Fire-fighting path planning system and method based on multi-objective optimization

Abstract

The invention belongs to the technical field of path planning, and discloses a fire-fighting path planning system and method based on multi-objective optimization, and the method comprises the steps: determining an influence coverage area based on the coordinates of a fire occurrence point, determining the number of personnel based on the influence coverage area, determining a personnel gathering area based on the personnel coordinates, and distinguishing the gathering crowd based on the personnel gathering area. Based on the positions of the gathered crowds, planning a fire-fighting escape path of each gathered crowd by using a path planning algorithm; if the fire-fighting escape paths corresponding to the different gathered crowds are overlapped, predicting crowd intersections of the different gathered crowds in the fire-fighting escape paths at future moments; the number of people borne in the next change area of the fire-fighting escape path is predicted based on the crowd intersection, and whether gathered crowds which are overlapped on the fire-fighting escape path and have the crowd intersection are shunted or not is determined based on the number of people borne; congestion and stampede events are avoided, the escape safety of personnel is guaranteed, and the reliability of the fire-fighting evacuation system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of path planning, and more specifically, to a fire path planning system and method based on multi-objective optimization. Background Art

[0002] In the traditional judgment mode of crowd convergence points, it mainly relies on real-time monitoring means. This means that only when people significantly show a gathering trend in a certain area, or even show signs of congestion, can relevant measures be initiated. This ex post facto processing method has significant drawbacks, and its reaction speed is extremely slow. The time consumed from detecting an abnormal situation to initiating countermeasures greatly compresses the precious time left for fire emergency response. In the event of an emergency of large-scale crowd convergence and congestion, due to the lag in the early response, it is very difficult to quickly organize an effective evacuation plan within a very limited time. At this time, the situation is extremely likely to fall into a "deadlock" state, where people cannot flow and evacuate as expected, and the evacuation work thus reaches a deadlock, ultimately making it difficult for people to evacuate in an orderly manner and posing a huge threat to life safety.

[0003] In the actual operation of fire emergency management, the lack of this predictive ability is directly reflected in the diversion decision-making. Due to the failure to detect potential congestion areas in a timely manner and the lack of reasonable diversion and re-planning of fire paths, a large number of people gather excessively in dangerous changing areas such as narrow corridor corners and stairwells. In these places with limited space and dense personnel, once a crush occurs, it is very likely to trigger serious stampede accidents, which undoubtedly greatly reduces the efficiency of fire evacuation work, poses a serious threat to the life safety of people, and greatly reduces the safety of fire evacuation.

[0004] In view of this, the present invention proposes a fire path planning system and method based on multi-objective optimization to solve the above problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art and achieve the above object, the present invention provides the following technical solution: A fire path planning method based on multi-objective optimization, including:

[0006] Determine the affected coverage area based on the coordinates of the fire occurrence point, determine the number of people based on the affected coverage area, determine the personnel gathering area based on the personnel coordinates, and distinguish the gathered crowd based on the personnel gathering area;

[0007] Based on the positions of the gathered crowd, use a path planning algorithm to plan the fire escape paths for each gathered crowd.

[0008] If there is an overlap in the fire escape paths corresponding to different gathered crowds, predict the crowd convergence points of different gathered crowds on the fire escape paths at future moments;

[0009] Predict the number of people that can be endured in the next changing area of the fire escape path based on the crowd intersection point, and determine whether to divert the crowded crowd on the overlapping fire escape paths with crowd intersection points based on the number of people that can be endured;

[0010] If diversion is to be carried out, determine the number of people to be diverted based on the number of people that can be endured in the changing area;

[0011] Select other fire escape paths that intersect with the fire escape path behind the crowd intersection point, and divert the corresponding crowded crowd, and re-plan the fire escape path, where the diversion quantity is the number of people to be diverted, and the direction behind is opposite to the advancing direction of the fire escape path.

[0012] Furthermore, the method for determining the affected coverage area based on the fire occurrence point coordinates includes:

[0013] A crowd determination module receives the fire occurrence point coordinates sent by the management server, takes the fire occurrence point coordinates as the origin, draws a circle with a preset influence radius to obtain a circular area, and uses the circular area as the affected coverage area; the fire occurrence point coordinates are determined by the management server according to the position sent by the fire information alarm device.

[0014] Furthermore, the method for determining the number of people based on the affected coverage area includes:

[0015] Receive the number of people determined in the affected coverage area sent by the management server; when the management server receives the fire information, continuously send statistical information to the communication server, and the statistical information includes the identifiers of all signal repeaters within the affected coverage area; the communication server obtains the corresponding number of devices connected to all signal repeaters within the affected coverage area according to the identifiers of all signal repeaters within the affected coverage area, and sends the number of devices connected to each signal repeater to the management server; the management server uses the number of devices as the number of people determined in the affected coverage area.

[0016] Furthermore, the method for determining the crowd gathering area based on the personnel coordinates includes:

[0017] Take the installation position coordinates of the signal repeater as the device coordinates of the devices connected to the same signal repeater, and the device coordinates are the corresponding personnel coordinates; take the personnel corresponding to the devices connected to the same signal repeater as the same crowd gathering area.

[0018] Furthermore, the position of the crowded crowd corresponds to the installation position of the signal repeater. According to the fire exits predetermined in the affected coverage area, use the A* algorithm to plan the fire escape paths for each crowded crowd based on the position of the crowded crowd and the fire exits.

[0019] Further, input the number of people, locations, panic coefficients, and the average width of the corresponding fire escape paths corresponding to different aggregated populations into the trained travel prediction model to obtain the travel of the corresponding populations at multiple future times;

[0020] Determine the travel cut-off points of the aggregated populations corresponding to multiple future times on the fire escape path according to the travel at multiple future times, and mark the cut-off points on the fire escape path; if the distance between the cut-off points at the same future time is lower than the predetermined distance threshold, then use the cut-off point in the front as the crowd intersection point, and the front direction is consistent with the forward direction of the fire escape path.

[0021] Further, the method for predicting the number of people that can be endured in the next change area of the fire escape path based on the crowd intersection point includes:

[0022] The direction of the next change area is consistent with the forward direction of the fire escape path; the change area is an area where the turning angle or width of the area is lower than the corresponding preset threshold;

[0023] Input the turning angle, width, average panic coefficient, and the sum of the number of people corresponding to the existence of the crowd intersection point of the next change area into the pre-trained endurance prediction model to obtain the number of people that can be endured in the change area; the average panic coefficient is the average panic coefficient of the aggregated populations corresponding to the crowd intersection point.

[0024] The panic coefficient is the number of screams per unit time,

[0025] Further, the method for determining whether to divert the aggregated populations on the overlapping fire escape paths and with crowd intersection points based on the number of people that can be endured includes:

[0026] If the sum of the number of people corresponding to the existence of the crowd intersection point exceeds the number of people that can be endured in the change area, then diversion is required, otherwise it is not.

[0027] Further, the method for determining the number of diverted people based on the number of people that can be endured in the change area includes:

[0028] Take the difference between the sum of the number of people of the aggregated populations corresponding to the existence of the crowd intersection point and the number of people that can be endured in the change area as the number of diverted people planned in advance.

[0029] A fire path planning system based on multi-objective optimization is used to implement the fire path planning method based on multi-objective optimization. The system includes:

[0030] A crowd determination module, which determines the affected coverage area based on the coordinates of the fire occurrence point, determines the number of people based on the affected coverage area, determines the personnel aggregation area based on the personnel coordinates, and distinguishes the aggregated populations based on the personnel aggregation area;

[0031] The first planning module plans the fire escape paths for each gathered crowd based on the locations of the gathered crowds using a path planning algorithm.

[0032] The first analysis module predicts the crowd intersection points of different gathered crowds on the fire escape paths at future moments if there are overlaps in the fire escape paths corresponding to different gathered crowds;

[0033] The second analysis module predicts the number of people that can be tolerated in the next changing area of the fire escape path based on the crowd intersection points, and determines whether to divert the gathered crowds that are on the overlapping fire escape paths and have crowd intersection points based on the number of people that can be tolerated;

[0034] The third analysis module determines the number of people to be diverted based on the number of people that can be tolerated in the changing area if diversion is to be carried out;

[0035] The second planning module selects other fire escape paths that have intersections with the fire escape path behind the crowd intersection points to divert the corresponding gathered crowds, re-plans the fire escape paths, and the number of diverted people is the number of people to be diverted, and the direction behind is opposite to the forward direction of the fire escape path.

[0036] Advantages of the fire path planning system and method based on multi-objective optimization of the present invention:

[0037] By inputting data such as real-time number of people, location, panic coefficient, and average width of the fire escape path into the travel prediction model, potential crowd intersection points can be accurately predicted in advance, gaining valuable time for fire emergency response, greatly advancing the warning timeliness, and avoiding getting into a "deadlock" state. At the same time, using the tolerance prediction model, based on the turning angle, width, average panic coefficient, and sum of the number of people in the changing area, the number of people that can be tolerated in the next changing area is accurately predicted, and based on this, it is judged whether diversion is needed. Once diversion is determined, the scheme will cleverly select the appropriate crowd for diversion, accurately allocate the number of diverted people, re-plan the fire escape paths and convey them to the management personnel in a timely manner to balance the flow of people on each path, avoid congestion and stampede incidents, ensure the safety of people's escape, and improve the reliability and emergency response ability of the fire evacuation system. Brief Description of the Drawings

[0038] Figure 1 Schematic diagram of the fire path planning system based on multi-objective optimization of the present invention;

[0039] Figure 2 Schematic diagram of the interaction between the alarm device, management server, communication server, and crowd determination module of the present invention;

[0040] Figure 3 Schematic diagram of the flow of the fire path planning method based on multi-objective optimization of the present invention. Detailed Embodiments

[0041] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0042] Embodiment 1

[0043] Refer to Figure 1 , the fire path planning system based on multi-objective optimization described in this embodiment includes:

[0044] A crowd determination module, which determines the affected coverage area based on the coordinates of the fire occurrence point, determines the number of people based on the affected coverage area, determines the crowd gathering area based on the coordinates of the people, and distinguishes the gathered crowd based on the crowd gathering area.

[0045] The method for determining the affected coverage area based on the coordinates of the fire occurrence point includes:

[0046] When the management server receives the fire information, it determines the coordinates of the fire occurrence point according to the location of the fire information alarm device. Taking the coordinates of the fire occurrence point as the origin, a circle is drawn with a preset influence radius to obtain a circular area, and the circular area is used as the affected coverage area. The preset initial influence radius can be preset according to the area or building type corresponding to the fire occurrence point.

[0047] The method for determining the number of people based on the affected coverage area includes:

[0048] Refer to Figure 2 , when the management server receives the fire information, it continuously sends statistical information to the communication server (such as a mobile communication server, a telecommunications communication server, a China Unicom communication server, etc.). The statistical information includes the identifiers of all signal repeaters within the affected coverage area. The communication server obtains the corresponding number of devices (such as the number of mobile phones connected to the signal repeater) currently connected to all signal repeaters within the affected coverage area according to the identifiers of all signal repeaters within the affected coverage area, and sends the number of devices connected to each signal repeater to the management server. The management server uses the number of devices as the number of people determined by the affected coverage area and sends it to the crowd determination module.

[0049] Utilizing the infrastructure of signal repeaters, which is widely distributed and closely connected to personnel and devices, naturally has the ability to track the connection status of personnel and devices in real time. The information interaction process between the management end and the communication server is designed to be simple and efficient. Relying on a mature communication network architecture, data acquisition and feedback are almost without delay. Taking the number of devices as the number of people in the coverage area, compared with identifying people through cameras and then determining the number of people, it does not involve images and personal identity information, has a lower privacy risk, and there are no blind spots in the field of vision; in the case of insufficient light, it can also estimate the number of people; it overcomes the limitations of environmental factors such as light on traditional visual recognition means, thus achieving rapid and accurate statistics of the number of people and privacy protection.

[0050] The method for determining the personnel gathering area based on personnel coordinates includes:

[0051] Taking the installation position coordinates of the signal repeater as the device coordinates of the devices connected to the same signal repeater, and the device coordinates are the corresponding personnel coordinates; taking the personnel corresponding to the devices connected to the same signal repeater as the same personnel gathering area.

[0052] The first planning module plans the fire escape paths for each gathering crowd using a path planning algorithm based on the position of the gathering crowd.

[0053] The position of the gathering crowd corresponds to the installation position of the signal repeater. According to the predetermined fire exits in the coverage area, use the A* algorithm to plan the fire escape paths for each gathering crowd based on the position of the gathering crowd and the fire exits.

[0054] The first analysis module, if there are overlaps in the fire escape paths corresponding to different gathering crowds, predicts the crowd intersection points of different gathering crowds at the future moment on the fire escape paths.

[0055] Input the number of people, positions, panic coefficients, and the average width of the corresponding fire escape paths of different gathering crowds into the trained travel prediction model to obtain the travel of the corresponding crowds at multiple future moments. The average width of the fire escape path is the average of the widths of multiple preset selected areas of the fire escape path.

[0056] Determine the travel cut-off points of the corresponding gathering crowds at multiple future moments on the fire escape paths according to the travel at multiple future moments, and mark the cut-off points on the fire escape paths; if the distance between the cut-off points at the same future moment is lower than the predetermined distance threshold, then take the cut-off point in the front as the crowd intersection point, and the front direction is the same as the forward direction of the fire escape path.

[0057] For example, the fire escape routes of the gathered crowd A and the gathered crowd B overlap. For the gathered crowd A, it moves 100 meters in the next minute. Starting from the current position of the gathered crowd A, the 100-meter distance at the position on the fire escape route is the first travel cut-off point; and so on, 180 meters in the next two minutes to obtain the second travel cut-off point, and 300 meters in the next three minutes to obtain the third travel cut-off point. Similarly, for the gathered crowd B, the first, second, and third travel cut-off points are obtained. The positions of the first and second travel cut-off points of the gathered crowd A and the gathered crowd B at the same moment on the fire escape route are relatively far apart, while the third travel cut-off points are relatively close, that is, lower than the predetermined distance threshold. If the third travel cut-off point of the gathered crowd B is in front of the third travel cut-off point of the gathered crowd A, then the third travel cut-off point of the gathered crowd B is the crowd intersection point.

[0058] Through the advance travel prediction in multiple time dimensions, the system can identify potential intersection points 1 - 3 future time units in advance. For example, the prediction of the intersection at the third time point in the example leaves a 3-minute golden disposal window for fire emergency response, improving the early warning and disposal efficiency compared with the traditional real-time monitoring mode. Based on the forward selection principle of the moving direction (such as selecting the forward cut-off point of the gathered crowd B), it ensures that the escape route maintains one-way fluidity and avoids the "deadlock" phenomenon caused by the cross of oncoming crowds. Based on the above-mentioned predicted crowd intersection points in advance, it provides a basis for the subsequent forward congestion prediction, re-planning of the fire escape route, diversion of the gathered crowd, and determination of the number of people to be diverted.

[0059] The training method of the travel prediction model is as follows:

[0060] Data collection: Collect a large amount of actual data including the number of people, positions, panic coefficients, average widths of fire escape routes, and corresponding future crowd travel of different gathered crowds. This data can come from actual fire drills, surveillance records in public places, simulation experiments, etc.

[0061] Data cleaning: Check for missing values and outliers in the data and perform corresponding processing, such as deleting or filling missing values, correcting or deleting outliers.

[0062] Data normalization: Normalize the data to map all feature values to the range of 0 to 1 or -1 to 1 to accelerate model convergence and improve model stability.

[0063] Data division: Divide the preprocessed data into a training set, a validation set, and a test set, usually in a ratio of 7:2:1.

[0064] Model selection and construction

[0065] Select a model architecture, such as LSTM, GRU, etc. These models can effectively process sequence data and are suitable for predicting the future crowd travel.

[0066] Taking LSTM as an example, a model including an input layer, multiple LSTM layers, a fully connected layer, and an output layer is constructed. The number of neurons in the input layer should be the same as the number of input features, namely the number of people, location, panic coefficient, and the mean width of the fire escape path. The number of neurons in the LSTM layer can be adjusted according to actual conditions, and 64, 128, 256, etc. are generally selected. The fully connected layer is used to further process and integrate the output of the LSTM layer. The number of neurons in the output layer should be the same as the predicted target number, namely the itinerary of the corresponding crowd at the future moment.

[0067] Model Training

[0068] Select a suitable loss function to measure the difference between the model prediction results and the actual results, such as mean square error (MSE), mean absolute error (MAE), etc.

[0069] Choose a suitable optimizer to update the parameters of the model, such as stochastic gradient descent (SGD), Adagrad, Adade lta, Adam, etc. The Adam optimizer usually performs better in most cases. It can adaptively adjust the learning rate and speed up the convergence of the model.

[0070] The training data in the training set is input into the model, the model's prediction results are calculated through forward propagation, and then the loss between the prediction results and the actual results is calculated according to the loss function, and then the model parameters are updated through back propagation. This process is repeated until the model converges or the preset number of training rounds is reached, and the training is completed.

[0071] The second analysis module predicts the number of people in the next change area of ​​the fire escape path based on the crowd intersection point, and determines whether to divert the gathered crowd on the overlapping fire escape path and with the crowd intersection point based on the number of people.

[0072] The method for predicting the number of people in the next change area of ​​the fire escape path based on the crowd intersection point includes:

[0073] The direction of the next change area is consistent with the direction of the fire escape path; the change area is the area where the turning angle and width are lower than the corresponding preset threshold. In such an area, the smaller the turning angle and width, the more people need to pass through at the same time, and the more likely congestion and trampling will occur. In order to avoid this situation from happening in the change area, advance prediction is required.

[0074] The sum of the turning angle, width, panic coefficient mean value, and the number of people corresponding to the crowd intersection point of the next change area is input into the pre-trained tolerance prediction model to obtain the number of people to be tolerated in the change area. The panic coefficient mean value is the panic coefficient mean value of the corresponding crowd gathered at the crowd intersection point.

[0075] The sum of the population numbers reflects the population density and affects the change of the panic coefficient. As fixed influencing factors, the turning angle and width limit the upper limit of the predicted number of people that can be accommodated, preventing the predicted number of people that can be accommodated from deviating from the actual situation and ensuring the prediction accuracy of the later accommodation prediction model.

[0076] For the same change area, under different panic coefficients and sums of population numbers, the actual number of people that can be accommodated is also different. When the panic coefficient is small and other factors remain unchanged, the number of people that can be accommodated is higher than that when the panic coefficient is large. During a panic, the personnel flow capacity decreases, increasing the risk of congestion and trampling. The number of people that can be accommodated here is the maximum number of people who can pass safely under the conditions of the turning angle, width, average panic coefficient, and sum of population numbers corresponding to the presence of a crowd intersection point in the change area.

[0077] The training method of the accommodation prediction model includes:

[0078] 1. Data collection:

[0079] In different actual scenarios (such as crowded places like shopping malls, stations, schools, etc.), data is collected for each change area (such as corridor corners, stairwells, etc.).

[0080] Record geometric feature data such as the turning angle and width of each change area, and simulate a fire.

[0081] At the crowd intersection point, count the number of gathered people and calculate the average panic coefficient.

[0082] Through a combination of on-site monitoring equipment and manual statistics, determine the actual maximum number of people who can pass safely (the number of people that can be accommodated) in each change area under different conditions.

[0083] 2. Data preprocessing

[0084] Clean the collected data to remove outliers and incorrect data.

[0085] Normalize the data, mapping data such as the turning angle, width, average panic coefficient, and sum of population numbers to a suitable interval (such as [0, 1]) to improve the training efficiency and performance of the model.

[0086] 3. Model selection

[0087] Select a suitable machine learning model, such as a linear regression model, decision tree model, random forest model, neural network model (such as a multi-layer perceptron, etc.). Here, the neural network model is used as an example for illustration.

[0088] 4. Model construction

[0089] Construct a multi-layer perceptron (MLP) model. The number of neurons in the input layer corresponds to the number of input features (turning angle, width, mean panic coefficient, sum of the number of people, a total of 4 features).

[0090] The number of hidden layers and neurons can be adjusted according to the complexity of the data and the performance of the model. For example, 2 - 3 hidden layers can be set, and each hidden layer contains 32 - 128 neurons.

[0091] The output layer has one neuron, which is used to output the predicted number of people that can be tolerated.

[0092] 5. Model Training

[0093] Divide the preprocessed data into a training set and a test set, for example, according to a ratio of 70% - 30%.

[0094] Use the training set to train the model, and select appropriate loss functions (such as the mean squared error loss function, which is used to measure the gap between the predicted value and the true value) and optimizers (such as the Adam optimizer).

[0095] During the training process, update the parameters of the model through the backpropagation algorithm, with the goal of minimizing the loss function value. The training is completed until the model reaches the preset loss function value on the test set.

[0096] The bearing prediction model focuses on the key influencing factors in the changing area. Through rigorous data collection, preprocessing, and reasonable model construction and training, it organically combines fixed scenario factors such as turning angle and width with dynamic variables such as panic coefficient and number of people, and quantifies their comprehensive effects on the personnel passage ability. The number of people that can be tolerated output by the model provides an accurate quantitative basis for the diversion decision, ensuring that the diversion operation is targeted.

[0097] The panic coefficient can be characterized by the number of screams per unit time, or it can be comprehensively characterized by the number of screams per unit time and the mean heart rate variability of the corresponding crowded people per unit time. Specifically, it can be determined according to the hardware support situation of the fire escape path.

[0098] The number of screams per unit time can be collected by acoustic sensors installed on the fire escape path. Then, the property server marks the sounds exceeding the preset decibel as screams, determines the number of screams per unit time based on the screams within a unit time, and sends it to the second analysis module. The occurrence frequency of screams is also closely related to the crowded people density, moving speed, and the environment where they are located. The number of screams per unit time has the characteristic of immediacy. Using the number of screams per unit time to characterize the panic coefficient can directly reflect the panic level of the instantaneous crowded people and reduce the computational workload of the computing device.

[0099] A method for determining whether to divert a crowded group on an overlapping fire escape path with a crowd intersection point based on the number of people that can be accommodated includes:

[0100] If the sum of the number of people corresponding to the crowd intersection point exceeds the number of people that can be accommodated in the changing area, diversion is required; otherwise, it is not.

[0101] A third analysis module, if diversion is to be carried out, determines the number of people to be diverted based on the number of people that can be accommodated in the changing area.

[0102] Take the difference between the sum of the number of people in the crowded group corresponding to the crowd intersection point and the number of people that can be accommodated in the changing area as the number of people to be diverted planned in advance.

[0103] A second planning module selects the crowded group corresponding to another fire escape path that intersects with the fire escape path behind the crowd intersection point for diversion and re-plans the fire escape path. The re-planned fire escape path can be sent to the management personnel's receiving device, and the management personnel can guide the crowded group in advance. The number of people diverted is the number of people to be diverted.

[0104] For example, for crowd A and crowd B, there is a crowd intersection point on the same fire escape path. There is no intersection with other fire escape paths in the area between the position of crowd A and the crowd intersection point, and there is an intersection with other fire escape paths in the area between the position of crowd B and the crowd intersection point. Then, select to divert the number of people to be diverted from crowd B to other fire escape paths. Plan the diversion in advance to avoid overcrowding on a certain fire escape path caused by different crowds being too concentrated, reduce the occurrence of stampede incidents, and improve the escape efficiency.

[0105] In this solution, by inputting the number of people, positions, panic coefficients, and the average width of the fire escape paths corresponding to different crowded groups into the trained travel prediction model, potential crowd intersection points can be accurately predicted in advance for multiple future time units. This advance prediction in multiple time dimensions wins a valuable "golden disposal window" for fire emergency response. Compared with the traditional real-time monitoring mode, the warning time is greatly advanced, allowing the management personnel to have sufficient time to take countermeasures.

[0106] According to the turning angle, width, average panic coefficient of the changing area, and the sum of the number of people corresponding to the presence of crowd intersection points, the pre-trained bearing prediction model is used to accurately predict the number of people that the next changing area can bear. Among them, the panic coefficient can immediately reflect the panic state of the crowd while taking into account the calculation efficiency. The model training is based on the actual data of the changing area under a large number of different scenarios, fully considering the comprehensive influence of various factors on the personnel passing capacity, ensuring that the prediction results are highly consistent with the actual situation. On this basis, by comparing the number of people that can be borne with the actual number of people, it is accurately judged whether diversion is needed, realizing the scientificity and rationality of the diversion decision-making, and avoiding unnecessary diversion operations or dangerous situations caused by failure to divert in time.

[0107] Once it is determined that diversion is needed, this solution diverts the gathered crowd corresponding to other fire escape paths that intersect with the fire escape path behind the crowd intersection point by cleverly selecting them, and accurately allocates according to the calculated number of people to be diverted. At the same time, it re-plans the fire escape path and promptly sends it to the receiving device of the management personnel for early guidance of the gathered crowd. In this way, it effectively avoids different crowds being overly concentrated on a single fire escape path, balances the personnel flow of each path, minimizes the probability of congestion and trampling incidents, effectively guarantees the orderliness and safety of personnel during the fire escape process, and comprehensively improves the reliability and emergency response ability of the entire fire evacuation system.

[0108] The diversion plan closely revolves around the crowd intersection point and the topological structure of the fire escape path. With the goal of balancing the flow of people on each path, by reasonably selecting the diversion objects and accurately calculating the number of people to be diverted, and cooperating with timely path re-planning and information transmission, the management personnel can intervene and guide in advance, nip the congestion hidden danger in the bud, and ensure the smoothness and safety of the entire evacuation process.

[0109] Embodiment 2

[0110] Refer to Figure 3 , this embodiment provides a fire path planning method based on multi-objective optimization, including:

[0111] Determine the affected coverage area based on the coordinates of the fire occurrence point, determine the number of people based on the affected coverage area, determine the personnel gathering area based on the personnel coordinates, and distinguish the gathered crowd based on the personnel gathering area;

[0112] Based on the positions of the gathered crowd, use the path planning algorithm to plan the fire escape paths for each gathered crowd;

[0113] If there is an overlap in the fire escape paths corresponding to different gathered crowds, predict the crowd intersection points of different gathered crowds on the fire escape paths at future moments;

[0114] Predict the number of people that can be endured in the next changing area of the fire escape route based on the crowd intersection point, and determine whether to divert the crowded crowd on the overlapping fire escape routes where there are crowd intersection points based on the number of people that can be endured.

[0115] If diversion is to be carried out, determine the number of people to be diverted based on the number of people that can be endured in the changing area.

[0116] Select other fire escape routes that intersect with the fire escape route behind the crowd intersection point, divert the corresponding crowded crowd, and re-plan the fire escape route. The diversion quantity is the number of people to be diverted, and the direction behind is opposite to the advancing direction of the fire escape route.

[0117] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

[0118] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A firefighting path planning method based on multi-objective optimization, characterized in that: include: Determine the affected coverage area based on the coordinates of the fire occurrence point, determine the number of people based on the affected coverage area, determine the gathering area of ​​people based on the coordinates of the people, and distinguish the gathering people based on the gathering area of ​​people; Based on the location of the crowd, a path planning algorithm is used to plan the fire escape path for each crowd; If the fire escape paths corresponding to different groups of people overlap, predict the crowd intersection points of different groups of people on the fire escape paths at future moments; Predicting the number of people in the next change area of ​​the fire escape path based on the crowd intersection point, and determining whether to divert the crowd gathered on the overlapping fire escape path and with the crowd intersection point based on the number of people; If diversion is carried out, the number of diversions shall be determined based on the number of people in the changed area; Select other fire escape paths that intersect with the fire escape path behind the crowd intersection point, divert the corresponding gathered crowd, and re-plan the fire escape path. The diversion number is the number of diverted people, and the rear is in the opposite direction of the fire escape path.

2. The firefighting path planning method based on multi-objective optimization according to claim 1 is characterized in that: Methods for determining the affected coverage area based on the coordinates of the fire occurrence point include: The crowd determination module receives the coordinates of the fire occurrence point sent by the management server, takes the coordinates of the fire occurrence point as the origin, draws a circle with a preset impact radius, obtains the circular area, and uses the circular area as the impact coverage area; the coordinates of the fire occurrence point are determined by the management server based on the position sent by the fire information alarm device.

3. The firefighting path planning method based on multi-objective optimization according to claim 2 is characterized in that: Methods for determining the number of personnel based on the impact coverage area include: Receive the number of personnel determined in the affected coverage area sent by the management server; when the management server receives the fire information, it continuously sends statistical information to the communication server, the statistical information includes the identification of all signal repeaters in the affected coverage area; the communication server obtains the corresponding number of devices connected to all signal repeaters in the affected coverage area at the current moment according to the signal repeater identification of all signal repeaters in the affected coverage area, and sends the number of devices connected to each signal repeater to the management server; the management server uses the number of devices as the number of personnel determined in the affected coverage area.

4. The firefighting path planning method based on multi-objective optimization according to claim 3 is characterized in that: Methods for determining personnel gathering areas based on personnel coordinates include: The coordinates of the signal repeater installation location are used as the coordinates of the device connected to the same signal repeater, and the device coordinates correspond to the personnel coordinates; the personnel corresponding to the equipment connected to the same signal repeater are regarded as the same personnel gathering area.

5. The firefighting path planning method based on multi-objective optimization according to claim 4 is characterized in that: The location of the crowd corresponds to the installation location of the signal repeater. According to the predetermined fire exits in the coverage area, the A* algorithm is used to plan the fire escape route of each crowd based on the location of the crowd and the fire exits.

6. The firefighting path planning method based on multi-objective optimization according to claim 4 is characterized in that: The number of people, location, panic coefficient, and the average width of the fire escape path corresponding to different crowds are input into the trained travel prediction model to obtain the travel of the corresponding crowds at multiple future moments; The end points of the travel for the crowd gathered at the corresponding multiple future moments on the fire escape path are determined according to the travel paths corresponding to the multiple future moments, and the end points are marked on the fire escape path; if the distance between the end points at the same moment in the future is lower than a predetermined distance threshold, the end point in front is used as the crowd intersection point, and the front direction is consistent with the forward direction of the fire escape path.

7. The firefighting path planning method based on multi-objective optimization according to claim 6 is characterized in that: The method for predicting the number of people in the next change area of ​​the fire escape path based on the crowd intersection point includes: The direction of the next change area is consistent with the direction of the fire escape path; the change area is an area where the turning angle or width of the area is lower than the corresponding preset threshold; The sum of the turning angle, width, average panic coefficient, and number of people corresponding to the crowd intersection of the next change area is input into the pre-trained tolerance prediction model to obtain the number of people to be tolerated in the change area; the average panic coefficient is the average panic coefficient of the corresponding crowd gathered at the crowd intersection; The panic coefficient is the number of screams per unit time.

8. The firefighting path planning method based on multi-objective optimization according to claim 7 is characterized in that: Methods for determining whether to divert crowds on overlapping fire escape routes and at crowd intersections based on the number of people to be supported include: If the sum of the number of people corresponding to the crowd intersection points exceeds the number of people that the change area can bear, diversion is required, otherwise it is not necessary.

9. The firefighting path planning method based on multi-objective optimization according to claim 8 is characterized in that: Methods for determining the number of diversions based on the number of people in the changing area include: The difference between the sum of the number of people in the crowd corresponding to the crowd intersection point and the number of people in the change area is used as the number of people to be diverted in advance.

10. The fire path planning system based on multi-objective optimization is characterized by: A firefighting path planning method based on multi-objective optimization for implementing any one of claims 1 to 9, the system comprising: A crowd determination module determines the affected coverage area based on the coordinates of the fire occurrence point, determines the number of people based on the affected coverage area, determines the gathering area of ​​people based on the coordinates of the people, and distinguishes the gathering crowd based on the gathering area of ​​people; The first planning module uses a path planning algorithm to plan the fire escape path for each crowd based on the location of the crowd; The first analysis module predicts the crowd intersection points of different gathered groups on the fire escape paths at future moments if the fire escape paths corresponding to different gathered groups overlap; The second analysis module predicts the number of people in the next change area of ​​the fire escape path based on the crowd intersection point, and determines whether to divert the crowd gathered on the overlapping fire escape path and with the crowd intersection point based on the number of people; The third analysis module determines the number of diversions based on the number of personnel in the changed area if diversion is required; The second planning module selects other fire escape paths that intersect with the fire escape path behind the crowd intersection point, diverts the corresponding gathered crowd, and re-plans the fire escape path. The diversion number is the number of diverted people, and the rear is in the opposite direction of the fire escape path.

Images