe grid SG, it means that this is a safe escape path, as shown in formula (3).

[0071]

[0072] Among them, SG is the set of safe grids, m is the total number of grids, G k is the grid k.

[0073] The time taken by each person to escape is expressed as formula (4), where d(G i , G j ) is the distance from grid i to grid j, c ij is a binary variable. If the grids are adjacent, then c ij =1. Otherwise c ij = 0. ms is the moving speed of the escapee, so the time T that a certain escapee spends in the escape process can be obtained by formula (4): q .

[0074]

[0075] Among them, T q is the time spent in the escape process, ms is the escapee's moving speed, c ij is a binary variable, and N is the total number of grids.

[0076] However, the time efficiency is lower, which means that the escapee spends less time escaping. E Time Efficiency is defined as the average time required to complete escape when a fire occurs. Reducing the time spent on escape means that the escapees can escape from the dangerous area faster, as shown in formula (5).

[0077]

[0078] Among them, T E is the average escape time, NP total is the total number of survivors.

[0079] The research objective of the escape path planning algorithm ERP proposed in this invention can be expressed as formula (6). It aims to design an efficient escape path planning algorithm to ensure that the escapee can reach a safe place in the shortest time, maximize the escape rate, and consider the limitations of various constraints at the same time.

[0080]

[0081] like Figure 2 As shown, the present invention also provides a real-time path planning system for fire evacuation. The system consists of three roles: a mobile phone, a server, and a sensor. The information transmission process is divided into four steps. When a fire occurs, (1) when the sensor receives abnormal temperature or smoke density, it sends the location of the dangerous grid where the abnormal temperature or smoke density is detected to the server. (2) The server broadcasts the emergency package to the user to notify the user that there is a dangerous grid in the current area. (3) After receiving the notification, the user transmits his own location to the server. (4) The server calculates an escape route based on the information collected by the sensor (including the location of the danger), the number of exits, and the degree of exit congestion. Finally, the escapeinstruction packet is sent to the user, and the user can escape from the fire scene according to the instructions.

[0082] Figure 3 It is an emergency packet format. The emergency packet records the sensor ID, Sen ID Record which sensor sent the data packet. ID Then record the grid ID of the dangerous area. t It records the temperature of the dangerous grid, A c The smoke density in the dangerous area is recorded.

[0083] When the server receives the emergency packet from the sensor, it broadcasts the emergency packet to the user, and the user transmits his or her location to the server, which then performs the path planning. The server calculates the escape path based on the user's location, the location of the fire, the total number of exits, and the degree of congestion at the exits. Finally, the escape path trajectory is recorded in the escape instruction package and sent to the user. Figure 4 The format of the escape instruction packet. Timestamp is the time when the escape instruction packet is sent, which is used to determine the newness of the information. us ID is the user's ID, exit ID is the exit ID, pathtrajectory records the trajectory of the escape path, that is, the grid G ​​where the escapee is located i To Exit G g The path trajectory, that is, G = {G i , G i+1 , .., G g-1 , G g}.

[0084] However, in order to accurately grasp the situation at the fire scene, as soon as a fire occurs, the server will regularly broadcast the escape instruction package to the user after receiving the user's location information. In addition, if a new fire point is detected, the server will immediately send it to the user to inform the user of the current situation in the entire area.

[0085] After the sensor sends the location to the server, the ERP program will be executed. ERP can be divided into two parts: path planning and escape diversion mechanism. By considering the actual cost and heuristic estimated cost, the ERP algorithm can prioritize the path with the greatest hope of finding an exit during the search process, thereby reducing the search space and improving the search efficiency. However, when applied to fire scenarios, the distance between the escapee and the fire source must also be considered to ensure the escapee's safety. Therefore, the issues of concern are (1) whether the path is safe and (2) whether the exit is congested.

[0086] Whether the path is safe is determined by whether there is a dangerous grid on the escape path, as shown in formula (3). As for the problem of whether the exit is congested, in Section 3.4, an escape diversion mechanism is proposed, called the balanced queue waiting time method, which takes into account the queue waiting time and the load of the exit. By multiplying the weight factor by the ratio of the equivalent number of people to the capacity, the individual waiting time and the overall diversion fairness can be balanced when selecting the exit. The exit with the minimum value will be regarded as the best exit to achieve a balanced distribution of tasks.

[0087] When it is necessary to find the shortest path between two nodes in a graph or network, the A* algorithm is a commonly used heuristic search algorithm. It combines the ideas of breadth-first search and greedy best-first search of the Dijkstra algorithm, and guides the search direction by estimating the cost from the starting point to the target node, thereby improving the search efficiency. However, the original A* algorithm does not consider the location of the fire point, which may cause the planned escape path to be unadoptable. Secondly, the existing A* algorithm does not give a specific method for the selection of escape exits. Therefore, in the proposed ERP algorithm, the escape exit is first determined, and then the path planning is performed. In the process of path planning, the estimated cost (distance) between the escapee and the nearest fire source is continuously updated to avoid dangerous grids.

[0088] When the user receives the emergency package notification from the server, the user transmits his position to the server, and the server begins the path planning task for the escapee. The ERP algorithm is divided into two steps: First, the escape exit is determined. Section 3.4 will explain how to determine the escape exit. After the exit is determined, the path planning procedure is performed. In order to clearly explain the path planning process of ERP, Figure 5For example, assume that the exit is at the grid position (7, 11) in the lower right corner of the figure, the fire source is at the grid position (2, 8), and there are 8 directions for escape (up, down, left, right and diagonal). Two data structures are used in the algorithm, namely the to-be-explored list and the explored list. The to-be-explored list is used to store grids to be explored, which have been discovered but not yet fully explored. The algorithm will select the next grid to be explored from the to-be-explored list. The explored list is used to store grids that have been fully explored, that is, their optimal path has been found and all possible neighbor grids have been considered. Once a grid is placed in a closed list, it will not be further explored or expanded.

[0089] First, initialize and set the starting grid to the current grid G i (Current location), G t Target grid (escape exit) and danger grid G f ,Step 1. The current grid is added to the to-be-explored list FS, as shown in formula (7). If the current grid is the target grid, it means that the path is found and the search ends.

[0090]

[0091] Step 2. View the current grid G i Adjacent (one-step reachable) grid G j , G j Add FS, G to the list to be explored j And record G i is the parent grid, as shown in formula (8).

[0092]

[0093] Due to G i It has been explored, so G i Remove it from the list to be explored FS and put G i Add it to the explored list ES, as shown in formula (9).

[0094]

[0095] Step 3. Take the grid with the minimum cost from the list to be explored FS and set it as the current grid G i , as shown in formula (10). After setting the new current grid G i Then go back to step 1. If the target grid is not found when the list to be explored is empty, it means there is no feasible path.

[0096] G i =min(cost(FS)) (10)

[0097] Formula (11) is the definition of the ERP cost function C(n), where AC(n) represents the actual cost and EC(n) represents the estimated cost. The estimated cost includes two parts: the estimated cost from the current grid where the escapee is to the exit grid and the estimated cost of the grid closest to the danger grid where the escapee is currently located.

[0098] C(n)=AC(n)+EC(n)

[0099] EC(n)=dist(G i , G t )+min(dist(G i , G f ))1≤f≤k (11)

[0100] The present invention uses Manhattan distance to calculate the distance between the current grid of the escapee and the exit or the minimum distance (shortest) between the current grid of the escapee and the dangerous grid, as shown in formula (12).

[0101] dist(G i , G j )=|x i -x j |+|y i -y j |, where 1≤i, j≤N (12)

[0102] Among them, G i The position of (x i ,y i ), G j The coordinate position is (x j ,y j ). Formula (13) is the calculation method for selecting the current grid. From the list to be explored, the grid with the minimum cost is found as the current grid for the next round.

[0103] min(C(n)) (13)

[0104] By tracing the parent grid in reverse, formula (14) is that each time a parent grid is found, it is added to the set P. Starting from the target grid, it is traced to the parent grid in sequence until the starting grid is reached to find the shortest path P.

[0105] P=P∪{G i} (14)

[0106] exist Figure 5 In (a), the survivor starts from grid (4, 3), calculates neighboring grids, and finds that grid (5, 4) has the smallest cost, so he chooses grid (5, 4) as the position to move to in the next round. Figure 5(b) shows the state of the escapee reaching the target grid. Algorithm I is the pseudocode of the ERP algorithm to find the shortest path, where the input is: the current grid G i , target grid G t , the dangerous grid set DG and the grid set G. Finally, the escape path P is output.

[0107] In ERP, the crowding degree of the escape exit is taken into consideration and the escapees are diverted to prevent all escapees from escaping to the same exit, which would cause exit congestion. Therefore, during the process of path planning, the server will continuously update the status of each exit and count the number of escapees at each exit.

[0108] Exits are allocated by the shortest queue time and the equivalent evacuation rate. The shortest queue time focuses on the waiting time of individual escapees, which can achieve balanced allocation of individuals, reduce waiting time, and improve individual satisfaction. The equivalent number of people focuses on the fairness of the overall diversion. By evenly allocating the number of people and the diversion ratio, the load of the escape exit can be relatively balanced to avoid overcrowding.

[0109] In the shortest queue time, observe the current queue waiting time of each exit and select the exit with the shortest queue time as the escape exit for the escapee. By observing the waiting time of each escapee in the queue, the escapee is assigned to the exit with the shortest waiting time to achieve a balanced distribution of tasks. QW ) can be defined as the longest escape time minus the individual escape time, as shown in formula (15):

[0110] T QW =T ML -T IE (15)

[0111] Where T IE For each person, the time it takes from receiving the escape instruction package from the server to successfully escaping the building can be recorded. ML For the longest escape time, determine the person with the longest escape time among all escapees, use it as a reference indicator, and use this time as the longest escape time (Longest Evacuation Time). For example, suppose a fire occurs, A takes 2 minutes to escape, and B takes 3 minutes to escape. The longest escape time is 3 minutes (determined by B's escape time). Then, A's waiting time in the queue is the longest escape time minus A's escape time, that is, 3 minutes - 2 minutes = 1 minute.

[0112] In order to make the load of the escape exit relatively balanced and avoid overcrowding, the balanced queue waiting time method is proposed. The queue waiting time and the load of the exit are comprehensively considered. By multiplying the weight factor by the ratio of the equivalent number of people to the capacity, the individual waiting time and the overall diversion fairness can be balanced when selecting the exit. The exit with the minimum value will be regarded as the best exit to achieve balanced distribution of tasks, as shown in formula (16). By observing the queue waiting time and equivalent number of people at each exit, the balanced queue waiting time method can dynamically adjust the exit allocation during the escape process to minimize the waiting time and ensure balanced exit load.

[0113]

[0114] Among them, 1≤i≤n, 0≤ω≤1, TQW i represents the waiting time in the queue at exit i. i The effective population of exit i, ec i represents the maximum capacity of each exit i. ω is a weight factor, and the calculation method of the equivalent number of people is shown in formula (17).

[0115]

[0116] Among them, 1≤i≤n.

[0117] To clearly illustrate how the balanced queue waiting time method is calculated, assume that there is a fire in a building with three escape exits available. The total number of people in the building is 300, and the maximum capacity of the three escape exits is 100, 150, and 200 respectively. First calculate the equivalent number of people at exits 1-3, and then use the balanced queue waiting time method to select the best exit.

[0118] (1) Calculate the equivalent number of people at exits 1-3

[0119] Equivalent number of people for exit 1:

[0120]

[0121] Equivalent number of people for exit 2:

[0122]

[0123] Equivalent number of people for exit 3:

[0124]

[0125] Use the formula of the balanced queue waiting time method to select the best exit. Assume that the queue waiting times of exits 1-3 are 5 minutes, 3 minutes, and 4 minutes respectively, and the weight factor ω is 0.5. According to formula (16), calculate the balanced queue waiting time of each exit.

[0126] (2) Calculate the waiting time for exits 1-3

[0127] Balanced queue waiting time at exit 1 (minutes):

[0128]

[0129] Balance queue waiting time at exit 2 (minutes):

[0130]

[0131] Balanced queue waiting time at exit 3 (minutes):

[0132]

[0133] According to the calculation results, it is found that Exit 2 has the smallest value, which is 3.17 minutes. Therefore, according to the balanced queue waiting time method, Exit 2 will be selected as the best exit. This means that during the escape process, people should give priority to Exit 2 to achieve a balanced distribution of tasks and reduce waiting time. It is worth noting that as the escapees are constantly assigned to each exit, the calculation results of the balanced queue waiting time will also change. The balanced queue waiting time method can dynamically adjust the exit allocation during the escape process to minimize the waiting time and ensure the balanced exit load.

[0134] Experimental parameters

[0135] This method uses an Intel core i9-12900k processor, DDR4 8GB memory, and Windows 11 operating system. The simulation is developed using Java, while data analysis and drawing are performed using Matlab and Gunplot. In the part of generating fire data, indoor fire data is generated by Fire Dynamics Simulator (FDS), which is an advanced indoor fire simulation software developed by the National Institute of Standards and Technology (NIST). This method simulates a plane at a height of 1.5 meters above the ground in the space, and cuts the entire area plane into N 1m*1m grids. Table 1 shows the experimental environment parameters and settings.

[0136] Table 1

[0137]

[0138]

[0139] In order to measure the effectiveness of the algorithm, the present invention uses three measurement criteria for comparison, namely escape rate, path length and time efficiency. Since the fire scene is a dynamic environment, the escape rate can reflect the adaptability and robustness of the algorithm in a dynamic environment. (1) Escape rate R E It can be defined as: When a fire occurs, the number of people who successfully escape NP suc NP total The first objective of the present invention is to improve the escape rate by planning a suitable path to ensure the safe evacuation of as many people as possible. The calculation method of the escape rate is shown in formula (18).

[0140]

[0141] (2) Path Length: The path length PL is the distance from the starting point to the safe exit. A shorter path length means a faster escape time and higher efficiency. The path length is defined as the sum of the path lengths of the escapees who successfully escaped divided by the total number of successful escapees. The calculation method is shown in formula (19). i is the path length of survivor i.

[0142]

[0143] (3) Time Efficiency: Time efficiency T E (Time Efficiency) is the average time required to complete escape when a fire occurs. Reducing the time spent on escape means that the escapees can escape from the dangerous area faster, as shown in formula (20).

[0144]

[0145] The impact of the number of fire points

[0146] This set of experiments aims to verify the impact of the number of fire points on the four algorithms. In order to simulate different scenarios, 20 obstacles of 6 grids in size were randomly generated, and the number of fire points was gradually increased from 0 to 9. The fire spread rate was fixed at 1 grid / second, the escape exit was fixed at 6, the exit capacity was 20 people, and the escaper's movement rate was fixed at 2 grids / second. The position of the escaper in the environment was also randomly generated. According to Figure 6 The experimental results in show that as the number of fire points increases, the escape rate gradually decreases. When the number of fire points reaches 9, the escape rate of ERP is about 51%, while the greedy method is about 21%. In comparison, the escape rate of ERP is about 30% higher. The experiment confirms that strategic escape routes help improve the escape rate.

[0147] Figure 7 It shows the path length required by the escapee during the escape process. It can be observed that as the number of fire points increases, the path length gradually decreases. This is because the escape rate decreases and only the path length of successful escapees is counted, so the required path length gradually decreases. The path lengths of the ERP and A* algorithms are close. The RRT algorithm has a 50% chance of choosing random movement, resulting in many unnecessary paths. The greedy method has a shorter path because only the path length of successful escapees is counted, so the path is shorter. Experiments have confirmed that the number of fire points has little effect on the path length.

[0148] according to Figure 8 The experimental results in show that when the number of fire points is 9, the time efficiency of ERP and A* algorithms is close, 9.64 and 10.57 respectively, while the time efficiency of RRT algorithm is the worst, about 17.57, that is, the time required for escape is longer. The time efficiency of the greedy method is the best, about 7.72. The experiment confirms that the number of fire points has little effect on time efficiency.

[0149] Effect of the number of escape exits

[0150] This set of experiments aims to verify the impact of the number of escape exits on the four algorithms. 20 obstacles of 6 grids in size were randomly generated, and the number of fire points was gradually increased to 3, the fire spread rate was fixed to 1 grid / second, and the escape exits were fixed from 2 to 11, with a capacity of 20 people per exit. The escapee's movement rate was fixed to 2 grids / second, and the escapee's position in the environment was randomly generated.

[0151] according to Fig. 9 The experimental results in show that the escape rate gradually increases with the increase in the number of escape exits. When the number of escape exits is 2, the escape rate of ERP is 78.5%, RRT and A* algorithms reach 72.5% and 73.7% respectively, while the greedy method is about 37.8%. In comparison, the escape rate of ERP is about 40.7% higher than Greedy. When the number of escape exits is 11, the escape rate of ERP is 97.7%, RRT and A* algorithms reach 92.9% and 92.7% respectively, while the greedy method is 74.5%. In comparison, the escape rate of ERP is about 23.2% higher than Greedy. The experiment confirms that a larger number of escape exits helps to improve the escape rate.

[0152] Fig.10The figure shows the path length required by the escapee during the escape process. It can be observed that as the number of escape exits increases, the path length gradually decreases. The path lengths of ERP, A* and Greedy are close. The RRT algorithm generates many unnecessary paths because it has a 50% chance of choosing random movement. The experiment confirms that as the number of escape exits increases, there is a greater probability of finding an exit that is shorter than the escapee.

[0153] according to Fig.11 The experimental results in show that when the number of escape exits increases, it helps the escapee find an exit that is closer to him, thus shortening the time required for escape. When the number of escape exits is 2, ERP, A* and Greedy are similar, 14.28, 14.52 and 14.47 respectively, while the time efficiency of the RRT algorithm is the worst, about 31.35. When the number of escape exits is 11, ERP, A* and Greedy are similar, 10.24, 10.7 and 10.47 respectively, while the time efficiency of the RRT algorithm is the worst, about 25.75, that is, the time required for escape is the longest. The experiment confirms that the more escape exits there are, the shorter the time it takes to escape.

[0154] Fig.13 This is the flow chart of the escape diversion mechanism of ERP. Since there may be multiple exits at the fire scene, before path planning, it is necessary to decide which exit is more suitable for escape (i.e., the target exit). First, initialize the equivalent number of people at each exit and count the capacity of each exit.

[0155] The input parameter is the total number of people in the building NP total The output of the function is the optimal exit BestExit. Create an empty array EP to store the equivalent number of people. Calculate the total capacity of all escape exits by adding all elements in the array ec.

[0156] Start a loop and traverse each element in the array ec, with indexes from 0 to n-1. First calculate the equivalent number of people EP at exit i i , by taking the total number of people in the building NP total Multiply by ec i The ratio to the total capacity is obtained.

[0157] The calculated equivalent number of people at each exit i is recorded in the EP array.

[0158] Initialize the minimum time min_time to infinity. Initialize the best exit BestExit to -1.

[0159] Start a loop and traverse each element in the array ec, with indexes from 0 to n-1. Calculate the time weight value TQW of each escape exiti . Calculate the waiting time of each exit, waiting_time, by using TQW i Plus (ω*EP i / ec i ) is obtained, where ω is the weight factor, which is set to 0.5 in the present invention.

[0160] If the waiting time waiting_time is less than the minimum time min_time, perform the following steps.

[0161] Update min_time to waiting_time and record the current best exit BestExit. When all exits have been considered, return BestExit as the output of the function.

[0162] like Fig.14 As shown, the present invention also provides a real-time path planning system for fire scene evacuation, including: a data acquisition module, a data broadcast module, a position positioning module and a path selection module;

[0163] The data acquisition module is used to collect abnormal temperature or smoke concentration and dangerous grid location, and send them to the data broadcast module;

[0164] The data broadcast module is used to send data to users through broadcast according to abnormal temperature or smoke concentration and dangerous grid location, informing users that there are dangerous grids in the current area;

[0165] The location positioning module is used to send the user's location information to the path selection module after receiving the notification;

[0166] The path selection module is used to calculate an escape path based on the collected dangerous grid information, the number of exits and the degree of exit congestion, and complete real-time path planning.

[0167] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A real-time path planning method for fire evacuation, characterized in that: include: Initialize the to-be-explored list FS, the explored list ES and the set P; Determine whether the list to be explored FS is empty. If it is empty, backtrack to find the parent node and output the path; If it is not empty, the grid with the minimum cost is set as the current grid G i , the neighbor grid G j Add to the list to be explored FS; Determine the current grid G j Is it equal to the target network G? t If yes, then backtrack to find the parent node and output the path; otherwise, check the grid G i The neighbor grid G j ; Determine the neighbor grid G j Is it not in the explored list ES? If yes, add it to the list to be searched FS, and continue to return to determine whether the list to be explored FS is empty; otherwise, ignore the grid G j , continue to return to determine whether the list to be explored FS is empty.

2. The real-time path planning method for fire evacuation according to claim 1, characterized in that: The escape rate is calculated as follows: Among them, R E is the escape rate, NP total is the total number of survivors, NP suc is the total number of people who successfully escaped.

3. The real-time path planning method for fire evacuation according to claim 1, characterized in that: Set D is the set of dangerous grids, and set D is calculated as follows: Where D is the set of dangerous grids, DG k is the dangerous grid k.

4. The real-time path planning method for fire evacuation according to claim 1, characterized in that: Assume an escape path sequence G = {G1, G2, .., G m }, when all grids in the sequence belong to the safe grid SG, it means that this is a safe escape path. The calculation is as follows: Among them, SG is the set of safe grids, m is the total number of grids, G k is the grid k.

5. The real-time path planning method for fire evacuation according to claim 1, characterized in that: The time T that a certain escapee spends in the process of escaping q , calculated as follows: s.t.c ij ={0,1},where i,j=1,2,3,..,N Among them, T q is the time spent in the escape process, ms is the escapee's moving speed, c ij is a binary variable, and N is the total number of grids.

6. The real-time path planning method for fire evacuation according to claim 1, characterized in that: The average time required to complete escape in the event of a fire is calculated as follows: Among them, T E is the average escape time, NP tatal is the total number of survivors.

7. The real-time path planning method for fire evacuation according to claim 1, characterized in that: G i =min(cost(FS)) Among them, FS is the list to be explored, ES is the explored list, and G is the set of all grids.

8. The system of the real-time path planning method for fire scene evacuation according to any one of claims 1 to 7, characterized in that: include: Data collection module, data broadcast module, location positioning module and path selection module; The data acquisition module is used to collect abnormal temperature or smoke concentration and dangerous grid location, and send them to the data broadcast module; The data broadcast module is used to send data to users in the form of broadcast according to abnormal temperature or smoke concentration and dangerous grid location, and inform users that there are dangerous grids in the current area; The location positioning module is used for the user to send its own location information to the path selection module after receiving the notification; The path selection module is used to calculate an escape path based on the collected dangerous grid information, the number of exits and the degree of exit congestion, and complete real-time path planning.

Images