Intelligent fire scene firefighting and rescue system and control method

Abstract

The application provides an intelligent fire rescue system and a control method. The system collects firefighter state data and temperature change information, uses gradient analysis reverse tracking technology to locate the fire source distribution, predicts the fire spread based on the fire source distribution data, constructs an evacuation path scheme from the dangerous area to the safe area, and formulates a safe operation range. The system breaks down the rescue target into a fragmented task set, dynamically recombines the task allocation scheme based on the firefighter state, establishes a two-way state confirmation channel between the rescuer and the rescued person, evaluates the rescue effect and analyzes the failure risk, optimizes the task combination, monitors the rescue process and identifies abnormal patterns, evaluates the safety according to the evacuation path scheme, adjusts the rescue target according to the safety state monitoring result, generates controllable risk parameters through risk conversion, formulates a rescue strategy, and realizes the intelligent and safe rescue of "first ensuring the ability to exit, and then maximizing the rescue effect".

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent fire rescue, in particular to an intelligent fire rescue system and a control method. BACKGROUND

[0002] With the acceleration of urbanization and the complexity of building structures, fire accidents show the characteristics of fast spreading speed, great difficulty in rescue, heavy casualties, etc., which puts higher requirements on the intelligent level and emergency response capability of fire rescue. Traditional fire rescue mainly relies on experience judgment and manual decision-making, and in the case of complex fire environment, incomplete information and time urgency, there are often problems such as decision lag, improper resource allocation and low rescue efficiency.

[0003] The existing fire rescue system has many limitations in technical implementation: lack of accurate identification and dynamic tracking ability of fire source position, leading to strong blindness of rescue action and inability to develop targeted rescue strategy; rescue task allocation mostly adopts static pre-plan or simple division, lacking intelligent dynamic adjustment mechanism based on personnel state and task complexity; safety evaluation and risk control means are single, lacking real-time monitoring and abnormal early warning ability of rescue process; rescue decision-making mostly depends on the personal experience of commanders, lacking scientific decision-making support based on data analysis and intelligent algorithm. These technical bottlenecks seriously restrict the rescue effect of fire rescue in complex places such as high-rise buildings, underground spaces and chemical enterprises, and it is urgent to develop a new generation of intelligent fire rescue system with intelligent perception, autonomous decision-making and dynamic optimization capability. SUMMARY

[0004] The present application provides an intelligent fire rescue system and a control method, aiming to realize intelligent rescue decision-making and precise rescue operation in complex fire environment, and to integrate core technologies such as intelligent fire source identification, safety path planning, task dynamic allocation, two-way state confirmation, real-time abnormal warning and risk adaptive control, to realize intelligent control of the whole fire rescue process, to realize accurate positioning of fire source position, autonomous allocation of rescue task, dynamic evaluation of safety risk and real-time optimization of rescue strategy, and to build an intelligent fire rescue system with environmental perception capability, intelligent decision-making capability, adaptive adjustment capability and collaborative operation capability.

[0005] The present application provides an intelligent fire rescue system and a control method, aiming to realize intelligent rescue decision-making and precise rescue operation in complex fire environment, and to integrate core technologies such as intelligent fire source identification, safety path planning, task dynamic allocation, two-way state confirmation, real-time abnormal warning and risk adaptive control, to realize intelligent control of the whole fire rescue process, to realize accurate positioning of fire source position, autonomous allocation of rescue task, dynamic evaluation of safety risk and real-time optimization of rescue strategy, and to build an intelligent fire rescue system with environmental perception capability, intelligent decision-making capability, adaptive adjustment capability and collaborative operation capability.

[0006] Collecting firefighter position information and vital sign data to generate firefighter state data, obtaining temperature change information around the firefighter, performing gradient analysis on the temperature change information to track the fire source position reversely and generating fire source distribution data;

[0007] generate a dangerous area distribution based on the fire source distribution data, identify a boundary area of the dangerous area distribution to generate a relatively safe area, construct an evacuation path scheme from the dangerous area distribution to the relatively safe area, and develop a rescue action boundary to generate a safe operation range according to the evacuation path scheme;

[0008] obtain rescue target information, task decompose the rescue target information to generate a fragmented rescue task set, dynamically recombine the fragmented rescue task set based on the firefighter state data to generate a personalized task allocation scheme;

[0009] execute rescue operations in the safe operation range according to the personalized task allocation scheme to generate rescue execution records, establish a two-way state confirmation channel between rescuers and the rescued based on the rescue execution records, and obtain two-way confirmation data through the two-way state confirmation channel;

[0010] based on the two-way confirmation data, analyze the effectiveness of the rescue operation to obtain rescue effect evaluation data, analyze the failure risk of the rescue effect evaluation data to identify potential failure factors, and adjust the fragmented rescue task set according to the potential failure factors to generate an optimized task combination;

[0011] monitor the optimized task combination to obtain rescue process data, identify abnormal patterns from the rescue process data to generate an abnormal early warning signal, and perform safety evaluation based on the evacuation path scheme combined with the abnormal early warning signal to generate a safety state monitoring result;

[0012] dynamically adjust the current rescue target according to the safety state monitoring result to generate a redefined rescue standard, convert the risk dimension of the redefined rescue standard to generate controllable risk parameters, develop a rescue strategy based on the controllable risk parameters to generate a final control instruction, and complete intelligent fire field fire rescue control.

[0013] The second aspect of the present application proposes an intelligent fire field fire rescue system, comprising:

[0014] The data acquisition module is used for collecting firefighter position information and vital sign data to generate firefighter state data, obtaining temperature change information around the firefighter, and performing gradient analysis on the temperature change information to track the fire source position in reverse to generate fire source distribution data;

[0015] The path planning module is used for generating a dangerous area distribution based on the fire source distribution data, identifying a boundary area of the dangerous area distribution to generate a relatively safe area, constructing an evacuation path scheme from the dangerous area distribution to the relatively safe area, and developing a rescue action boundary to generate a safe operation range according to the evacuation path scheme;

[0016] a task allocation module configured to obtain rescue target information, generate a fragmented rescue task set by task decomposition on the rescue target information, and generate a personalized task allocation scheme by dynamic reorganization of the fragmented rescue task set based on the firefighter state data;

[0017] a rescue execution module configured to execute rescue operations within the safe operation range according to the personalized task allocation scheme to generate rescue execution records, establish a two-way state confirmation channel between rescuers and rescued persons based on the rescue execution records, and obtain two-way confirmation data through the two-way state confirmation channel;

[0018] an effect evaluation module configured to analyze the effectiveness of rescue operations based on the two-way confirmation data to obtain rescue effect evaluation data, identify potential failure factors by failure risk analysis on the rescue effect evaluation data, and generate an optimized task combination by adjusting the fragmented rescue task set according to the potential failure factors;

[0019] a safety monitoring module configured to execute monitoring on the optimized task combination to obtain rescue process data, generate an abnormality early warning signal by abnormal pattern recognition on the rescue process data, and generate a safety state monitoring result by safety evaluation based on the evacuation path scheme combined with the abnormality early warning signal;

[0020] a strategy control module configured to dynamically adjust a current rescue target according to the safety state monitoring result to generate a redefined rescue standard, generate controllable risk parameters by risk dimension conversion on the redefined rescue standard, formulate a rescue strategy based on the controllable risk parameters to generate a final control instruction, and complete intelligent fire field fire rescue control.

[0021] The beneficial effects of the present application are embodied in the following aspects: first, by temperature gradient analysis reverse tracking fire source positioning technology, the fundamental change from "blind search" to "accurate positioning" is realized, combined with the real-time monitoring of the multi-dimensional state of firefighters, the key technical bottleneck of unknown fire location and insufficient personnel state in traditional rescue is broken through, the pertinence and scientificity of rescue operation are significantly improved, and a solid foundation is laid for formulating accurate rescue strategy. Secondly, the fragmented task dynamic allocation and bidirectional state confirmation mechanism are innovatively established, realizing a major breakthrough from "extensive people tactics" to "fine collaborative operation", through intelligent matching of tasks and personnel capabilities and real-time bidirectional feedback of the rescue process, the core problem of unreasonable task allocation and opaque execution state in traditional rescue is fundamentally solved, and the rescue efficiency and success rate are effectively improved. Finally, a safety guarantee system based on abnormal early warning and risk self-adaptive control is built, realizing the rescue safety mode of "active safety protection" instead of "passive experience dependence", through the closed-loop control mechanism of real-time monitoring, dynamic early warning and intelligent adjustment, the technical difficulties of single safety guarantee means and insufficient risk control capability in traditional rescue are fundamentally solved, ensuring the maximization of rescue effect under the premise of "rescuer safety first", and forming an intelligent fire rescue technology system with self-decision and self-adaptive adjustment capability.

[0022] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0023] The drawings herein show specific examples of the technical solutions described in the present application, and constitute part of the specification together with the specific embodiments, for explaining the technical solutions, principles and effects of the present application.

[0024] Unless specifically stated or defined otherwise, the same reference signs in different drawings represent the same or similar technical features, and different reference signs may also be used to represent the same or similar technical features.

[0025] Figure 1 is a flowchart of an intelligent fire field fire rescue control method of the present application.

[0026] Figure 2 is a structural block diagram of an intelligent fire field fire rescue system of the present application. DETAILED DESCRIPTION

[0027] In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular architectures, technologies, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

[0028] It is to be understood that the terminology “includes”, “has”, “holds”, “comprises”, “containing”, “having” or “including” when used in this specification and / or the claims indicates an inclusion of one or more features, integers, steps, operations, elements, and / or components but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0029] It is also to be understood that the terminology “and / or” when used in this specification and / or the claims indicates an inclusion of one or more of the associated listed items and all possible combinations of those items.

[0030] As used in this specification and claims, the terms “if’ and “when” can be interpreted to mean “upon” or “in response to a determination” or “in response to a detection” depending on the context. Similarly, the phrase “if it is determined” or “if [a described condition or event] is detected” can be interpreted to mean “upon determining” or “in response to a determination” or “upon detecting [the described condition or event]” or “in response to a detection [of the described condition or event]” depending on the context.

[0031] In addition, the terms “first”, “second”, “third”, etc. in the description of the present application and the appended claims are only used to distinguish descriptions, and cannot be understood as indicating or implying relative importance.

[0032] Reference in the specification to “one embodiment” or “some embodiments” etc. means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application. The appearances of the phrases “in one embodiment”, “in some embodiments”, “in other embodiments”, “in additional embodiments”, etc. in various places in the specification are not necessarily all referring to the same embodiment, although they can. The terms “comprising”, “including”, “having” and the like are meant to be interpreted as “including but not limited to”, unless otherwise specifically noted.

[0033] The technical solutions of the embodiments of the present application are introduced as follows.

[0034] As Figure 1As shown, the embodiment of the present application provides a kind of intelligent fire field fire fighting rescue control method, comprising the following steps S110-step S170:

[0035] Step S110, collection firefighter position information and vital sign data generation firefighter state data, obtain the temperature variation information around firefighter, gradient analysis is carried out to temperature variation information, and fire source distribution data is generated by reverse tracking fire source position.

[0036] Specifically, collection firefighter position information and vital sign data generation firefighter state data.Firefighter position information collection uses Beidou satellite positioning and inertial navigation unit fusion positioning technology, provides meter-level positioning accuracy in outdoor environment, cooperates with UWB ultra-wideband indoor positioning to realize accurate positioning in building interior.Positioning equipment continuously updates the three-dimensional coordinate position, movement speed vector and orientation angle of firefighter, and through unified time base, data synchronism is ensured.Vital sign data collection relies on physiological monitoring sensor array integrated in protective clothing, heart rate monitoring uses photoelectric plethysmogram sensor, body temperature monitoring is realized through core temperature sensor and skin temperature sensor cooperation, blood oxygen saturation monitoring uses dual-wavelength pulse oximetry technology, respiratory rate is obtained by chest impedance change detection, and motion state detection is realized by three-axis accelerometer and gyroscope combination.Data fusion processing uses multi-source fusion algorithm to carry out space-time alignment, physiological data fluctuation is smoothed by Kalman filtering algorithm, and the fatigue index, stress level and health status rating of firefighter are calculated.For example, in the fire rescue of certain high-rise office building, when firefighter Zhang enters the twelve floor area full of smoke, monitoring data shows that its heart rate rises from 78 times per minute when calm to 142 times, core temperature rises from 36.6 degrees to 38.1 degrees, and blood oxygen saturation drops from 98% to 91%, and data fusion algorithm calculates its fatigue index to reach high level.After multidimensional data collection and intelligent fusion analysis, firefighter state data is formed.

[0037] The temperature change information of the firefighters is acquired. The temperature change information acquisition is based on a dual monitoring architecture combining a distributed sensor network and a mobile thermal imaging device. The distributed temperature sensor network uses industrial-grade digital temperature sensors and is deployed in key locations of the building according to a grid layout, including stairwells, main evacuation routes, equipment rooms and other important areas, forming a temperature monitoring network covering the building space. The mobile temperature acquisition is realized by a portable infrared thermal imaging device carried by the firefighters. The device is installed on the side of the firefighter's helmet or on the chest protector, and real-time two-dimensional temperature distribution images within the field of view are collected. The thermal imaging data is transmitted in real time to the data processing equipment of the on-site command vehicle through a wireless transmission link. The received temperature data is comprehensively preprocessed, including time stamp synchronization correction, GPS coordinate unification, data integrity inspection and other key steps. The data standardization processing converts the temperature data in different formats from different sensors into a unified standard format. For example, in a fire scene of a large shopping mall, the fixed sensor network provides real-time temperature data of key points on each floor, and the thermal imaging devices carried by five firefighters collect a large amount of temperature image information from different angles. After processing, a temperature distribution data set covering the entire area of the shopping mall is constructed. After multi-link processing and quality control, the temperature change information is obtained.

[0038] In some embodiments, the gradient analysis of the temperature change information reversely traces the fire source position to generate fire source distribution data, including: performing multi-point temperature gradient analysis on the temperature change information to obtain a gradient vector field; constructing a temperature change trend map based on the gradient vector field; generating a fire source candidate position by reverse path tracking using the temperature change trend map; and generating fire source distribution data according to the fire source candidate position.

[0039] First, multi-point temperature gradient analysis is performed on the temperature change information to obtain a gradient vector field. Temperature gradient analysis uses numerical differentiation method to determine the rate of change of temperature field in each direction by calculating the temperature difference between adjacent measurement points and the corresponding spatial distance. In a two-dimensional plane, the temperature partial derivatives in x and y directions are calculated to form a plane gradient vector; in a three-dimensional space, the temperature gradient in z direction is further calculated to construct a complete three-dimensional gradient vector field. First, multi-point temperature gradient analysis is performed on the temperature change information to obtain a gradient vector field. Temperature gradient analysis uses numerical differentiation method to determine the rate of change of temperature field in each direction by calculating the temperature difference between adjacent measurement points and the corresponding spatial distance. The mathematical expression of the temperature gradient vector is: Gradient strength calculation formula: where T(x, y, z) represents the temperature value at the spatial position (x, y, z), respectively represent the partial derivative of temperature in x, y, z direction, represents the strength of temperature gradient. In two-dimensional plane, the temperature partial derivatives in x and y direction are calculated to form the planar gradient vector; in three-dimensional space, the temperature gradient in z direction is further calculated to construct the complete three-dimensional gradient vector field. Central difference scheme is adopted in numerical calculation to improve the calculation accuracy, and data quality issues are handled in the process of constructing the gradient vector field, including identifying and removing abnormal measurement values, and handling data missing caused by sensor failure. Spatial interpolation algorithm is used to fill in the data missing area, and low-pass filtering technique is used to smooth the gradient field. Finally, the spatial gridded gradient data structure is established, and the size and direction information of the gradient vector are stored for each grid node. For example, in a fire accident in a chemical plant, the gradient analysis identifies a significant high temperature gradient center in the southeast corner of the reaction workshop, and the temperature gradient vector in this area presents a typical radial distribution pattern. Through numerical calculation and data processing, the gradient vector field is established.

[0040] Then, the temperature change trend map is constructed based on the gradient vector field. The visualization method combining isotheral line drawing technique and temperature distribution cloud map is adopted to convert the discrete distributed temperature measurement data and gradient information into continuous spatial distribution image. The isotheral line generation adopts the contour tracing algorithm, and reasonable temperature interval is set according to the temperature range to generate a complete isotheral line family from ambient temperature to high temperature area in fire field. The temperature distribution cloud map adopts color mapping technology to map temperature values to visually friendly color distribution, usually using a gradual color spectrum from blue to red. The gradient vector information is integrated in the trend map, and the direction and intensity of temperature change are displayed through arrow symbol, and the length of arrow is proportional to the gradient intensity. The time series temperature change trend map is constructed to show the spatio-temporal evolution process of temperature field. For example, in a fire rescue in an underground parking garage, the temperature change trend map shows that the high temperature isotheral line starts from a parking space in area B and gradually spreads outward in an elliptical pattern, and the diffusion direction is along the ventilation duct and lane layout of the parking garage. The temperature change trend map is constructed by using visualization technology and data fusion method.

[0041] Then, the temperature change trend map is constructed based on the gradient vector field. The visualization method combining isotheral line drawing technique and temperature distribution cloud map is adopted to convert the discrete distributed temperature measurement data and gradient information into continuous spatial distribution image. The isotheral line generation adopts the contour tracing algorithm, and reasonable temperature interval is set according to the temperature range to generate a complete isotheral line family from ambient temperature to high temperature area in fire field. The temperature distribution cloud map adopts color mapping technology to map temperature values to visually friendly color distribution, usually using a gradual color spectrum from blue to red. The gradient vector information is integrated in the trend map, and the direction and intensity of temperature change are displayed through arrow symbol, and the length of arrow is proportional to the gradient intensity. The time series temperature change trend map is constructed to show the spatio-temporal evolution process of temperature field. For example, in a fire rescue in an underground parking garage, the temperature change trend map shows that the high temperature isotheral line starts from a parking space in area B and gradually spreads outward in an elliptical pattern, and the diffusion direction is along the ventilation duct and lane layout of the parking garage. The temperature change trend map is constructed by using visualization technology and data fusion method. Path integral equation: where x(t) represents the path position at time t, x0is the starting point coordinate, the negative sign represents moving in the opposite direction of the gradient, and the integration process starts from the starting point x0and traces towards the fire source. The path tracking uses the fourth-order Runge-Kutta numerical integration method, and the algorithm uses a multi-start tracking strategy to simultaneously start the reverse tracking process from multiple high-temperature regions identified in the trend map. Various termination conditions are set during the tracking process, including when the gradient intensity falls below a pre-set threshold and when the tracking distance exceeds a reasonable range. The path aggregation analysis uses a density clustering algorithm to identify and analyze the spatial convergence patterns of multiple tracking paths, and determines the credibility of the fire source candidate location based on the density and concentration of path convergence. The algorithm also considers the local features of the gradient field to comprehensively evaluate and rank the candidate locations. For example, in a fire in a high-rise residential building, 15 reverse tracking paths were started from the multiple high-temperature regions identified at the 12th floor. Among them, 11 paths clearly converged to the kitchen area of room 1202, 3 paths converged to the vicinity of the distribution box in the corridor, and 1 path converged to the stairwell location. After numerical calculation and clustering analysis, the fire source candidate location was determined.

[0042] Finally, fire source distribution data is generated based on the fire source candidate location. Based on the fire source candidate location and its credibility evaluation results, a probability distribution modeling method is used to quantitatively process the candidate location. The path convergence density, convergence path number, and comprehensive evaluation ranking results of the candidate location are used to establish a probability weight allocation model for the candidate location. The probability distribution modeling extends the discrete candidate location to a continuous spatial probability field, and establishes a probability gradient distribution around the candidate location through spatial interpolation and probability diffusion techniques. Bayesian fusion method is used to integrate the probability contributions of multiple candidate locations to generate a unified fire probability distribution map. The fire distribution data is stored in a grid form, and each grid cell records the fire existence probability and confidence information. For example, in a fire case in a high-rise residential building, the kitchen area of room 1202 is assigned the highest fire probability due to the convergence of 11 paths, the corridor distribution box area obtains a medium probability due to the convergence of 3 paths, and the stairwell location is assigned a lower probability due to the convergence of only 1 path. Based on the credibility of the candidate location and the probability modeling, fire distribution data is generated.

[0043] Step S120, based on the fire source distribution data, generate the fire spread prediction to generate the dangerous area distribution, identify the boundary area of the dangerous area distribution to generate the relatively safe area, construct the evacuation path scheme from the dangerous area distribution to the relatively safe area, and develop the rescue action boundary to generate the safe operation range according to the evacuation path scheme.

[0044] In particular, the fire spread prediction based on the fire source distribution data generates the dangerous area distribution. The fire spread prediction adopts a prediction method based on a physical combustion model and numerical simulation, uses the fire source distribution data as the initial condition, combines building structure information, combustible material distribution, and environmental conditions to perform the space-time prediction of fire development. The combustion model considers three main ways of flame propagation: heat conduction, heat convection, and heat radiation, and describes the space-time evolution process of temperature field, velocity field, and concentration field through a system of partial differential equations. The numerical simulation solves the combustion equation set by using the finite element method, and the mesh is adaptively adjusted according to the building geometry. In the prediction process, the effects of the burning characteristics of building materials, the influence of ventilation conditions on the speed of fire spread, and the effects of fire extinguishing measures such as fire sprinkler devices are considered. The prediction results are represented in the form of temperature distribution, smoke concentration distribution, and flame front position, and the prediction area is divided into different danger levels by setting a danger threshold. For example, in the fire prediction of an office building, based on the fire source distribution data of the conference room on the fifteenth floor, the prediction model shows that the fire will spread rapidly along the open office area, and it is predicted that the flame front will reach the east side of the floor within 15 minutes, and the high-temperature area will cover two-thirds of the area of the entire floor within 30 minutes. Through physical modeling and numerical simulation calculation, the dangerous area distribution is generated.

[0045] The boundary area of the dangerous area distribution is identified to generate the relatively safe area. The contour tracing technology is adopted to determine the outer boundary of the dangerous area by setting a safety threshold, and the threshold selection considers the personnel safety standards and rescue operation requirements. The generation of the relatively safe area is based on the buffer analysis of the dangerous area boundary, and a safety buffer distance is set on the periphery of the dangerous area, and the buffer distance is calculated according to the speed of fire spread and the speed of personnel movement. The safety area evaluation considers the spatial distance from the dangerous source, the ventilation condition, the structural stability, the coverage of fire fighting facilities, and the accessibility of evacuation passages. The evaluation algorithm calculates a comprehensive safety index for each spatial position, and preferentially selects areas with multiple evacuation exits, solid structures, and good ventilation as the relatively safe area. For example, in the fire scene of a shopping center, the dangerous area distribution shows that the clothing area and the electrical appliance area on the first floor east side are high-risk areas, the boundary identification algorithm extracts the irregular boundary line of the dangerous area, and identifies that the atrium area on the first floor west side, the north side of the dining area on the second floor, and part of the underground parking area are relatively safe areas. Based on the boundary identification and safety evaluation analysis, the relatively safe area is generated.

[0046] In some embodiments, the construction of the evacuation path scheme from the dangerous area distribution to the relatively safe area includes: boundary identification of the dangerous area distribution to obtain dangerous boundary coordinates; determining the shortest evacuation distance based on the dangerous boundary coordinates and the relatively safe area; designing multiple candidate evacuation paths according to the shortest evacuation distance; and safety evaluation of the multiple candidate evacuation paths to generate an evacuation path scheme.

[0047] The boundary recognition of the dangerous area distribution obtains the dangerous boundary coordinates. The boundary extraction adopts the contour tracing technique and the contour detection algorithm, and determines the area boundary by setting the dangerous threshold. The algorithm can process complex irregular boundary shapes and multiple separated dangerous areas. The coordinate extraction process discretizes the boundary line into a series of key control points, and each boundary point records the accurate two-dimensional or three-dimensional spatial coordinate information. The coordinate system adopts a unified coordinate system inside the building to ensure compatibility with other spatial data. The boundary coordinate data structure contains attribute information such as boundary point sequence, boundary segment direction, boundary type, and dangerous level. The algorithm also performs boundary smoothing to eliminate boundary jaggedness caused by data noise. For example, in a fire in a certain chemical plant, the dangerous area boundary recognition algorithm extracts the complex polygon boundary of the reaction workshop, which contains several main boundary control points, and the boundary coordinates are accurate to the centimeter level, forming a complete dangerous boundary contour description. After boundary detection and coordinate extraction, the dangerous boundary coordinate data is obtained.

[0048] The shortest evacuation distance is determined based on the dangerous boundary coordinates and the relatively safe area. A topological network model of the building is established, with corridors, rooms, staircases, etc. as network nodes, and doors, passages, stairs, etc. as network edges. Each edge is assigned weights such as length, travel time, and safety factor. The distance calculation uses a multi-source shortest path algorithm, with the dangerous boundary coordinates as the starting point set and the relatively safe area as the target point set, to find the shortest path from all starting points to target points. The algorithm considers the actual travel conditions inside the building, including constraints such as passage width, staircase capacity, door opening status, etc. The calculation result generates a distance field distribution map, showing the evacuation distance and optimal direction from each location in the dangerous area to the nearest safe area. The distance data is stored in matrix form, recording the shortest distance, path length, and estimated travel time between boundary points and safe areas. For example, in a fire in a certain shopping center, the shortest evacuation distance from the key control points of the dangerous boundary on the second floor east side to the atrium safe area on the first floor west side is distributed between 65-95 meters, with an average evacuation distance of 78 meters and an estimated travel time of 4-6 minutes. Through network modeling and path calculation, the shortest evacuation distance distribution is determined.

[0049] Multiple alternative evacuation paths are designed according to the shortest evacuation distance. First, multiple local optimal path corridors in the distance field are identified, which represent different evacuation directions and path choices. Path generation adopts the K-shortest path algorithm and path diversity optimization techniques to maximize the spatial separation degree between paths while ensuring reasonable path length. The design process considers path load balancing, and path allocation is performed based on expected traffic volume and passage capacity to avoid the concentration of all personnel using a single path. The algorithm also considers path fault tolerance to ensure that other alternative paths are still available when a main path is blocked. Each designed path contains detailed path node sequences, segment distances, cumulative lengths, estimated travel times, and recommended traffic capacities, etc. Path representation uses a directed path graph to clearly identify evacuation directions and key decision points. For example, in a fire scenario in an office building, based on shortest evacuation distance analysis, three main alternative evacuation paths are designed from the eighth floor dangerous area: the east side fire staircase path (total length 52 meters), the west side evacuation staircase path (total length 48 meters), and the central staircase path (total length 65 meters), each with good spatial separation and effective dispersion of traffic. Based on distance optimization and diversity principles, multiple alternative evacuation paths are designed.

[0050] The safety of multiple alternative evacuation paths is evaluated to generate an evacuation path scheme. Evaluation indicators include path safety, travel efficiency, structural reliability, environmental adaptability, and emergency availability, etc. Path safety evaluation analyzes the fire risk level of the area, the influence of smoke diffusion, the stability of the structure, and the coverage of fire facilities. Travel efficiency evaluation includes path total length, expected evacuation time, passage bottleneck identification, and traffic carrying capacity analysis. The evaluation algorithm uses a weighted scoring method and a risk assessment matrix to calculate the comprehensive safety index and priority ranking for each path. The evaluation process also considers the time factor of fire spread to analyze the availability and safety changes of each path at different time periods. A hierarchical configuration scheme is finally formed, including main paths, alternative paths, and emergency paths, with clear usage conditions, applicable scenarios, and switching standards for each path. The evacuation path scheme also includes detailed traffic allocation strategies, path capacity control, and dynamic scheduling mechanisms. For example, in a hotel fire, the safety evaluation result shows that the east side fire staircase path is the priority path due to its independent smoke control system and shorter evacuation distance, the west side evacuation staircase path is the main alternative path, and the central staircase path is set as an emergency backup path due to possible smoke impact. After completing safety evaluation and path configuration, an evacuation path scheme is formed.

[0051] In some embodiments, the generating a safe operation range according to the evacuation path scheme includes: obtaining time safety parameters based on evacuation time requirements analysis of the evacuation path scheme; determining a rescue action boundary according to the time safety parameters; and generating a safe operation range based on the rescue action boundary.

[0052] First, time safety parameters are obtained based on evacuation path scheme analysis. Time safety analysis utilizes the aforementioned generated evacuation path scheme to conduct detailed time calculation, and determines the time constraints of rescue operations by analyzing the evacuation time requirements of different areas of personnel. Time calculation considers the influence of personnel density distribution, moving speed characteristics, path capacity and bottleneck effect on evacuation efficiency. Personnel evacuation time is calculated using crowd evacuation dynamics model, considering the moving characteristics of different crowds and behavior patterns in emergency situations. Fire spread time prediction is based on the aforementioned dangerous area distribution results to determine the critical time point when each area becomes impassable. Safety margin setting ensures that rescue personnel have sufficient evacuation time, usually set as a certain proportion of the minimum evacuation time. Time safety parameters include key indicators such as regional evacuation time, path effective time window, rescue task execution time and safety evacuation reserved time. For example, in a factory fire, it takes 12 minutes for the personnel in a workshop to evacuate completely, and the safe and effective time of the main evacuation path is 18 minutes. Considering the safety margin, the maximum allowed time for rescue operations is determined to be 8 minutes. After detailed time analysis and safety margin calculation, key time safety parameters are obtained.

[0053] Then, rescue action boundaries are determined according to time safety parameters. Rescue action boundaries are determined based on the aforementioned time safety parameters, and the safe time window and spatial range of rescue operations are determined by spatiotemporal analysis method. Time boundary specifies the safe operation time limit of each area, and spatial boundary determines the range of operation area that can be safely entered within a given time. Boundary determination uses dynamic analysis method, and the action boundary needs to be adjusted and updated in real time as the fire develops and time passes. Rescue action boundaries are represented in the form of space-time diagram, which clearly describes the trend of safe operation area and boundary evolution with time. Boundary determination considers the time requirements and risk characteristics of different rescue tasks, and sets corresponding action boundaries for personnel search and rescue, equipment rescue, fire reconnaissance and other task types. Boundary information includes accurate time nodes, spatial coordinates and safety level identification. For example, in a residential building fire, based on an 8-minute safe operation time window, the rescue action boundaries including the first floor lobby, the second floor west side and the third floor north side are determined, and the boundaries gradually shrink over time. Based on the above spatiotemporal analysis and dynamic modeling, the safety boundaries of rescue action are determined.

[0054] Finally, the safe operation range is generated based on the rescue action boundary. The generation of the safe operation range utilizes the previously determined rescue action boundary to develop detailed rescue operation area planning and operation guidance schemes. The safe operation range contains complete information such as the space area that can be safely entered, the recommended operation time, the entry path, the evacuation path, and the emergency plan. The operation range planning considers the priority and resource requirements of different rescue tasks, prioritizing the rescue operation of high-value targets and easy-to-rescue objects. Safety monitoring points and communication relay points are set within the range to ensure real-time communication between rescue personnel and the command center. The safe operation range adopts a hierarchical management strategy, dividing the operation area into core operation zones, buffer operation zones, and emergency operation zones according to safety levels. The operation guidance includes specific task allocation, personnel configuration, equipment requirements, and safety protection measures. The range planning also includes a dynamic adjustment mechanism to update the operation range and safety measures in a timely manner according to changes in the field situation. For example, in a fire rescue at a chemical plant, the safe operation range includes the first half of the reaction workshop, the main corridor, and the adjacent office area, with two entry paths and three evacuation paths, and five safety monitoring points and three communication relay points. The safe operation range is generated. Combining boundary constraints and actual task requirements, a complete safe operation range is finally established.

[0055] In step S130, the rescue target information is obtained, the rescue target information is task-decomposed to generate a fragmented rescue task set, and the fragmented rescue task set is dynamically reorganized based on the firefighter state data to generate a personalized task allocation scheme.

[0056] Specifically, the rescue target information is obtained. The on-site reconnaissance information is obtained through video equipment, thermal imaging instruments, and acoustic detectors carried by firefighters, and real-time collection of the location, number, state, and surrounding environment information of trapped personnel. Building information includes floor structure diagrams, room layout, personnel capacity, and functional use, etc. static data, obtained through building information modeling and fire archives. Personnel registration information is obtained from access control records, attendance clocking, visitor registration, and security monitoring channels, providing basic data on personnel distribution. Real-time monitoring information is obtained through smoke detectors, emergency call devices, mobile signal positioning, and life detection equipment to dynamically update data. Information fusion algorithms perform spatio-temporal correlation and consistency checking on multi-source data, identify and eliminate conflicting information, and establish a credibility evaluation mechanism. Target priority evaluation is based on factors such as personnel life safety, property value, environmental impact, and rescue difficulty for comprehensive sorting. Information representation uses a structured data format, including target location, target type, urgency, rescue difficulty, and time constraints. For example, in a hotel fire, the rescue target information shows that there are 12 trapped guests in the guest rooms on the eighth floor, including 3 elderly people in room 808 who need to be rescued first, and 8 trapped personnel in the conference room on the fifth floor but are close to the safe exit, with relatively low rescue difficulty. Through multi-source information fusion and priority analysis, the rescue target information is formed.

[0057] In some embodiments, the task decomposition of the rescue target information to generate a fragmented rescue task set comprises: performing complexity analysis on the rescue target information to obtain a task complexity evaluation; decomposing the rescue task into the smallest execution unit based on the task complexity evaluation; performing correlation analysis on the smallest execution unit to generate a task dependency relationship; and generating a fragmented rescue task set according to the task dependency relationship.

[0058] Performing complexity analysis on the rescue target information to obtain a task complexity evaluation. The time complexity evaluation considers the expected execution time, time window constraints and timing dependency relationship of the rescue task, analyzes the time sensitivity and urgency of the task. The spatial complexity evaluation analyzes the spatial distribution, accessibility and environmental complexity of the rescue target, including factors such as floor height, passage narrowness, obstacle distribution and spatial restrictions. The resource complexity evaluation considers the human resources, equipment resources and technical resources required by the rescue task, analyzes the complexity of resource allocation and the possibility of resource conflicts. The coordination complexity evaluation analyzes the mutual influence between multiple rescue targets, the coordination needs between rescue teams and the complexity of command and control. The task complexity comprehensive evaluation formula is C = w1·C t +w2·C s +w3·C r +w4·C o , constraint condition: w1+w+w3+w4 = 1, and w1, w2, w3, w4 ≥ 0, wherein C represents the comprehensive complexity index, C t is the time complexity, C s is the spatial complexity, C r is the resource complexity, and C o is the coordination complexity, w i is the weight coefficient of each dimension. The complexity calculation adopts a weighted scoring model, and each dimension complexity is quantitatively evaluated through expert knowledge base and historical case data. The algorithm also considers the influence of uncertainty factors on complexity, and processes the complexity increase caused by incomplete information and environmental changes through fuzzy logic and probability analysis method. For example, in a fire in a comprehensive mall, the rescue task of the electrical appliance store on the first floor is evaluated as a high complexity task because it involves flammable and explosive goods, open space but heavy smoke, requires professional equipment and multi-team coordination, while the rescue task of the dining area on the third floor is evaluated as a medium complexity task because it involves a large number of people, relatively simple passages and standard equipment requirements. Based on multi-dimensional analysis and quantitative evaluation, the task complexity evaluation result is obtained.

[0059] The rescue task is decomposed into the smallest execution units based on task complexity assessment. The decomposition principles follow the standards of single responsibility, independent execution, clear input and output, and measurable completion. Hierarchical decomposition starts from the main stages of the rescue task and gradually refines into specific operation steps, each level of decomposition maintains logical integrity and operational feasibility. Functional decomposition splits the composite functions into basic function combinations according to the functional characteristics of the rescue task, ensuring that each smallest unit has a clear functional definition and execution standard. The decomposition process considers the granularity balance of the execution unit, which should be fine enough to support flexible combination, and avoid excessive decomposition leading to management complexity. The smallest execution unit contains detailed task description, execution conditions, required resources, expected time, completion standards, and quality requirements, etc. The unit definition adopts a standardized format to facilitate subsequent combination and scheduling. For example, in a certain office building fire, the complex task of "rescuing trapped personnel on the eighth floor" is decomposed into "arriving on the eighth floor", "searching for trapped personnel", "evaluating personnel status", "selecting rescue method", "performing rescue operation", "escorting to the safe area", etc. Each unit has specific operation specifications and completion markers. Through hierarchical and functional decomposition, a library of the smallest execution units is established.

[0060] The smallest execution units are analyzed for relevance to generate task dependency relationships. Logical relationship analysis identifies the prerequisites, trigger conditions, and constraint conditions between execution units, determines which units must be executed after the completion of other units. Time sequence relationship analysis determines the time sequence, parallel execution possibility, and time window constraints of execution units, and establishes a time dependency graph. Resource relationship analysis identifies execution units that need to share resources, analyzes resource conflicts and resource coordination needs, and establishes a resource dependency matrix. Spatial relationship analysis considers the spatial location requirements and movement path dependencies of execution units, identifies spatial conflicts and spatial coordination needs. Dependency relationship modeling uses a directed acyclic graph structure, with nodes representing the smallest execution units and edges representing dependency relationships. The attributes of the edges include dependency type, strength, and constraint conditions. The algorithm also performs consistency checking of dependency relationships, identifies and solves problems such as circular dependency, conflict dependency, and unsatisfiable dependency. Dependency relationship optimization finds the most efficient dependency configuration through critical path analysis and parallelism analysis. For example, in a certain factory fire rescue, the "searching for trapped personnel" unit must be executed after the "arriving at the accident area" unit is completed, "evaluating personnel status" and "preparing rescue equipment" can be executed in parallel, and "performing rescue operation" depends on the completion of the previous two units, forming a clear dependency relationship network. After relevance mining and dependency modeling, the task dependency relationship graph is generated.

[0061] According to the task dependency relationship, a fragmented rescue task set is generated. The set generation considers the constraint conditions of the dependency relationship, ensures that the execution units within each task fragment can be executed in the correct order, and the fragments can be combined in parallel or series. The algorithm uses topological sorting and strong connected component decomposition techniques to identify task subgraphs that can be executed independently, forming relatively independent task fragments. Fragment division balances execution efficiency and management complexity, taking full advantage of parallel execution opportunities while controlling the number of fragments and the coordination complexity between fragments. Each task fragment contains complete execution unit sequences, resource requirement lists, execution time estimates, preconditions, and output results. The task set also contains coordination mechanisms between fragments, synchronization point settings, and exception handling plans. Set optimization uses critical path analysis and load balancing techniques to find the shortest execution time and optimal resource utilization fragment combination scheme. For example, in a hospital fire rescue, the fragmented task set includes "emergency transfer of intensive care unit personnel", "surgery room equipment protection", "medicine warehouse safety disposal", "evacuation passage cleaning", and other relatively independent task fragments, each containing several minimum execution units. Fragments are synchronized through time nodes and resource coordination. Through dependency analysis and fragmentation organization, a fragmented rescue task set is constructed.

[0062] In some embodiments, the dynamic reorganization of the fragmented rescue task set based on the firefighter state data generates a personalized task allocation scheme, including: performing capability evaluation on the firefighter state data to obtain personal capability indicators; matching the fragmented rescue tasks based on the personal capability indicators to obtain task matching results; dynamically reorganizing according to the task matching results to obtain a reorganized task sequence; and generating a personalized task allocation scheme based on the reorganized task sequence.

[0063] The personal ability index is obtained by performing ability evaluation on the firefighter state data. The evaluation dimensions include physical ability, technical ability, experience ability, and psychological ability, etc. The physical ability evaluation is based on vital sign data, fatigue index, physical reserve, and health status to analyze the physical endurance and continuous operation ability of the firefighter. The technical ability evaluation is based on professional skill level, equipment operation proficiency, special skill certification, and training record to evaluate the technical competence of the firefighter in different rescue tasks. The experience ability evaluation is based on historical rescue participation record, similar scene disposal experience, emergency disposal success rate, and team cooperation performance to quantify the practical experience value of the firefighter. The psychological ability evaluation is based on stress bearing index, decision reaction speed, risk judgment ability, and team communication ability to analyze the psychological adaptability of the firefighter in high-pressure environment. The ability calculation adopts fuzzy comprehensive evaluation and neural network method to comprehensively analyze static ability profile and dynamic state monitoring data to generate real-time ability evaluation results. The evaluation results also consider the timeliness and decay characteristics of the ability, and predict the ability change trend according to the task duration and intensity. For example, in a high-rise residential fire, the physical ability evaluation of firefighter Li shows that the current fatigue level is low, the physical energy is sufficient, the technical ability is outstanding in high-altitude rescue, the experience ability has rich experience in similar building rescue, the psychological ability remains calm in complex environment, and the comprehensive ability index is suitable for undertaking high-difficulty rescue tasks. Based on multi-dimensional evaluation and dynamic analysis, the personal ability index is obtained.

[0064] The task matching result is obtained by matching the fragmented rescue tasks based on the personal ability index. First, the ability requirement of each task segment is analyzed to identify the specific requirements and importance of the task on physical ability, technical ability, experience ability, and psychological ability. The ability requirement modeling adopts multi-attribute decision theory to establish a task-ability requirement matrix to clearly define the demand level of each type of task on different ability dimensions. The personnel and task matching degree calculation formula is: where the adaptability function is: Constraint condition: w1+w2+...+w n =1, and all w k ≥0, where M(i,j) represents the matching degree of personnel i and task j, represents the kth ability index of personnel i, represents the demand level of the kth ability of task j, and w kis the weight coefficient of the kth ability. The matching calculation adopts a multi-objective optimization algorithm to find the personnel allocation scheme with the highest ability utilization efficiency and the strongest ability complementarity under the premise of meeting the task ability requirements. The algorithm considers multiple optimization objectives such as ability matching degree, ability margin, ability complementarity, and load balancing, and finds a balanced solution through the Pareto optimal solution set. The matching results are represented in the form of a matching matrix and a matching score, recording the matching degree and recommended priority of each firefighter to each task segment. For example, in a shopping mall fire, the high-altitude rescue task segment has the highest matching degree with firefighter Wang, who has high-altitude operation certification and rich high-rise rescue experience, while the equipment operation task segment has the optimal matching degree with firefighter Zhang, who is proficient in professional rescue equipment. The chemical disposal task forms the best match with firefighter Liu, who has hazardous chemical disposal qualifications. After intelligent matching and multi-objective optimization, the task matching results are obtained.

[0065] According to the task matching results, the dynamic reorganization obtains the reorganized task sequence. According to the matching results, the urgency and executability of the task segments are re-evaluated, and the task segments with high matching degree and strong criticality are prioritized. The sequence optimization adopts dynamic programming and genetic algorithm to find the shortest total execution time and the least resource conflict in the task execution order under the premise of meeting the task dependency relationship. The reorganization algorithm also considers load balancing and ability complementarity to avoid overloading some personnel while others are idle, while ensuring the complementarity and synergy of team capabilities. The dynamic adjustment mechanism updates the task sequence and personnel allocation in real time according to the feedback information during task execution, adapting to changes in the field situation. The reorganization results are represented in the form of a time Gantt chart and a task network diagram, clearly showing the time arrangement and personnel allocation of task execution. For example, in a factory fire rescue, the originally scheduled equipment rescue task is executed in advance because it is matched to professional technical personnel, and the personnel search and rescue task is re-grouped according to personnel ability, forming two parallel search and rescue teams, significantly improving the overall rescue efficiency. Through dynamic scheduling and sequence optimization, the reorganized task sequence is obtained.

[0066] A personalized task allocation scheme is generated based on the reorganized task sequence. With the reorganized task sequence, a detailed personalized rescue action plan and execution guidance scheme are developed. The allocation scheme contains complete information of each firefighter's specific task arrangement, execution time, work location, required equipment, collaboration partners, and safety precautions. Personalized design tailors task execution strategies and operation specifications according to each firefighter's ability characteristics, current state, and task matching situation. The scheme generation considers the operability and safety of task execution, and configures sufficient capacity margin and safety measures for each task. The allocation strategy also contains task rotation mechanism and fatigue management measures, and reasonably arranges personnel rotation and post exchange according to task intensity and duration. Emergency plan design targets possible personnel injury, equipment failure, environmental changes and other unexpected situations, and develops corresponding task reallocation and personnel scheduling schemes. The scheme representation adopts a multi-level structure, containing overall allocation overview, group task division, individual detailed arrangement, and emergency scheduling plan, and other information at different levels. Quality control mechanism ensures the execution effect of the task allocation scheme through task completion standards, quality checkpoints, and performance evaluation indicators. For example, in a hospital fire rescue, the personalized task allocation scheme allocates complex critical patient transfer tasks to experienced firefighters, arranges high-intensity equipment moving tasks for young firefighters with abundant physical energy, and configures professional equipment operation tasks for firefighters with outstanding technical expertise. Each firefighter has clear task objectives, execution paths, and collaboration requirements. Combined with the ability characteristics of personnel and the requirements of task execution, the personalized task allocation scheme is finally established.

[0067] Step S140, execute rescue operation in the safe operation range according to the personalized task allocation scheme to generate rescue execution record, and establish a bidirectional state confirmation channel between the rescuer and the rescued person based on the rescue execution record, and obtain bidirectional confirmation data through the bidirectional state confirmation channel.

[0068] The rescue execution record is generated according to the personalized task allocation scheme within the safe operation range. According to the time arrangement, personnel configuration and task priority in the task allocation scheme, specific execution instructions and operation guidance are issued to each rescue team. The instruction content includes task target, execution path, operation specification and safety requirement information. The safety operation range constraint ensures that all rescue operations are carried out within the predetermined safety boundary, and the position and activity range of the rescue personnel are monitored in real time through GPS positioning technology. During the operation execution process, the firefighters upload the execution state information in real time through the head-mounted camera equipment, voice recorder and operation sensor. The video record captures the visual information of the rescue site, the voice record captures the communication and on-site description between the rescue personnel, and the operation sensor monitors the use state and function execution of the rescue equipment. The execution record generation adopts multi-modal data fusion technology to synchronize and integrate video, audio, sensor data and location information according to the time axis. The record structure storage contains key information such as execution time, execution location, execution personnel, execution task, operation steps and completion status. For example, in a high-rise apartment fire rescue, according to the personalized task allocation scheme, the firefighter search team enters the twelfth floor area according to the predetermined route, searches for trapped personnel through breaking doors and windows, and the record of the whole process shows that 3 trapped residents are successfully found, and the voice record shows effective communication between the teams. Through intelligent scheduling and multi-modal record, detailed rescue execution records are formed.

[0069] Based on the rescue execution record, a two-way state confirmation channel is established between the rescuer and the rescued person. Communication technology adopts multiple alternative schemes according to the on-site environmental conditions, including a wireless intercom system providing short-distance voice communication, a mobile communication network using mobile phone signals for long-distance contact, an emergency communication device including a portable signal transmitter and a life detection communicator, and a simple signal device using knocking sound and light signals in non-electronic ways. The interactive interface design takes into account the physical and psychological state of the rescued personnel, and uses simple and intuitive operation methods and clear and explicit information expression. The rescuer interface is integrated into a portable device, which displays the basic information, health status, location information and communication status of the rescued personnel. The rescued person interface selects voice dialogue, key confirmation, gesture recognition and physiological signal detection according to specific circumstances. Channel testing and optimization ensure the stability and reliability of two-way communication, and signal strength detection and backup scheme switching are used to ensure communication effect. For example, in a fire rescue in an underground shopping mall, the rescue personnel confirmed the location of the 2 workers trapped in the storage room on the first floor through the execution record, and established voice contact with the trapped personnel using a portable emergency communication device. The trapped personnel can feed back their physical condition and needs through simple key operation. Through the integration of communication technology and interactive design, a two-way state confirmation channel is established.

[0070] The bidirectional confirmation data is obtained through the bidirectional state confirmation channel. The data collection framework establishes two parallel data streams of rescuer state confirmation and rescued person state confirmation. The rescuer state confirmation includes rescue progress report, records the degree of completion of task execution and subsequent plans, field environment assessment, including temperature, smoke and structural stability, observation of the state of the rescued personnel, records the state of consciousness and physical condition, and safety status report. The information categories include the state of consciousness and physical condition, and safety status report. The rescued person state confirmation includes self-health status, collects pain degree and breathing condition through simplified question and answer, degree of consciousness and physical activity ability, psychological state and emotional response, surrounding environment perception, including environmental factors such as temperature and smoke, and rescue demand expression, collects specific requirements such as rescue method and medical needs. The data standardization processing converts the confirmation information of different formats and sources into a unified data structure, establishes standard fields such as time stamp, data source, confirmation type, confirmation content and credibility, and improves the accuracy of the confirmation data through redundant confirmation and consistency check. Real-time data synchronization ensures that the rescue command center can obtain the latest bidirectional confirmation information in time. For example, in a factory fire rescue, the rescuer confirms that the location of the trapped personnel has been reached and the environment is safe and controllable, the three trapped workers confirm that their physical condition is good, but one of them has mild smoke poisoning and needs medical attention, the rescuer confirms that he has the ability to provide on-site medical treatment, and the trapped personnel confirms that they are willing to cooperate with the rescue operation. Using the standardized collection and quality control mechanism, complete bidirectional confirmation data is finally obtained.

[0071] In step S150, based on the bidirectional confirmation data, the effectiveness analysis of the rescue operation is performed to obtain rescue effect evaluation data, the failure risk analysis of the rescue effect evaluation data is performed to identify potential failure factors, and the set of fragmented rescue tasks is adjusted according to the potential failure factors to generate an optimized task combination.

[0072] Specifically, the effectiveness analysis of the rescue operation based on the bidirectional confirmation data obtains rescue effect evaluation data. The analysis framework establishes a cross-validation mechanism for rescuer confirmation data and rescued person confirmation data, identifies the actual effect and existing problems of the rescue operation by comparing the different perspectives of both parties on the same rescue event. The rescue progress effectiveness analysis evaluates the completion quality, time efficiency, and goal achievement of task execution according to the progress report of the rescuer and the perception feedback of the rescued person. The operation quality evaluation judges the improvement degree of the rescue operation on the safety condition of the rescued personnel by analyzing the on-site environment assessment of the rescuer and the physical condition change of the rescued person. The communication effectiveness analysis evaluates the information transmission quality, response timeliness, and understanding accuracy of the bidirectional confirmation channel. The resource utilization efficiency analysis evaluates the rationality of personnel allocation, equipment use, and time allocation by combining the resource demand and actual use of the rescuer. The algorithm adopts a fuzzy comprehensive evaluation method to establish a multi-dimensional effect evaluation index system for each rescue operation. For example, in a certain office building fire rescue, through bidirectional confirmation data analysis, it is found that the rescuer confirmation of the eighth floor personnel search and rescue task shows that the room search is completed according to the plan, but the rescued person confirmation reflects that 2 people in the waiting process have anxiety. The analysis result shows that the search and rescue operation technology is effective, but the psychological support link needs to be improved. Based on the cross-validation of bidirectional data and multi-dimensional evaluation, rescue effect evaluation data is formed.

[0073] The failure risk analysis of the rescue effect evaluation data identifies potential failure factors. Comprehensive risk scanning is performed from the dimensions of technical failure, process failure, personnel failure, and environmental failure. Technical failure analysis identifies potential problems such as equipment failure, communication interruption, positioning deviation, etc. on the technical level, and gives early warnings through abnormal patterns in the effect evaluation data. Process failure analysis checks the weaknesses of task execution sequence, coordination, time arrangement, etc. in the process link, and identifies process defects that may cause rescue delay. Personnel failure analysis evaluates the capacity matching, fatigue state, psychological pressure, etc. of the rescue personnel, and identifies insufficient personnel capacity through performance deviation in the effect evaluation. Environmental failure analysis considers the negative impact of environmental factors such as fire change, structure damage, weather influence, etc. on the rescue effect. The algorithm establishes a correlation analysis model of failure factors to identify the combined risk that may be caused by multiple factors. For example, in a certain factory fire rescue, the failure risk analysis finds that there are technical risks of communication equipment performance degradation in high temperature environment, process risks of time conflict between personnel search and rescue and equipment rescue tasks, and personnel risks of insufficient experience of some rescue personnel in chemical disposal. Through systematic risk scanning and correlation analysis, the key potential failure factors are identified.

[0074] In some embodiments, the adjusting the set of fragmented rescue tasks according to the potential failure factors to generate an optimized task combination comprises: performing impact degree evaluation on the potential failure factors to obtain a failure risk level; filtering task fragments that need to be adjusted based on the failure risk level; reconfiguring the task fragments that need to be adjusted to obtain an adjustment scheme; and reorganizing fragmented rescue tasks according to the adjustment scheme to generate an optimized task combination.

[0075] The impact degree evaluation on the potential failure factors obtains a failure risk level. The evaluation dimensions include key indicators such as impact range, impact severity, occurrence probability, and response difficulty. The impact range evaluation analyzes the number of tasks, the number of personnel, and the spatial range that may be affected by the failure factors. The impact severity evaluation considers the potential damage degree of the failure factors to the rescue target, personnel safety, and time delay. The occurrence probability evaluation predicts the possibility of actual occurrence of the failure factors based on historical data, current conditions, and environmental change trends. The algorithm uses a fuzzy evaluation method to calculate a comprehensive risk index for each failure factor. The risk level classification uses a three-level classification system of high, medium, and low. High-risk factors have significant impact and high occurrence probability, medium-risk factors have moderate impact or probability, and low-risk factors have limited impact and low occurrence probability. For example, in a hospital fire rescue, the high-temperature failure of communication equipment is rated as a high-risk level because it affects global coordination and has a high occurrence probability, and the shortage of transfer equipment for critically ill patients is rated as a medium-risk level because it affects the local area but has a backup solution. After multidimensional evaluation and level classification, the failure risk level classification is obtained.

[0076] The task fragments that need to be adjusted are filtered based on the failure risk level. The filtering strategy focuses on high-risk level failure factors and the task fragments directly affected by them, while also considering the potential threat of medium-risk factors. The impact propagation analysis establishes the mapping relationship between failure factors and each task fragment in the set of fragmented rescue tasks, and identifies the specific task fragments that may be affected by each failure factor and the degree of influence. The task fragment vulnerability evaluation analyzes the sensitivity of each task fragment in the set of fragmented rescue tasks to different failure factors, and identifies task fragments with high vulnerability as the focus of adjustment. The algorithm considers the dependency relationship between task fragments to avoid negative impact on other fragments in the set when adjusting a fragment. The screening results are sorted according to the adjustment priority to form a task list that needs to be adjusted from the set of fragmented rescue tasks. For example, in a shopping mall fire rescue, based on the high-risk communication interruption factor, the task fragments that need to be adjusted from the set of fragmented rescue tasks include the search and rescue coordination tasks on each floor, the equipment deployment tasks, and the personnel status monitoring tasks. Through risk propagation analysis and priority sorting, the key task fragments that need to be adjusted from the set of fragmented rescue tasks are selected.

[0077] The reconfiguration of the task fragments that need to be adjusted obtains an adjustment scheme. The configuration strategy adopts different adjustment methods according to the characteristics of the failure factors, including task decomposition, task merging, task replacement, task enhancement and task rearrangement. Task decomposition splits complex or high-risk task fragments in the fragmented rescue task set into multiple simple sub-tasks, reducing the influence range of single-point failure. Task merging integrates related small task fragments in the set into a unified task, improving execution efficiency. Task replacement replaces the original task fragments in the set with lower-risk alternative solutions. Task enhancement strengthens the reliability of the task fragments in the set by increasing backup resources and safety guarantees. The reconfiguration process considers actual constraint conditions such as personnel capability matching, equipment resource constraints and time window limitations. The adjustment scheme contains detailed task descriptions, personnel allocation, equipment configuration, time scheduling and other elements. For example, in a hotel fire rescue, in view of the high-risk problem of the elevator shaft search and rescue task fragment in the fragmented rescue task set, the adjustment scheme includes splitting the original single team search and rescue into a double team cooperative search and rescue, increasing backup communication equipment, and adjusting the execution time to avoid unstable structure periods. Based on the risk characteristics and constraint conditions, a detailed adjustment scheme for the fragmented rescue task set is formed.

[0078] According to the adjustment scheme, the fragmented rescue task is reorganized to generate an optimized task combination. The reorganization process first updates the task dependency relationship graph in the fragmented rescue task set, reflecting the influence and changes of the adjusted task fragments on the original dependency relationship. The task scheduling optimization uses a heuristic algorithm to find the optimal task execution order and resource allocation scheme under the premise of meeting the dependency relationship. The optimization objectives include multiple objectives such as the shortest total execution time, the lowest risk level and the optimal resource utilization. The algorithm considers the robustness of the task combination to ensure that the optimized task combination can cope with environmental changes and unexpected situations. The final optimized task combination is an improved version of the original fragmented rescue task set, containing complete task list, execution plan, resource configuration and risk control measures, etc. For example, in a factory fire rescue, after reorganizing and optimizing the fragmented rescue task set, the original 6 parallel task fragments are adjusted to 4 serial-parallel mixed task fragments, and the high-risk chemical disposal task in the set is divided into three sub-tasks of environmental assessment, professional disposal and safety monitoring, and the risk level of the whole task combination is reduced from high risk to medium risk. Through the integration and combination optimization of the fragmented rescue task set, the optimized task combination is finally established.

[0079] In step S160, the optimized task combination is executed to obtain rescue process data, the rescue process data is subjected to abnormal pattern recognition to generate an abnormal early warning signal, and a safety state monitoring result is generated based on the evacuation path scheme combined with the abnormal early warning signal.

[0080] Specifically, the monitoring obtains rescue process data for the optimization task combination. A multidimensional data collection system is established, including key dimensions such as task execution progress monitoring, personnel state monitoring, environmental change monitoring, and resource use monitoring. Task execution progress monitoring tracks the completion status and time progress of the task, and real-time monitoring of the execution of each task segment in the optimization task combination ensures that the optimization task combination progresses in an orderly manner according to the established plan. Taking the personnel search and rescue task in the fire rescue of a high-rise office building as an example, the monitoring module can collect the execution progress of the task segment in the optimization task combination in real time, which is 75%, indicating that the rescue team has completed the search of the east area of the eighth floor. Personnel state monitoring continuously tracks the physiological state, location information, equipment state, and communication status of firefighters executing the optimization task combination. When the monitoring finds that the personnel participating in the execution of the optimization task combination are in good physical condition but the pressure of the protective equipment gas cylinder is decreasing, timely replacement of the equipment is prompted. Environmental change monitoring collects fire field environmental parameters that affect the execution effect of the optimization task combination, including dynamic changes in temperature, smoke, gas concentration, and evaluates the influence of the environment on the execution of the optimization task combination. Resource use monitoring records the use status and performance of rescue equipment, firefighting equipment, medical supplies, and communication equipment during the execution of the optimization task combination, ensuring sufficient resource support for the optimization task combination. Through multidimensional real-time monitoring and data collection, comprehensive rescue process data for the execution of the optimization task combination is obtained.

[0081] Anomaly pattern recognition is performed on the rescue process data to generate an abnormal warning signal. A multi-level anomaly detection system is established using a combination of statistical analysis, machine learning, and expert rules. Statistical analysis identifies abnormal fluctuations and trend deviations in rescue process data by setting normal value ranges, change rate thresholds, and statistical control limits. In the actual operation of the high-rise office building fire rescue, when the algorithm detects intermittent communication signal interruptions during the execution of the eighth floor personnel search and rescue task, it can identify this communication anomaly pattern. Time series anomaly detection uses sliding window analysis techniques to identify sudden changes and abnormal trends in the data. When the personnel transfer speed is 15% slower than expected, trend analysis can capture this efficiency decline anomaly pattern. Multivariate anomaly detection considers the correlation between different monitoring indicators and identifies abnormal combinations in multidimensional data. Anomaly severity assessment grades the detected anomalies according to their type, duration, impact range, and development trend. The warning signal generation uses a hierarchical warning mechanism, including minor anomaly reminders, important anomaly warnings, and severe anomaly alarms. When some rescue personnel show signs of mild fatigue, combined with multiple anomalies such as communication interruptions and efficiency declines, the algorithm generates an important anomaly warning signal, reminding the command center to pay attention to communication support and personnel rotation needs, and finally generates a hierarchical abnormal warning signal.

[0082] The safety state monitoring result is generated based on the evacuation path scheme combined with the abnormal early warning signal. A dual evaluation system of abnormal influence analysis and evacuation capacity evaluation is established. The abnormal influence analysis evaluates the specific influence of the abnormality on rescue safety according to the type and severity of the abnormal early warning signal. The evacuation capacity evaluation evaluates the safety evacuation capacity under abnormal conditions by using the path configuration, traffic capacity and time parameters in the evacuation path scheme, combined with the current personnel distribution and environmental conditions. In the specific application of high-rise office building fire rescue, when the evaluation is based on the communication interruption abnormal early warning signal, the algorithm will check the communication coverage of each path in the evacuation path scheme, and find that the east main evacuation path is difficult to coordinate due to communication failure, but the communication equipment of the west standby path can still be used normally. The time safety analysis calculates the time required for safe evacuation under the current abnormal state, and compares it with the fire spread prediction to evaluate the time safety margin. By calculating the current personnel distribution and available path capacity, the evaluation result shows that the safe evacuation can be completed within 12 minutes, while the fire spread prediction shows that there is still 20 minutes of safety time, and there is sufficient time safety margin. The comprehensive safety evaluation adopts a multi-factor weight model to comprehensively calculate the abnormal influence degree, evacuation capacity, path availability and time safety, and generate a safety state level. The safety state classification adopts a four-level classification system of green safety, yellow attention, orange warning and red danger. Based on the above analysis, the comprehensive evaluation result is yellow attention state, and it is recommended to strengthen the personnel guidance of the west path and restore the east communication as soon as possible, and finally form the safety state monitoring result.

[0083] Step S170, according to the safety state monitoring result, dynamically adjust the current rescue target to generate a redefined rescue standard, convert the risk dimension of the redefined rescue standard to generate a controllable risk parameter, develop a rescue strategy based on the controllable risk parameter to generate a final control instruction, and complete the intelligent fire field fire rescue control.

[0084] Specifically, the current rescue target is adjusted and optimized in real time by dynamic assessment and target reconstruction technology using the safety state monitoring results. The adjustment strategy determines the direction and amplitude of target adjustment according to the safety level, main risk factors and dynamic trend in the safety state monitoring results. When the safety state level is green safety, the original rescue target and standard are maintained; when it is yellow attention state, the target priority and execution standard are adjusted moderately; when it is orange warning or red danger state, the rescue target is adjusted significantly to prioritize personnel safety. The target adjustment process considers factors such as rescue resource constraints, time window limitations and environmental change trends. In a high-rise office building fire rescue, when the safety state monitoring result shows yellow attention state and the main risk factor is communication interruption on the east side, the target adjustment algorithm adjusts the original "comprehensive search and rescue of all areas on the eighth floor" to "priority search and rescue of the west side communicable area, and temporary suspension of the east side area until communication is restored". The rescue standard redefinition includes adjusted target content, execution priority, time requirement, safety boundary and resource allocation elements. The standard reconstruction uses a multi-objective optimization method to maximize rescue effectiveness while ensuring safety. The redefined rescue standard also contains a dynamic adjustment mechanism that can be adjusted again according to subsequent safety state changes. Based on safety state analysis and target optimization, the redefined rescue standard is generated.

[0085] In some embodiments, the risk dimension conversion of the redefined rescue standard generates controllable risk parameters, including: risk element identification of the redefined rescue standard to obtain risk constituent elements; establishing a multi-dimensional risk assessment system based on the risk constituent elements; utilizing the multi-dimensional risk assessment system for risk quantification processing; and generating controllable risk parameters according to the risk quantification processing results.

[0086] The redefined rescue standard is used to systematically identify various risk factors affecting the implementation of the rescue standard through risk decomposition and factor extraction techniques. The identification dimensions include key categories such as personnel risk factors, environmental risk factors, technical risk factors, and time risk factors. Personnel risk factor analysis includes factors such as the ability of rescue personnel to match, fatigue level, equipment status, and collaboration efficiency. In the redefined standard for high-rise office building fire rescue, when the west side area is given priority for search and rescue, the personnel risk factors include whether the number of personnel in the west rescue team is sufficient and whether they have the familiarity with the area. Environmental risk factors assess the potential threats of environmental factors such as fire development, smoke diffusion, structural stability, and weather conditions to the implementation of the standard. Technical risk factors identify the reliability and applicability of technical factors such as communication equipment, rescue equipment, monitoring equipment, and information processing technology. Time risk factors analyze the matching relationship between the time required to implement the redefined standard and the available time window. The factor identification method combines expert knowledge base and historical case analysis to ensure the comprehensiveness and accuracy of risk factor identification. The identification results are stored in the form of a structured factor list, each factor containing type, impact degree, occurrence probability, and control difficulty attributes. Through systematic risk decomposition and factor extraction, risk constituent factors are obtained.

[0087] Based on the risk constituent factors, a comprehensive evaluation framework covering various risk factors is established through analytic hierarchy process and weight allocation techniques. A multi-level architecture is adopted, including risk category layer, risk factor layer, and risk indicator layer at different levels. The risk category layer classifies and summarizes the risk constituent factors according to their nature and characteristics, the risk factor layer further decomposes each type of risk, and the risk indicator layer establishes specific quantitative evaluation indicators. Weight allocation uses a combination of analytic hierarchy process and expert evaluation to assign appropriate weight coefficients to risk factors at different levels. In the case of high-rise office building fire rescue, the evaluation system may allocate 30% weight to personnel risk factors, 40% weight to environmental risk factors, 20% weight to technical risk factors, and 10% weight to time risk factors, reflecting the dominant position of environmental risk in this scenario. System verification through historical case review and expert review ensures the scientificity and applicability of the evaluation system. Standardized processing of evaluation indicators ensures the comparability and operability of different types of indicators. Finally, through analytic hierarchy process and weight allocation techniques, a complete multi-dimensional risk evaluation system is established.

[0088] The multi-dimensional risk assessment system is used to convert the qualitative risk description into quantitative risk values through numerical calculation and statistical analysis techniques. The quantitative method uses a combination of fuzzy comprehensive evaluation, analytic hierarchy process, and Monte Carlo simulation techniques. Fuzzy comprehensive evaluation handles the uncertainty and fuzziness in risk assessment by establishing a fuzzy evaluation matrix and fuzzy operations to obtain the fuzzy evaluation results of each risk factor. The analytic hierarchy process handles the weight relationship between risk factors by constructing a judgment matrix and consistency check to determine the relative importance of each factor. Monte Carlo simulation handles the randomness and variability in risk assessment by random sampling and probability distribution to estimate the probability distribution and confidence interval of risk indicators. In the specific application of fire rescue in high-rise office buildings, the quantitative processing may result in a personnel risk index of 0.3, an environmental risk index of 0.6, a technical risk index of 0.4, a time risk index of 0.2, and a comprehensive risk index of 0.43. The quantitative results also include risk level classification, which maps continuous risk values to discrete risk levels, facilitating decision-making and control. The quantitative processing results are output in the form of a risk assessment report, including quantitative values, level assessment, probability distribution, and trend analysis of various risks. Through numerical calculation and statistical analysis, the risk quantification results are obtained.

[0089] For example, generating controllable risk parameters based on the risk quantification results includes: performing risk threshold division on the risk quantification results to obtain a risk control boundary; establishing a risk adjustment mechanism based on the risk control boundary to generate an adjustment strategy; using the adjustment strategy to convert the risk level to obtain a conversion coefficient; and generating controllable risk parameters based on the conversion coefficient.

[0090] Firstly, the risk threshold is divided to obtain the risk control boundary, and the technical route of combining the percentile analysis, K-means clustering analysis and expert experience is adopted to determine the critical threshold and control boundary of different risk levels through the distribution characteristics of historical risk data and the safety standards of fire rescue. In the fire rescue of high-rise office buildings, the low-risk boundary is set to 0.0-0.3, the low-moderate-risk boundary is set to 0.3-0.5, the moderate-high-risk boundary is set to 0.5-0.7, and the high-risk boundary is set to 0.7-1.0. The boundary setting also considers the different characteristics of different risk types. Then, the risk adjustment mechanism is established based on the risk control boundary to generate the adjustment strategy, and the hierarchical response and step-up control mode is adopted to establish a complete adjustment system including the early warning mechanism, active adjustment mechanism and emergency response mechanism. When the current risk index is 0.43 and is in the low-moderate-risk boundary, the early warning mechanism is started, and the adjustment strategy includes increasing the personnel configuration in the west area, strengthening the backup deployment of communication equipment, shortening the state reporting interval to every 5 minutes, and establishing a feedback control loop for dynamic adjustment. Next, the risk level is converted to obtain the conversion coefficient by using the adjustment strategy, and the risk reduction ability and execution efficiency of each adjustment strategy are evaluated through control effect analysis and historical data verification technology. Considering the control ability, response time and resource constraints, the mathematical model of risk conversion is R'=R×∏ i＝1 m(1-η i ·α i ), and the conversion coefficient calculation formula is k=R' / R=∏ i＝1 m(1-η i ·α i ), where R is the original risk level, R' is the risk level after adjustment, η i is the maximum risk reduction ability of the i th adjustment strategy, α i is the implementation intensity coefficient of the i th strategy (0≤α i ≤1), and m is the total number of adjustment strategies. Finally, the controllable risk parameters are generated according to the conversion coefficient, and the function relationship between the conversion coefficient and the control parameters is established through parameter mapping and standardization technology to convert the risk control information into the standardized parameter form directly used for rescue strategy formulation. The controllable risk parameters include a complete parameter system such as risk control level, adjustment intensity coefficient, personnel resource allocation ratio, equipment resource allocation ratio, logistics resource allocation ratio, time constraint coefficient and safety margin coefficient. For example, in the fire rescue of a certain high-rise office building, when the risk conversion coefficient indicates that the risk level is effectively controlled, the corresponding risk control level of "moderate controllable" state is automatically generated, and the appropriate adjustment intensity is determined to reasonably allocate personnel resources to the search and rescue tasks in the communicable area, and equipment resources are mainly used for communication recovery and life detection. The logistics resources are reserved for support and emergency response, and reasonable time constraints and sufficient safety margins are set,

[0091] The final control instruction is generated based on the controllable risk parameter to formulate the rescue strategy. A method combining multi-objective optimization and constraint programming is adopted to maximize the rescue effect and resource utilization efficiency under the premise of meeting safety constraints. The optimization objectives include maximizing the success rate of rescue, minimizing the execution time, minimizing the resource consumption, and minimizing the risk level. The constraint conditions include personnel capability constraints, equipment resource constraints, time window constraints, and safety boundary constraints. In the fire rescue of a high-rise office building, based on the controllable risk parameter, the strategy formulation may determine the following scheme: 65% of personnel resources are preferentially allocated to the west communicable area, 25% of equipment resources are deployed for communication recovery and life detection, 10% of logistics resources are reserved for support and emergency, the rescue time target is set to be completed within 15 minutes in the west area, and the safety margin coefficient is 1.2 to ensure sufficient safety time. The control instruction generation converts the abstract rescue strategy into specific operation instructions and task allocation. The instruction content includes personnel allocation instructions, equipment deployment instructions, time control instructions, communication coordination instructions, and safety monitoring instructions. The instruction format adopts a standardized command control protocol to ensure accurate communication and efficient execution of the instructions. The instruction priority setting ensures the priority execution of key instructions, and the instruction tracking mechanism monitors the execution status and completion of the instructions. The final control instruction is issued to each execution unit in the form of an instruction package, including complete task description, execution time, resource allocation, cooperation requirements, and safety reminders. Through strategy optimization and instruction generation, intelligent fire field fire rescue control is completed, and full-process automatic control from risk perception to intelligent decision-making is realized.

[0092] In order to perform the intelligent fire field fire rescue control method corresponding to the method embodiment, the corresponding functions and technical effects are realized. Referring to Figure 2 , Figure 2 The structure block diagram of the intelligent fire field fire rescue system 200 provided by the embodiment of the present application is shown. For ease of illustration, only the parts related to the present embodiment are shown. The intelligent fire field fire rescue system 200 provided by the embodiment of the present application includes:

[0093] The data acquisition module 201 is configured to acquire firefighter position information and vital sign data to generate firefighter state data, acquire temperature change information around the firefighters, perform gradient analysis on the temperature change information, and reversely track a fire source position to generate fire source distribution data.

[0094] The path planning module 202 is configured to perform fire spread prediction based on the fire source distribution data to generate a dangerous area distribution, identify a boundary area of the dangerous area distribution to generate a relatively safe area, construct an evacuation path scheme from the dangerous area distribution to the relatively safe area, and formulate a rescue action boundary based on the evacuation path scheme to generate a safe operation range.

[0095] The task allocation module 203 is configured to acquire rescue target information, decompose the rescue target information to generate a fragmented rescue task set, dynamically recombine the fragmented rescue task set based on the firefighter state data to generate a personalized task allocation scheme;

[0096] The rescue execution module 204 is configured to execute rescue operations in the safe operation range according to the personalized task allocation scheme to generate a rescue execution record, establish a bidirectional state confirmation channel between rescuers and rescued persons based on the rescue execution record, and acquire bidirectional confirmation data through the bidirectional state confirmation channel;

[0097] The effect evaluation module 205 is configured to analyze the effectiveness of rescue operations based on the bidirectional confirmation data to acquire rescue effect evaluation data, analyze the failure risk of the rescue effect evaluation data to identify potential failure factors, and adjust the fragmented rescue task set according to the potential failure factors to generate an optimized task combination.

[0098] The safety monitoring module 206 is configured to monitor the optimized task combination to acquire rescue process data, identify abnormal patterns based on the rescue process data to generate an abnormal early warning signal, and perform safety evaluation based on the evacuation path scheme and the abnormal early warning signal to generate a safety state monitoring result.

[0099] The strategy control module 207 is configured to dynamically adjust a current rescue target according to the safety state monitoring result to generate a redefined rescue standard, convert the redefined rescue standard to a controllable risk parameter to generate a rescue strategy, and generate a final control instruction to complete intelligent fire field fire rescue control.

[0100] The intelligent fire field fire rescue system 200 described above can implement the intelligent fire field fire rescue control method of the method embodiment described above. The optional items in the method embodiment described above are also applicable to this embodiment, and will not be described in detail here. The remaining content of the embodiment of the present application can refer to the content of the method embodiment described above, and will not be described in detail in this embodiment.

[0101] The purpose of the above embodiments is to exemplarily reproduce and deduce the technical solutions of the present application, and to completely describe the technical solutions, purposes and effects of the present application. The purpose is to make the public understand the disclosure of the present application more thoroughly and comprehensively, and does not limit the protection scope of the present application.

[0102] The above embodiments are not based on an exhaustive enumeration of the present application, and there can be many other unlisted embodiments. Any substitution and improvement made without violating the concept of the present application is within the protection scope of the present application.

Claims

1. A smart fire scene fire rescue control method, characterized in that, include: Collect firefighter location information and vital sign data to generate firefighter status data, obtain temperature change information around the firefighter, perform gradient analysis on the temperature change information to trace the fire source location in reverse and generate fire source distribution data; Based on the fire source distribution data, fire spread is predicted to generate a dangerous area distribution, the boundary areas of the dangerous area distribution are identified to generate a relatively safe area, an evacuation route plan from the dangerous area distribution to the relatively safe area is constructed, and the boundary of the rescue operation is determined according to the evacuation route plan to generate a safe operating range. Obtain rescue target information, decompose the rescue target information into a fragmented rescue task set, and dynamically reorganize the fragmented rescue task set based on the firefighter status data to generate a personalized task allocation scheme; According to the personalized task allocation scheme, rescue operations are performed within the safe operating range to generate rescue execution records. A two-way status confirmation channel is established between the rescuer and the rescued based on the rescue execution records, and two-way confirmation data is obtained through the two-way status confirmation channel. Based on the two-way confirmation data, effectiveness analysis of rescue operations is performed to obtain rescue effect evaluation data. Failure risk analysis is then performed on the rescue effect evaluation data to identify potential failure factors. Based on the potential failure factors, the fragmented rescue task set is adjusted to generate an optimized task combination. The optimized task combination is monitored to obtain rescue process data, abnormal pattern recognition is performed on the rescue process data to generate abnormal early warning signals, and a safety assessment is performed based on the evacuation route plan and the abnormal early warning signals to generate safety status monitoring results. Based on the safety status monitoring results, the current rescue target is dynamically adjusted to generate a redefined rescue standard. This redefined rescue standard undergoes risk dimension transformation to generate controllable risk parameters, including: identifying risk elements to obtain risk components; establishing a multi-dimensional risk assessment system based on these risk components; quantifying risks using the multi-dimensional risk assessment system; dividing the risk quantification results into risk thresholds to obtain risk control boundaries; establishing a risk adjustment mechanism based on the risk control boundaries to generate adjustment strategies. These risk adjustment mechanisms adopt a tiered response and progressive escalation control mode, including an early warning mechanism, an active adjustment mechanism, and an emergency response mechanism. The adjustment strategies are used to controllably transform the risk level to obtain a transformation coefficient, calculated using the formula: k = R' / R = π. i=1 m(1-ηi·αi), where R is the original risk level, R' is the adjusted risk level and R'=R×k, ηi is the maximum risk reduction capability of the i-th adjustment strategy, αi is the implementation intensity coefficient of the i-th strategy and 0≤αi≤1, and m is the total number of adjustment strategies; controllable risk parameters are generated based on the conversion coefficients, including risk control level, adjustment intensity coefficient, personnel resource allocation ratio, equipment resource allocation ratio, logistics resource allocation ratio, time constraint coefficient, and safety margin coefficient; rescue strategies are formulated based on the controllable risk parameters to generate final control instructions and complete intelligent fire scene fire rescue control.

2. The method according to claim 1, characterized in that, The step of performing gradient analysis on the temperature change information to reverse-track the fire source location and generate fire source distribution data includes: Multi-point temperature gradient analysis is performed on the temperature change information to obtain the gradient vector field; A temperature change trend diagram is constructed based on the gradient vector field; The temperature change trend graph is used to perform reverse path tracing to generate candidate fire source locations; Fire source distribution data is generated based on the candidate fire source locations.

3. The method according to claim 1, characterized in that, The scheme for constructing evacuation routes from the dangerous area to the relatively safe area includes: The dangerous area distribution is analyzed to identify the dangerous boundary coordinates. The shortest evacuation distance is determined based on the coordinates of the dangerous boundary and the relatively safe area; Based on the shortest evacuation distance, multiple alternative evacuation routes are designed; A safety assessment is performed on the multiple alternative evacuation routes to generate an evacuation route plan.

4. The method according to claim 1, characterized in that, The step of defining the rescue operation boundary and generating a safe operating area based on the evacuation route plan includes: Based on the evacuation route plan, analyze the evacuation time requirements and obtain time safety parameters; The boundaries of the rescue operation are determined based on the aforementioned time safety parameters; A safe operating area is generated based on the boundaries of the rescue operation.

5. The method according to claim 1, characterized in that, The step of breaking down the rescue target information into fragmented rescue task sets includes: A complexity analysis is performed on the rescue target information to obtain a task complexity assessment; Based on the aforementioned task complexity assessment, the rescue task is decomposed into the smallest execution unit; Perform correlation analysis on the smallest execution unit to generate task dependencies; A fragmented set of rescue tasks is generated based on the task dependencies.

6. The method according to claim 1, characterized in that, The process of dynamically reorganizing the fragmented rescue task set based on the firefighter status data to generate a personalized task allocation scheme includes: The firefighters' status data is used to conduct a capability assessment to obtain individual capability indicators; The fragmented rescue tasks are matched based on the individual ability indicators to obtain task matching results; Based on the task matching results, a dynamic reorganization is performed to obtain a reorganization task sequence; A personalized task allocation scheme is generated based on the recombined task sequence.

7. The method according to claim 1, characterized in that, The step of adjusting the fragmented rescue task set to generate an optimized task combination based on the potential failure factors includes: The potential failure factors are assessed to determine their impact level and thus the failure risk level. Based on the aforementioned failure risk level, select the task segments that need adjustment; The task segments that need adjustment are reconfigured to obtain an adjustment plan; Based on the aforementioned adjustment scheme, fragmented rescue missions are reorganized to generate optimized mission combinations.

8. An intelligent fire scene firefighting and rescue system, characterized in that, include: The data acquisition module is used to collect firefighter location information and vital sign data to generate firefighter status data, obtain temperature change information around the firefighter, perform gradient analysis on the temperature change information to trace the fire source location in reverse and generate fire source distribution data. The path planning module is used to predict the spread of fire based on the fire source distribution data, generate a dangerous area distribution, identify the boundary areas of the dangerous area distribution to generate a relatively safe area, construct an evacuation path plan from the dangerous area distribution to the relatively safe area, and formulate the rescue operation boundary and generate a safe operation range based on the evacuation path plan. The task allocation module is used to acquire rescue target information, decompose the rescue target information into a set of fragmented rescue tasks, and dynamically reorganize the set of fragmented rescue tasks based on the firefighter status data to generate a personalized task allocation scheme. The rescue execution module is used to perform rescue operations within the safe operating area according to the personalized task allocation scheme, generate rescue execution records, establish a two-way status confirmation channel between the rescuer and the rescued based on the rescue execution records, and obtain two-way confirmation data through the two-way status confirmation channel. The effectiveness evaluation module is used to analyze the effectiveness of rescue operations based on the two-way confirmation data to obtain rescue effectiveness evaluation data, perform failure risk analysis on the rescue effectiveness evaluation data to identify potential failure factors, and adjust the fragmented rescue task set according to the potential failure factors to generate an optimized task combination. The safety monitoring module is used to monitor the optimized task combination to obtain rescue process data, perform abnormal pattern recognition on the rescue process data to generate abnormal early warning signals, and perform safety assessment based on the evacuation route plan and the abnormal early warning signals to generate safety status monitoring results. The strategy control module is used to dynamically adjust the current rescue target and generate a redefined rescue standard based on the safety status monitoring results. It then performs risk dimension transformation on the redefined rescue standard to generate controllable risk parameters. This includes: identifying risk elements in the redefined rescue standard to obtain risk components; establishing a multi-dimensional risk assessment system based on the risk components; quantifying the risk using the multi-dimensional risk assessment system; dividing the risk quantification results into risk thresholds to obtain risk control boundaries; establishing a risk adjustment mechanism based on the risk control boundaries to generate adjustment strategies. The risk adjustment mechanism adopts a graded response and progressive escalation control mode, including an early warning mechanism, an active adjustment mechanism, and an emergency response mechanism. Finally, it uses the adjustment strategy to controllably transform the risk level to obtain a conversion coefficient. The formula for calculating the conversion coefficient is: k = R' / R = π. i=1 m(1-ηi·αi), where R is the original risk level, R' is the adjusted risk level and R'=R×k, ηi is the maximum risk reduction capability of the i-th adjustment strategy, αi is the implementation intensity coefficient of the i-th strategy and 0≤αi≤1, and m is the total number of adjustment strategies; controllable risk parameters are generated based on the conversion coefficients, including risk control level, adjustment intensity coefficient, personnel resource allocation ratio, equipment resource allocation ratio, logistics resource allocation ratio, time constraint coefficient, and safety margin coefficient; rescue strategies are formulated based on the controllable risk parameters to generate final control instructions and complete intelligent fire scene fire rescue control.

Images