Emergency rescue management method and system for first aid equipment

Abstract

The present invention discloses an emergency rescue management method and system for first aid equipment, belonging to the field of emergency rescue technology. The specific steps of the management method are as follows: S1, real-time collection and analysis of environmental data and image information; S2, evaluation and ranking of different rescue plans; S3, rescue experience learning and optimization of rescue plans; S4, evaluation of different decision paths and selection of the best action plan; the present invention can more objectively and comprehensively select rescue plans, improve the system's resilience and adaptability, improve the system's stability and reliability, improve the overall rescue efficiency, make emergency response more timely and effective, enable the system to maintain efficient operation in a complex and dynamic environment, improve the quality and reliability of real-time rescue actions, reduce dependence on human intervention, reduce errors and risks caused by human judgment, and improve the scientificity and objectivity of rescue decisions.

Description

Technical Field

[0001] The present invention relates to the field of emergency rescue technology, and in particular to an emergency rescue management method and system for first aid equipment. Background Art

[0002] Emergency rescue management involves complex decision-making and execution processes, which requires processing a large amount of information in a short period of time to make scientific and reasonable decisions. Traditional rescue methods rely on human judgment, which are inefficient and have large errors, and are difficult to cope with the complexity and urgency of emergencies. In recent years, with the development of technologies such as the Internet of Things, big data, and artificial intelligence, intelligent emergency rescue management systems have gradually become possible. At present, many emergency rescue systems have initially realized data collection and monitoring functions, but there is still much room for improvement in decision support and optimization. Especially in complex and changeable rescue environments, how to quickly and accurately evaluate rescue plans and select the optimal action path is a problem that needs to be solved urgently. Existing research and applications show that a single technical means is difficult to meet actual needs, and a variety of technologies must be used in combination to form a complete system solution.

[0003] The existing emergency rescue management methods and systems are unable to objectively and comprehensively select rescue plans, which reduces the system's resilience and adaptability and is unable to obtain emergency responses in a timely and effective manner. In addition, the quality and reliability of real-time rescue actions in the existing emergency rescue management methods and systems are low, and they rely heavily on human intervention, increasing the errors and risks caused by human judgment. To this end, we propose an emergency rescue management method and system for first aid equipment. Summary of the invention

[0004] The purpose of the present invention is to solve the defects in the prior art and to propose an emergency rescue management method and system for first aid equipment.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] An emergency rescue management method for first aid equipment, the specific steps of the management method are as follows:

[0007] S1. Collect and analyze environmental data and image information in real time.

[0008] As a further solution of the present invention, the specific steps of collecting and analyzing the environmental data and image information described in S1 are as follows:

[0009] Step 1: Survey the emergency scene, draw a floor plan, determine the best placement of sensors and cameras, allocate equipment locations based on environmental characteristics and rescue needs, cover all key rescue areas, and then calibrate each group of sensors and cameras. Collect environmental data in real time through each sensor, and collect image information in real time through the camera;

[0010] Step 2: Use OpenCV or PIL library to read the image file, convert the image to RGB color space, perform multi-scale scaling on the image, generate image pyramids of different sizes, set the sliding window size according to the expected size of the target object, and use 10 to 20% of the window size as the step size of the sliding window to determine the number of pixels that the sliding window moves each time;

[0011] Step 3: Starting from the upper left corner of each group of image data, move the window according to the step size. Each time it moves, extract the sub-image in the window to generate multiple groups of sub-images. Use the TensorFlow framework to load the trained deep learning model, input the sub-image into the deep learning model, obtain the classification results, and output the target category and confidence.

[0012] Step 4: Sort all detection frames by confidence, check the detection frames one by one, retain the detection frames whose overlap with the current detection frame is lower than the threshold, retain the detection frames with the highest confidence and no duplication, store the detection results of the coordinates, category and confidence of the detection frame in a list or dictionary, and output the detection results to a file or database;

[0013] Step 5: Draw a detection box on the original image and mark the category and confidence of the detection box. At the same time, display the recognition results through map annotations and text information, and provide real-time updates of the location and status of key targets on site.

[0014] S2. Evaluate and rank different rescue plans.

[0015] As a further solution of the present invention, the specific steps of rescue plan evaluation and sorting in S2 are as follows:

[0016] Step 1: Discuss with rescue experts and rescuers to collect potential decision criteria, and determine various rescue criteria according to the characteristics of the rescue mission. Determine the weight of each criterion through the hierarchical analysis method, and normalize the weight of each criterion so that its sum is 1;

[0017] Step 2: Use sensors and cameras to obtain real-time data from the rescue site and the location information of emergency equipment, investigate and analyze historical data, evaluate the effectiveness and safety of each rescue plan, and conduct rescue simulations to obtain performance data of rescue plans in different scenarios;

[0018] Step 3: Organize the collected performance data of each rescue plan on each decision criterion into a decision matrix, where rows represent different rescue plans and columns represent different decision criteria, and fill in the performance values ​​of specific rescue plans on the corresponding criteria into the matrix;

[0019] Step 4: Normalize the decision matrix through Z-score normalization, multiply each element of the normalized matrix by the weight of the corresponding criterion to generate a new matrix, and each element represents the weighted performance value of the corresponding rescue plan on a certain criterion;

[0020] Step 5: Sum up the data in each row of the weighted decision matrix and calculate the comprehensive score of each rescue plan. The higher the comprehensive score, the better the plan. Sort all rescue plans from high to low according to the comprehensive score, and select the plan with the highest comprehensive score as the recommended rescue plan. Record and save the sorting results, and verify the feasibility and effectiveness of the optimal plan through simulation and field exercises.

[0021] As a further solution of the present invention, the rescue criteria in step 1 specifically include response time, resource consumption, rescue effect, safety and coverage.

[0022] S3. Learn rescue experience and optimize rescue plans.

[0023] As a further solution of the present invention, the specific steps of optimizing the rescue solution described in S3 are as follows:

[0024] Step I: Collect the generated recommended rescue solutions, and randomly perturb the rescue solutions to generate multiple groups of rescue solutions. Then, each solution is used as an antibody, and then the affinity function is defined by comprehensively considering multiple evaluation criteria.

[0025] Step II: Calculate the affinity of each group of antibodies through the affinity function, record the affinity values ​​of all antibodies at the same time, perform proportional selection based on the affinity, select antibodies with affinity higher than the preset threshold for cloning, and then mutate each clone to generate new antibodies;

[0026] Step III: Calculate the affinity of the mutated clones, sort them by affinity, select multiple antibodies with affinity higher than the preset screening threshold to enter the new generation population, repeat the antibody iteration until the maximum number of iterations is reached or the change rate is lower than the set threshold, and output the antibody with the highest affinity in the current population as the optimal rescue solution.

[0027] S4. Evaluate different decision paths and choose the best course of action;

[0028] S5. Simulate the actual situation at the rescue site to verify the optimized rescue plan.

[0029] An emergency rescue management system for first aid equipment, comprising an acquisition processing module, a target detection module, a decision selection module, a solution optimization module, a simulation learning module, a simulation exploration module, a communication module, a user platform, a storage management module and a report generation module;

[0030] The acquisition and processing module is used to acquire environmental data and image information in real time and pre-process the data;

[0031] The target detection module is used to analyze the collected images and videos to identify the locations of injured persons, obstacles and emergency equipment;

[0032] The decision selection module is used to set multiple evaluation criteria and assign weights according to rescue needs, and evaluate and sort different rescue plans;

[0033] The solution optimization module is used to optimize the recommended rescue solution generated by the decision selection module;

[0034] The simulation learning module is used to create a virtual simulation environment, conduct online learning, and simulate and practice different rescue plans;

[0035] The simulation exploration module is used to evaluate different decision paths of rescue plans in real time during the actual rescue process and select the optimal action plan.

[0036] As a further solution of the present invention, the specific steps of selecting the optimal action plan of the simulation exploration module are as follows:

[0037] Step ①: Create a root node representing the current rescue status, initialize the root node, and add it to the tree structure as the starting point. Starting from the root node, select a child node with the highest upper confidence bound value for expansion according to the upper confidence bound selection strategy;

[0038] Step 2: Generate all possible child node states based on the state of the current node, add the newly generated child node to the tree structure, initialize its visit count and reward value, and then randomly select a child node state for simulation starting from the current node until the simulated rescue is completed, and record the accumulated reward value during the simulation process;

[0039] Step 3: Starting from the terminal state of the simulation, trace back to the root node along the path, update the number of visits and cumulative reward value of each node, and repeat the selection, expansion, simulation and backtracking update steps until the predetermined number of iterations or calculation time is reached;

[0040] Step ④: Among all the child nodes of the root node, select the child node with the most visits as the starting point of the optimal decision path, take the rescue action represented by this node as the current optimal decision, and dynamically adjust the current rescue plan.

[0041] The communication module is used for data transmission and communication between modules in the system, and for data interaction with external devices and systems;

[0042] The user platform is used to provide an operation and monitoring interface for users;

[0043] The storage management module is used to store all data generated by the system and provide data query, retrieval and management functions;

[0044] The report generation module is used to automatically generate a report according to the rescue process and results.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. The emergency rescue management method of the first aid equipment determines various rescue criteria according to the characteristics of the rescue mission, determines the weight of each criterion through the hierarchical analysis method, obtains data and the location information of the first aid equipment from the rescue site in real time through sensors and cameras, investigates and analyzes historical data, evaluates the effectiveness and safety of various rescue plans, and conducts rescue simulation at the same time to obtain the performance data of the rescue plans in different situations. The performance data of each rescue plan on each decision criterion collected are sorted into a decision matrix. After preprocessing the decision matrix, the comprehensive score of each rescue plan is calculated. The higher the comprehensive score, the better the plan. All rescue plans are sorted from high to low according to the comprehensive score, and the plan with the highest comprehensive score is selected as the recommended rescue plan. The generated recommended rescue plan is collected. It recommends a rescue plan, and randomly perturbs the rescue plan to generate multiple groups of rescue plans. It then takes each plan as an antibody, calculates the affinity of each group of antibodies, selects antibodies with affinity higher than a preset threshold for cloning, and then mutates each clone to generate a new antibody. It calculates the affinity of the mutated clones, and selects multiple antibodies with affinity higher than a preset screening threshold to enter the new generation population. It repeats the antibody iteration until the maximum number of iterations is reached or the change rate is lower than the set threshold, and outputs the antibody with the highest affinity in the current population as the optimal rescue plan. It can select rescue plans more objectively and comprehensively, improve the system's resilience and adaptability, improve the system's stability and reliability, improve the overall rescue efficiency, and make emergency response more timely and effective.

[0047] 2. The emergency rescue management system of the first aid equipment creates a root node representing the current rescue state, initializes the root node, and adds it to the tree structure as the starting point. Starting from the root node according to the upper confidence bound selection strategy, a child node with the highest upper confidence bound value is selected for expansion. According to the state of the current node, all possible child node states are generated, and the newly generated child node is added to the tree structure, and its access count and reward value are initialized. Starting from the current node, a child node state is randomly selected for simulation until the simulated rescue ends, and the cumulative reward value during the simulation is recorded. Starting from the terminal state of the simulation, the path is traced back to the root node, and the access count and cumulative reward value of each node are updated. The selection, expansion, simulation and backtracking update steps are repeated until the predetermined number of iterations or calculation time is reached. Among all the child nodes of the root node, the child node with the most access times is selected as the starting point of the optimal decision path, and the rescue action represented by the node is taken as the current optimal decision. At the same time, the current rescue plan is dynamically adjusted, which can enable the system to maintain efficient operation in a complex and dynamic environment, improve the quality and reliability of real-time rescue actions, reduce dependence on human intervention, reduce errors and risks caused by human judgment, and improve the scientificity and objectivity of rescue decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0049] Figure 1 A flowchart of an emergency rescue management method for first aid equipment proposed by the present invention;

[0050] Figure 2 This is a system block diagram of an emergency rescue management system for first aid equipment proposed by the present invention. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0052] Example 1

[0053] Reference Figure 1 This embodiment discloses an emergency rescue management method for first aid equipment, and the specific steps of the management method are as follows:

[0054] Collect and analyze environmental data and image information in real time.

[0055] Specifically, survey the emergency scene, draw a floor plan, determine the best layout of sensors and cameras, allocate equipment locations according to environmental characteristics and rescue needs, cover all key rescue areas, calibrate each group of sensors and cameras, collect environmental data in real time through each sensor, and collect image information in real time through the camera. Use OpenCV or PIL library to read image files, convert images to RGB color space, perform multi-scale scaling on images, generate image pyramids of different sizes, set the sliding window size according to the expected size of the target object, and use 10 to 20% of the window size as the step size of the sliding window to determine the number of pixels moved each time the sliding window moves. Start from the upper left corner of each group of image data and move the window according to the step size. Each time it moves, it extracts the subgraph within the window and generates multiple groups of subgraphs. It uses the TensorFlow framework to load the trained deep learning model, inputs the subgraph into the deep learning model, obtains the classification results, outputs the target category and confidence, sorts all detection frames according to the confidence, checks the detection frames in turn, retains the detection frames whose overlap with the current detection frame is lower than the threshold, retains the detection frames with the highest confidence and no duplication, stores the detection results of the coordinates, category and confidence of the detection frame in a list or dictionary, outputs the detection results to a file or database, draws the detection frame on the original image, and marks the category and confidence of the detection frame. It also displays the recognition results through map annotations and text information, and provides real-time updates of the location and status of key targets on site.

[0056] Evaluate and rank different rescue options.

[0057] Specifically, discuss with rescue experts and rescuers to collect potential decision criteria, and determine various rescue criteria according to the characteristics of the rescue mission. Determine the weight of each criterion through the hierarchical analysis method, and normalize the weight of each criterion so that its sum is 1. Use sensors and cameras to obtain real-time data and emergency equipment location information from the rescue site, investigate and analyze historical data, evaluate the effectiveness and safety of each rescue plan, and conduct rescue simulations to obtain performance data of rescue plans in different situations. Organize the performance data of each rescue plan on each decision criterion into a decision matrix, in which rows represent different rescue plans, columns represent different decision criteria, and The performance value of a specific rescue plan on the corresponding criterion is filled in the matrix. The decision matrix is ​​normalized through Z-score normalization. Each element of the normalized matrix is ​​multiplied by the weight of the corresponding criterion to generate a new matrix, and each element represents the weighted performance value of the corresponding rescue plan on a certain criterion. The data in each row of the weighted decision matrix is ​​summed up to calculate the comprehensive score of each rescue plan. The higher the comprehensive score, the better the plan. All rescue plans are sorted from high to low according to the comprehensive score. The plan with the highest comprehensive score is selected as the recommended rescue plan. The sorting results are recorded and saved, and the feasibility and effectiveness of the optimal plan are verified through simulation and field exercises.

[0058] In this embodiment, the rescue criteria specifically include response time, resource consumption, rescue effect, safety, and coverage.

[0059] Learn from rescue experience and optimize rescue plans.

[0060] Specifically, the generated recommended rescue plans are collected, and the rescue plans are randomly perturbed to generate multiple groups of rescue plans. Each plan is then used as an antibody. After that, an affinity function is defined by comprehensively considering multiple evaluation criteria. The affinity of each group of antibodies is calculated through the affinity function. At the same time, the affinity values ​​of all antibodies are recorded, and proportional selection is performed according to the affinity. Antibodies with an affinity higher than a preset threshold are selected for cloning, and each clone is mutated to generate new antibodies. The affinity of the mutated clones is calculated, and they are sorted by affinity. Multiple antibodies with an affinity higher than a preset screening threshold are selected to enter a new generation population, and the antibody iteration is repeated until the maximum number of iterations is reached or the rate of change is lower than the set threshold, and the antibody with the highest affinity in the current population is output as the optimal rescue plan.

[0061] Evaluate different decision paths and select the best course of action;

[0062] Simulate the actual situation at the rescue site to verify the optimized rescue plan.

[0063] Example 2

[0064] Reference Figure 2, this embodiment discloses an emergency rescue management system for first aid equipment, including an acquisition processing module, a target detection module, a decision selection module, a solution optimization module, a simulation learning module, a simulation exploration module, a communication module, a user platform, a storage management module and a report generation module;

[0065] The acquisition and processing module is used to collect environmental data and image information in real time and pre-process the data; the target detection module is used to analyze the collected images and videos and identify the locations of injured persons, obstacles and first aid equipment.

[0066] The decision-making selection module is used to set multiple evaluation criteria, assign weights, and evaluate and rank different rescue plans according to rescue needs; the plan optimization module is used to optimize the recommended rescue plan generated by the decision-making selection module.

[0067] The simulation learning module is used to create a virtual simulation environment, conduct online learning, and simulate different rescue plans; the simulation exploration module is used to evaluate different decision paths of rescue plans in real time during the actual rescue process and select the optimal action plan.

[0068] Specifically, a root node representing the current rescue state is created, the root node is initialized, and it is added to the tree structure as the starting point. Starting from the root node according to the upper confidence bound selection strategy, a child node with the highest upper confidence bound value is selected for expansion. According to the state of the current node, all possible child node states are generated, and the newly generated child nodes are added to the tree structure. At the same time, their visit times and reward values ​​are initialized. Starting from the current node, a child node state is randomly selected for simulation until the simulated rescue ends, and the accumulated reward value during the simulation is recorded. Starting from the terminal state of the simulation, the path is traced back to the root node, and the visit times and accumulated reward values ​​of each node are updated. The selection, expansion, simulation and backtracking update steps are repeated until the predetermined number of iterations or calculation time is reached. Among all the child nodes of the root node, the child node with the most visits is selected as the starting point of the optimal decision path, and the rescue action represented by the node is used as the current optimal decision. At the same time, the current rescue plan is dynamically adjusted.

[0069] The communication module is used for data transmission and communication between modules within the system, and for data interaction with external devices and systems; the user platform is used to provide users with an operation and monitoring interface; the storage management module is used to store all data generated by the system and provide data query, retrieval and management functions; the report generation module is used to automatically generate reports based on the rescue process and results.

Claims

1. An emergency rescue management method for first aid equipment, characterized in that: The specific steps of this management method are as follows: S1, real-time collection and analysis of environmental data and image information; S2. Evaluate and rank different rescue plans; S3. Learn rescue experience and optimize rescue plans; S4. Evaluate different decision paths and choose the best course of action; S5. Simulate the real situation of the rescue site to verify the optimized rescue plan; The specific steps of rescue plan evaluation and ranking described in S2 are as follows: Step 1: Discuss with rescue experts and rescuers to collect potential decision criteria, and determine various rescue criteria according to the characteristics of the rescue mission. Determine the weight of each criterion through the hierarchical analysis method, and normalize the weight of each criterion so that its sum is 1; Step 2: Use sensors and cameras to obtain real-time data from the rescue site and the location information of emergency equipment, investigate and analyze historical data, evaluate the effectiveness and safety of each rescue plan, and conduct rescue simulations to obtain performance data of rescue plans in different scenarios; Step 3: Organize the collected performance data of each rescue plan on each decision criterion into a decision matrix, where rows represent different rescue plans and columns represent different decision criteria, and fill in the performance values ​​of specific rescue plans on the corresponding criteria into the matrix; Step 4: Normalize the decision matrix through Z-score normalization, multiply each element of the normalized matrix by the weight of the corresponding criterion to generate a new matrix, and each element represents the weighted performance value of the corresponding rescue plan on a certain criterion; Step 5: Sum up the data of each row of the weighted decision matrix and calculate the comprehensive score of each rescue plan. The higher the comprehensive score, the better the plan. Sort all rescue plans from high to low according to the comprehensive score, select the plan with the highest comprehensive score as the recommended rescue plan, record and save the sorting results, and verify the feasibility and effectiveness of the optimal plan through simulation and field exercises; The specific steps for optimizing the rescue plan described in S3 are as follows: Step I: Collect the generated recommended rescue solutions, and randomly perturb the rescue solutions to generate multiple groups of rescue solutions. Then, each solution is used as an antibody, and then the affinity function is defined by comprehensively considering multiple evaluation criteria. Step II: Calculate the affinity of each group of antibodies through the affinity function, record the affinity values ​​of all antibodies at the same time, perform proportional selection based on the affinity, select antibodies with affinity higher than the preset threshold for cloning, and then mutate each clone to generate new antibodies; Step III: Calculate the affinity of the mutated clones, sort them by affinity, select multiple antibodies with affinity higher than the preset screening threshold to enter the new generation population, repeat the antibody iteration until the maximum number of iterations is reached or the change rate is lower than the set threshold, and output the antibody with the highest affinity in the current population as the optimal rescue solution; The specific steps of evaluating different decision paths and selecting the best action plan as described in S4 are as follows: Step ①: Create a root node representing the current rescue status, initialize the root node, and add it to the tree structure as the starting point. Starting from the root node, select a child node with the highest upper confidence bound value for expansion according to the upper confidence bound selection strategy; Step 2: Generate all possible child node states based on the state of the current node, add the newly generated child node to the tree structure, initialize its visit count and reward value, and then randomly select a child node state for simulation starting from the current node until the simulated rescue is completed, and record the accumulated reward value during the simulation process; Step 3: Starting from the terminal state of the simulation, trace back to the root node along the path, update the number of visits and cumulative reward value of each node, and repeat the selection, expansion, simulation and backtracking update steps until the predetermined number of iterations or calculation time is reached; Step ④: Among all the child nodes of the root node, select the child node with the most visits as the starting point of the optimal decision path, take the rescue action represented by this node as the current optimal decision, and dynamically adjust the current rescue plan.

2. The emergency rescue management method of first aid equipment according to claim 1, characterized in that: The specific steps for collecting and analyzing the environmental data and image information described in S1 are as follows: Step 1: Survey the emergency scene, draw a floor plan, determine the best placement of sensors and cameras, allocate equipment locations based on environmental characteristics and rescue needs, cover all key rescue areas, and then calibrate each group of sensors and cameras. Collect environmental data in real time through each sensor, and collect image information in real time through the camera; Step 2: Use OpenCV or PIL library to read the image file, convert the image to RGB color space, perform multi-scale scaling on the image, generate image pyramids of different sizes, set the sliding window size according to the expected size of the target object, and use 10 to 20% of the window size as the step size of the sliding window to determine the number of pixels that the sliding window moves each time; Step 3: Starting from the upper left corner of each group of image data, move the window according to the step size. Each time it moves, extract the sub-image in the window to generate multiple groups of sub-images. Use the TensorFlow framework to load the trained deep learning model, input the sub-image into the deep learning model, obtain the classification results, and output the target category and confidence. Step 4: Sort all detection frames by confidence, check the detection frames one by one, retain the detection frames whose overlap with the current detection frame is lower than the threshold, retain the detection frames with the highest confidence and no duplication, store the detection results of the coordinates, category and confidence of the detection frame in a list or dictionary, and output the detection results to a file or database; Step 5: Draw a detection box on the original image and mark the category and confidence of the detection box. At the same time, display the recognition results through map annotations and text information, and provide real-time updates of the location and status of key targets on site.

3. The emergency rescue management method of first aid equipment according to claim 1, characterized in that: The rescue criteria described in step 1 specifically include response time, resource consumption, rescue effect, safety, and coverage.

4. An emergency rescue management system for first aid equipment, used to implement the emergency rescue management method according to any one of claims 1 to 3, characterized in that: It includes acquisition and processing module, target detection module, decision selection module, solution optimization module, simulation learning module, simulation exploration module, communication module, user platform, storage management module and report generation module; The acquisition and processing module is used to acquire environmental data and image information in real time and pre-process the data; The target detection module is used to analyze the collected images and videos to identify the locations of injured persons, obstacles and emergency equipment; The decision selection module is used to set multiple evaluation criteria and assign weights according to rescue needs, and evaluate and sort different rescue plans; The solution optimization module is used to optimize the recommended rescue solution generated by the decision selection module; The simulation learning module is used to create a virtual simulation environment, conduct online learning, and simulate and practice different rescue plans; The simulation exploration module is used to evaluate different decision paths of rescue plans in real time during the actual rescue process and select the best action plan; The communication module is used for data transmission and communication between modules in the system, and for data interaction with external devices and systems; The user platform is used to provide an operation and monitoring interface for users; The storage management module is used to store all data generated by the system and provide data query, retrieval and management functions; The report generation module is used to automatically generate a report according to the rescue process and results.

5. The emergency rescue management system for first aid equipment according to claim 4, characterized in that: The specific steps of selecting the optimal action plan of the simulation exploration module are as follows: Step ①: Create a root node representing the current rescue status, initialize the root node, and add it to the tree structure as the starting point. Starting from the root node, select a child node with the highest upper confidence bound value for expansion according to the upper confidence bound selection strategy; Step 2: Generate all possible child node states based on the state of the current node, add the newly generated child node to the tree structure, initialize its visit count and reward value, and then randomly select a child node state for simulation starting from the current node until the simulated rescue is completed, and record the accumulated reward value during the simulation process; Step 3: Starting from the terminal state of the simulation, trace back to the root node along the path, update the number of visits and cumulative reward value of each node, and repeat the selection, expansion, simulation and backtracking update steps until the predetermined number of iterations or calculation time is reached; Step ④: Among all the child nodes of the root node, select the child node with the most visits as the starting point of the optimal decision path, take the rescue action represented by this node as the current optimal decision, and dynamically adjust the current rescue plan.

Images