Intelligent fire fighting method and system based on multi-modal perception

Abstract

The application relates to an intelligent fire-fighting method and system based on multi-modal perception, and belongs to the field of intelligent fire-fighting. The method comprises the following steps: acquiring a global map, presetting a patrol route and a patrol point, moving according to the patrol route, reaching the patrol point through autonomous path planning, and obtaining environmental perception data; identifying a fire source, obtaining fire scene perception data, positioning the fire source, mapping the fire source positioning to a smart fire-fighting equipment coordinate system; the smart fire-fighting equipment reaches the fire source positioning, sends a signal to a fire-fighting system, cuts off the power supply of the fire scene, and starts smoke exhaust; based on the environmental perception data and the fire scene perception data, the best fire extinguishing medium is selected through multi-dimensional fire extinguishing medium adaptability selection; the best fire extinguishing medium is used for fire extinguishing, after the fire is extinguished, the fire extinguishing is confirmed through fire source identification, and a standby state is returned. The application introduces multi-modal perception and an intelligent decision mechanism, and realizes all-around perception, accurate identification and optimal response of a fire situation.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent fire fighting, and particularly relates to an intelligent fire fighting method and system based on multi-modal perception. BACKGROUND

[0002] The traditional fire fighting system mainly relies on fixed spray systems, smoke detectors, manual fire extinguishers and other equipment. Although these systems have played an important role in history, they also have significant shortcomings: the fixed spray system usually adopts a one-size-fits-all water-based fire extinguishing method, which not only has poor adaptability to electrical equipment, high-value assets or specific types of fire (such as metal fire, grease fire), but also may cause secondary damage (such as circuit short circuit, damage to cultural relics, chemical reaction, etc.) due to water spraying; the smoke detector is slow to respond and prone to false positives, and cannot provide accurate fire source positioning and fire assessment; existing fire fighting robots mainly focus on remote monitoring or single fire extinguishing medium injection, lack intelligent decision-making and multi-medium switching capabilities, and are difficult to make the optimal response in complex fire environment. In addition, the traditional system lacks the ability of multi-dimensional perception of the environment (such as airtightness, personnel distribution, asset value), and the selection of fire extinguishing medium often depends on artificial experience, which is low in response efficiency and high in risk. Therefore, there is an urgent need for an intelligent fire fighting system that can integrate multi-modal perception, autonomous navigation, intelligent identification and dynamic medium selection to improve fire extinguishing efficiency, reduce secondary damage and ensure personnel safety. SUMMARY

[0003] To solve the above problems in the prior art, the application provides an intelligent fire fighting method and system based on multi-modal perception.

[0004] The object of the application can be achieved by the following technical solutions:

[0005] An intelligent fire fighting method based on multi-modal perception, the implementation of the intelligent fire fighting method based on multi-modal perception includes the following steps:

[0006] Step S1: Obtain a global map and preset a patrol route and patrol points, and the intelligent fire fighting device moves according to the patrol route and reaches the patrol points through autonomous path planning to obtain environment perception data, the environment perception data including environment airtightness, asset value density, personnel risk factor and surface complexity;

[0007] Step S2: Perform fire source identification, and when a fire source is identified, obtain fire field perception data, the fire field perception data including fire type, fire intensity index and estimated intervention time, and locate the fire source to obtain fire source positioning and map the fire source positioning to the intelligent fire fighting device coordinate system;

[0008] Step S3: The intelligent fire fighting device reaches the fire source positioning and sends a signal to the fire fighting system to cut off the power supply of the fire field and start smoke evacuation;

[0009] Step S4: selecting the best fire extinguishing medium based on the environment perception data and the fire field perception data through multi-dimensional fire extinguishing medium adaptability selection;

[0010] Step S5: extinguishing the fire through the best fire extinguishing medium, and after the fire is extinguished, confirming the fire extinguishing through fire source identification, and returning to standby state.

[0011] Preferably, the multi-dimensional fire extinguishing medium adaptability selection in the step S4 is specifically:

[0012] Step S401: constructing a fire field feature vector based on the environment perception data and the fire field perception data, and establishing a medium ontology knowledge base;

[0013] Step S402: obtaining a dynamic effective efficiency index based on the fire field feature vector and the medium ontology knowledge base;

[0014] Step S403: obtaining a residual impact index based on the asset value density and the surface complexity;

[0015] Step S404: obtaining a comprehensive survival safety index based on the environment airtightness, the personnel risk factor, the fire intensity index and the dynamic effective efficiency index;

[0016] Step S405: constructing a multi-objective optimization decision function based on the dynamic effective efficiency index, the residual impact index and the comprehensive survival safety index, and selecting the best fire extinguishing medium, the mathematical description of the multi-objective optimization decision function is , wherein, is the fire extinguishing medium score, is the dynamic effective efficiency index, is the residual impact index, is the comprehensive survival safety index, is the cost index, is the maximum value of the task priority weight, is the task priority weight; traversing all fire extinguishing media to obtain the fire extinguishing medium with the highest medium score as the best fire extinguishing medium.

[0017] Preferably, the step S401 specifically includes:

[0018] A set of pre-set fire extinguishing media is set, and each fire extinguishing medium is labeled with fire extinguishing medium characteristics, the fire extinguishing medium characteristics including fire extinguishing efficiency value, electrical conductivity risk factor, residual index, toxicity index and cost index;

[0019] constructing the fire field feature vector based on the environment perception data and the fire field perception data, and establishing the medium ontology knowledge base for each fire extinguishing medium based on the fire extinguishing medium characteristics, the mathematical description of the fire field feature vector is wherein, is the fire field feature vector, is the fire type conversion six-dimensional vector, is the environment airtightness, is the asset value density, is the personnel risk factor, is the surface complexity, is the fire intensity index, is the expected intervention time, the medium ontology knowledge base comprises the fire extinguishing medium characteristics and the dynamic response function of the fire extinguishing medium , the dynamic response function describes the decay of the fire extinguishing medium efficiency from spraying to acting on the fire source.

[0020] Preferably, the step S402 specifically comprises:

[0021] obtaining an environment-fire extinguishing medium dynamic action item according to the electrical conductivity risk factor and the environment airtightness, the mathematical description is wherein, is the electrical conductivity risk factor;

[0022] obtaining the dynamic effective efficiency index based on the fire extinguishing efficiency value, the dynamic response function and the environment-fire extinguishing medium dynamic action item, the mathematical description is wherein, is the dynamic effective efficiency index, is the fire extinguishing efficiency value, is the dynamic response function.

[0023] Preferably, the step S403 specifically comprises:

[0024] introducing a reversible recovery, the value range is [0, 1]; obtaining the residual impact index based on the reversible recovery, the mathematical description is wherein, is the residual impact index, is the residual index, is the adjustment coefficient, is the reversible recovery.

[0025] Preferably, the step S404 specifically comprises:

[0026] constructing a medium toxicity risk item (F) based on the toxicity index, the environment airtightness and the personnel risk factor, wherein, is the toxicity index, environmental tightness weight;

[0027] constructing a fire extinguishing failure risk term based on the personnel risk factor, the fire intensity index and the dynamic effective performance index ];

[0028] obtaining the comprehensive survival safety index based on the medium toxicity risk term and the fire extinguishing failure risk term, and the mathematical description is , wherein, is a comprehensive survival safety index.

[0029] An intelligent fire extinguishing system based on multi-modal perception is used to execute the intelligent fire extinguishing method based on multi-modal perception, and comprises an environment perception module, a fire source identification module, a fire extinguishing medium selection module and a fire elimination module.

[0030] The environment perception module is used to obtain a global map and preset a patrol route and a patrol point, and the intelligent fire extinguishing equipment moves according to the patrol route and reaches the patrol point through autonomous path planning to obtain environment perception data, wherein the environment perception data comprises environmental tightness, asset value density, a personnel risk factor and surface complexity.

[0031] The fire source identification module is used to identify a fire source, and when the fire source is identified, fire scene perception data is obtained, wherein the fire scene perception data comprises a fire type, a fire intensity index and an expected intervention time, and the fire source is located to obtain a fire source location, and the fire source location is mapped to an intelligent fire extinguishing equipment coordinate system; the intelligent fire extinguishing equipment reaches the fire source location and sends a signal to the fire extinguishing system to cut off the power supply of the fire scene and start smoke exhaust.

[0032] The fire extinguishing medium selection module is used to select the best fire extinguishing medium through multi-dimensional fire extinguishing medium adaptability selection based on the environment perception data and the fire scene perception data.

[0033] The fire elimination module is used to extinguish the fire through the best fire extinguishing medium, and after the fire is extinguished, the fire source identification is used to confirm that the fire is extinguished, and the system returns to a standby state.

[0034] The present application has the following advantages:

[0035] (1) Through dynamic fire intensity evaluation and fire extinguishing medium selection, the most suitable fire extinguishing medium is used at the best time, and the fire extinguishing efficiency and success rate are improved.

[0036] (2) Through quantification of residual impact, asset value and surface complexity, high-value asset loss caused by improper selection of fire extinguishing medium is avoided, and secondary damage is reduced.

[0037] (3) By comprehensively considering the toxicity risk and the fire extinguishing failure risk, the personnel living environment is preferentially ensured, and the personnel safety is enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0038] For the convenience of those skilled in the art, the present application will be further described below with reference to the accompanying drawings.

[0039] Figure 1 A step flow chart of an intelligent fire extinguishing method based on multi-modal perception. DETAILED DESCRIPTION

[0040] For a better understanding of the present application, various aspects of the present application will be described in more detail with reference to the accompanying drawings. It should be understood that these detailed descriptions are merely descriptive of illustrative embodiments of the present application, and are not intended in any way to limit the scope of the present application. Throughout the specification, the expression "and / or" includes any and all combinations of one or more of the associated listed items. As used in this document, the words "substantially", "approximately", and similar expressions are used as terms of approximation and not as terms of degree, and are intended to account for the inherent deviations in a measuring or computing process that would be recognized by those of ordinary skill in the art. In addition, in the present application, the order of the steps of the process described does not necessarily represent the order in which the processes appear in actual operation, unless there is an explicit other limitation or it can be derived from the context.

[0041] It should also be understood that expressions such as "include", "including", "have", "has", "contain" and / or "containing" and the like, are open-ended terms that are intended to mean that there are other items or components that are not expressly mentioned, but that are still within the scope of the present application. In addition, when expressions such as "at least one of" appear after a list of items, it modifies the entire list of items and not just the individual items in the list. In addition, when describing embodiments of the present application, the use of "may" indicates that one or more embodiments of the present application. And, the word "exemplary" is intended to mean example or illustrative.

[0042] Unless otherwise defined, all terms used in this document, including engineering terms and scientific terms, have the same meaning as commonly understood by those of ordinary skill in the art to which the present application belongs. It should also be understood that, unless explicitly stated otherwise in the present application, words defined in common dictionaries should be interpreted to have meanings consistent with their meanings in the context of the relevant technology, and should not be interpreted in an idealized or overly formalized sense.

[0043] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0044] Embodiment 1:

[0045] Please refer to Figure 1 An intelligent fire-fighting method based on multi-modal perception, comprising:

[0046] S1: Obtain a global map and preset a patrol route and a patrol point, move the robot according to the patrol route, and reach the patrol point through autonomous path planning to obtain environmental perception data, wherein the environmental perception data includes environmental airtightness, asset value density, personnel risk factor, and surface complexity;

[0047] S2: Identify a fire source through a deep learning algorithm (such as a YOLO algorithm) and devices such as a purple-infrared sensor, obtain fire field perception data when the fire source is identified, wherein the fire field perception data includes a fire type (A solid, B liquid, C gas, D metal, E live equipment, and F cooking oil), a fire intensity index (estimated based on flame height, temperature, and spread speed, with a value range of [0, 1]), and a predicted intervention time, simultaneously locate the fire source through the fusion of vision, laser radar, and the purple-infrared sensor to obtain a fire source location, map the fire source location to a robot coordinate system, and the predicted intervention time is the estimated time from the discovery of the fire to the start of the rescue, which is shorter when an automatic fire extinguishing system is started or a firefighter is already on the scene, and can be 0-30s, and is longer when professional fire-fighting forces need to be on the scene, and needs to be more than 120s;

[0048] S3: Drive the robot moving chassis to quickly reach the fire source location, and send a signal to the fire-fighting system to cut off the power supply of the fire field and start smoke exhaust;

[0049] S4: Based on the environmental perception data and the fire field perception data, select the best fire extinguishing medium through multi-dimensional fire extinguishing medium adaptability selection;

[0050] S5: Switch to the best fire extinguishing medium through control of a switching valve group, align the fire source through a rotatable spray head, start spray fire extinguishing, and after the fire is extinguished, confirm the extinguishment of the fire through fire source identification (YOLO, high temperature detection, combustion spectrum analysis, etc.), and return to a standby state.

[0051] In the present embodiment, the step S1 can be implemented through the following steps:

[0052] S101: Real-time locate the robot position through a laser slam algorithm, and control the robot motion trajectory through a MCP operation and control algorithm, a PID feedback control algorithm, a side offset correction algorithm, etc.

[0053] S102: The autonomous path planning is performed by an A* algorithm or a Dijkstra algorithm, and obstacle avoidance is performed by a DWA dynamic window algorithm;

[0054] S103: After reaching the inspection point, environment perception is performed by using a multi-modal sensor to obtain environment perception data, the environment perception includes but is not limited to visible light, thermal imaging, laser, ultraviolet detection, infrared detection, the environment airtightness value range is [0, 1], and the environment airtightness can be perceived by air flow or judged by experience, for example, a large open space (a factory building, an outdoor space), a large number of doors and windows are open, and the value is low; a small airtight space (a safe deposit, a device cabin, a basement), no window or a small window, and the value is high; the mathematical description of the asset value density is normalized to the interval [0, 1]; the mathematical description of the personnel risk factor is (potential personnel presence probability + environment airtightness 0.5) / 1.5, and the value range is [0, 1]; the potential personnel presence probability is the possibility of the presence of personnel in the space when a fire occurs, and the value is low in a device room, a warehouse, an unmanned workshop at night, and the value is high in a personnel-intensive place such as a dormitory, a residence, a shopping mall, a theater or a place where personnel are known to be trapped; the surface complexity is used to describe the complexity of the surface of the fire scene, and the value range is [0, 1], for example, many precision instruments and cables are 0.9, and flat walls and floors are 0.1.

[0055] In the embodiment, the step S3 can be implemented by the following steps:

[0056] With the fire source positioning as a target and the robot position as a starting point, a global path is re-planned by autonomous path planning, in the process of movement of the robot, dynamic obstacle information is detected by an obstacle avoidance sensor (such as a laser radar, an ultrasonic wave), and real-time obstacle avoidance path planning is performed by a DWA dynamic window algorithm, and after bypassing the obstacle, the robot returns to the global path.

[0057] In the embodiment, the multi-dimensional fire extinguishing medium adaptability selection is specifically:

[0058] S401: A fire scene feature vector is constructed based on the environment perception data and the fire scene perception data, and a medium ontology knowledge base is established;

[0059] S402: A dynamic effective efficiency index is obtained based on the fire scene feature vector and the medium ontology knowledge base;

[0060] S403: A residual impact index is obtained based on the asset value density and the surface complexity;

[0061] S404: obtaining a comprehensive survival safety index based on the environment tightness, the personnel risk factor, the fire intensity index, and the dynamic effective efficiency index;

[0062] S405: constructing a multi-objective optimization decision function based on the dynamic effective efficiency index, the residual impact index, and the comprehensive survival safety index, and selecting the optimal fire extinguishing medium, the mathematical description of the multi-objective optimization decision function being wherein, is a fire extinguishing medium score, is a dynamic effective efficiency index, is a residual impact index, is a comprehensive survival safety index, is a cost index, is a maximum value of a task priority weight, is a task priority weight (for example, 1-3), input by a decision maker, is an asset protection priority mode, 2 is a balance mode, 3 is a life protection priority mode, is a geometric mean, and all indexes cannot have a short board, and any index close to 0 will drag the total score close to 0, thereby vetoing the medium. The index weights of each item are different, reflecting the importance of different goals. Survival safety ( ) is given a fixed high weight ( ), highlighting its eternal highest priority. Traversing all fire extinguishing media, the fire extinguishing medium with the highest medium score is obtained as the optimal fire extinguishing medium. When the compatibility of a certain medium to the environment is extremely poor, and the residual impact is unacceptable, the residual impact index of the medium may be greater than or equal to 1, at which time the medium can be directly removed from the candidate list, or its medium score is forcibly set to 0.

[0063] In the embodiment, the step S401 can be implemented by the following steps:

[0064] S401-1: presetting a set of fire extinguishing media and marking each fire extinguishing medium with fire extinguishing medium characteristics, the fire extinguishing medium characteristics including fire extinguishing efficiency value, conductivity risk factor, residue index, toxicity index and cost index; the fire extinguishing efficiency value is the fire extinguishing ability of the fire extinguishing medium for different fire types, the value range is [0, 1], which is a benchmark value obtained from national standards or experimental data, for example, for class A fire, the fire extinguishing efficiency value of water can be set to 1.0, and the fire extinguishing efficiency value of dry powder is 0.8; the conductivity risk factor is used to judge whether the fire extinguishing medium is conductive, and the insulation (such as gas) is 0, the conductive (such as water) is 1, and the semi-insulation (such as fog water) is 0.5; the residue index is the cleaning difficulty and corrosiveness of the residue after the use of the fire extinguishing medium itself, the value range is [0, 1], no residue is 0 (such as gas), and it is difficult to clean and corrosive is 1 (such as dry powder); the toxicity index is the toxicity degree of the fire extinguishing medium itself or its decomposition product to personnel, the value range is [0, 1], no toxicity is 0, and high toxicity is 1; the cost index is a normalized index of the comprehensive unit fire extinguishing agent cost, system installation and maintenance cost, the value range is [0, 1], the lowest cost is 0, and the highest cost is 1;

[0065] S401-2: constructing the fire field feature vector based on the environment perception data and the fire field perception data, and establishing the medium ontology knowledge base for each fire extinguishing medium based on the fire extinguishing medium characteristics, the mathematical description of the fire field feature vector is , wherein, is the fire field feature vector, is the six-dimensional vector converted from the fire type (A, B, C, D, E, F), such as class A fire [1, 0, 0, 0, 0, 0], is the environment airtightness, is the asset value density, is the personnel risk factor, is the surface complexity, is the fire intensity index, is the predicted intervention time, and the medium ontology knowledge base includes the fire extinguishing medium characteristics (static attributes) of the fire extinguishing medium and the dynamic response function f decay ( , ), which describes the decay of the fire extinguishing medium efficiency from spraying to acting on the fire source, for example, for gas medium, the dynamic response function can be defined as f decay_gas = e^(-a·TTI ), a is an empirical decay constant, is the standard value, and the typical value is 120s, and for dry powder medium, it decays fast with time, and the greater the fire, the stronger the disturbance airflow, the faster the decay, so the dynamic response function can be defined as f decay_powder =e^[-a·(TTI )·(1 + )).

[0066] In this embodiment, step S402 can be implemented through the following steps:

[0067] S402-1: Based on the conductivity risk factor and the environmental airtightness, the dynamic action term of the environment-extinguishing medium is obtained, mathematically described as follows: ,in, As a conductivity risk factor, in a confined space ( Approaching 1), if the dielectric is conductive ( When the dynamic interaction term of the environment and the extinguishing medium approaches 1, it will approach 0, which will greatly inhibit the effectiveness of the extinguishing medium. However, in open spaces, the dynamic interaction term of the environment and the extinguishing medium is about 1, which has almost no effect on the effectiveness of the extinguishing medium.

[0068] S402-2: The dynamic effective efficiency index is obtained based on the fire extinguishing efficiency value, the dynamic response function, and the dynamic interaction term between the environment and the fire extinguishing medium. Mathematically, it is described as follows: ,in, It is a dynamic effectiveness index. This corresponds to the fire extinguishing efficiency value. This is a dynamic response function. Example: a large space warehouse ( Rapidly developing Class B fire (=0.2) =0.8) Select medium, expected intervention time TTI = 60 seconds. For carbon dioxide, its =0.95, =0, If ≈0.94, then its DEEI is 0.95·0.94·(1-0.2·0)≈0.89. For dry powder, its... =0.85, =0.1 (slight conductivity) If the value is approximately 0.58, then its DEEI is 0.85·0.58·(1-0.2·0.1)≈0.48. This indicates that under this dynamic scenario, the effective performance of the dry powder drops significantly to 0.48 due to attenuation and slight environmental interference, while the performance of the gas medium remains good.

[0069] In this embodiment, step S403 can be implemented through the following steps:

[0070] S403-1: In order not to look at the extinguishing medium residue in isolation, but to calculate its impact cost with the whole value system, reversible recovery is introduced, which represents whether the residue can be recovered by simple physical means (such as purging, suction) or cause permanent chemical corrosion or pollution, with a value range of [0, 1], and reversible recovery of 1 represents 100% reversibility (such as inert gas, which disappears without a trace), and reversible recovery of 0 represents almost irreversibility (such as dry powder invading precision circuits, causing permanent corrosion and short circuit);

[0071] S403-2: Based on the reversible recovery, the residual impact index is obtained, which is mathematically described as , wherein, is the residual impact index, is the residue index, is the adjustment coefficient (0~1), representing the contribution of reversibility to the reduction of residue, is the reversible recovery; Example: compared with the rescue of precision instrument room ( =0.95, =0.9) fire, new gaseous medium =0, =1, then its is 0, which can be perfectly compatible; water-based foam =0.7, =0.3 (water stains and foam residues may cause circuit board oxidation, which is difficult to recover completely), then its is 0.7·(0.95+0.9)·(1-0.7·0.3) =1.02, indicating that the impact cost of water-based foam on this environment is very high (1.02 is much larger than 0), and the system will strongly reject this option.

[0072] In this embodiment, the step S404 can be implemented by the following steps:

[0073] S404-1: Based on the toxicity index, the environmental airtightness and the personnel risk factor, a medium toxicity risk term ( ) is constructed, wherein, is the toxicity index, is the environmental airtightness weight (the higher the environmental airtightness, the higher the weight), representing that in airtight and crowded places, the toxicity risk is high;

[0074] S404-2: Based on the personnel risk factor, the fire intensity index and the dynamic effective efficiency index, a fire extinguishing failure risk term [ ] is constructed, if the dynamic effective efficiency of the medium is very low, which means that the fire may not be extinguished quickly ( very high), the fire intensity ( ) itself poses a huge threat to personnel ( ) (high temperature, lack of oxygen, structure collapse, etc.);

[0075] S404-3: obtaining the comprehensive survival safety index based on the medium toxicity risk term and the fire extinguishing failure risk term, mathematically described as , wherein, is the comprehensive survival safety index. The dual risks of fire extinguishing medium toxicity and fire extinguishing failure are integrated, forcing the system to make a trade-off when choosing, that is, sometimes choosing a slightly toxic but efficient and fast fire extinguishing medium, and the comprehensive survival safety index of the medium may be higher than that of a non-toxic but slow fire extinguishing medium.

[0076] Example: inside the cabin ( = 0.9, = 0.8) electrical fire (0.7) occurs = 0.7), for efficient HCFC gas (0.6, , ), its ; for non-toxic but less efficient inert gas (0.4, , ), its It can be seen that the efficient but slightly toxic medium HCFC can greatly reduce the threat of the fire itself, and the comprehensive survival safety index is higher than that of the non-toxic but slow inert gas.

[0077] In this embodiment, the step S5 can be implemented by the following steps:

[0078] The robot controls the switching valve group to select the corresponding fire extinguishing agent tank according to the calculated optimal fire extinguishing medium. Through visual servo algorithm and PID control, etc., the nozzle is accurately controlled to aim at the root of the fire source for spraying. After the open fire is extinguished, the thermal imager is used to continuously monitor whether the temperature has dropped below the ignition point, and the visible light camera and ultraviolet sensor are used to monitor whether there is a sign of rekindling. When all indicators confirm that the fire has been completely extinguished, the robot returns to the standby point and continues the inspection task.

[0079] Embodiment 2:

[0080] An intelligent fire extinguishing system based on multi-modal perception, comprising an environment perception module, a fire source identification module, a fire extinguishing medium selection module, and a fire elimination module;

[0081] The environment perception module is used to obtain a global map and preset an inspection route and inspection points. The intelligent fire extinguishing device moves according to the inspection route and reaches the inspection points through autonomous path planning to obtain environment perception data, which includes environment tightness, asset value density, personnel risk factor, and surface complexity.

[0082] The fire source identification module is used for fire source identification, when the fire source is identified, fire field perception data is obtained, the fire field perception data includes fire type, fire intensity index and predicted intervention time, and the fire source is located to obtain fire source positioning, and the fire source positioning is mapped to an intelligent fire fighting equipment coordinate system; the intelligent fire fighting equipment reaches the fire source positioning, and sends a signal to the fire fighting system to cut off the power supply of the fire field and start smoke exhaust;

[0083] The fire extinguishing medium selection module is used for selecting the best fire extinguishing medium based on the environment perception data and the fire field perception data through multi-dimensional fire extinguishing medium adaptability selection;

[0084] The fire elimination module is used for extinguishing the fire through the best fire extinguishing medium, after the fire is extinguished, the fire source identification is used to confirm that the fire is extinguished, and the standby state is returned.

[0085] The above is only a preferred embodiment of the present application, and does not limit the present application in any form. Although the preferred embodiment of the present application has been disclosed as above, it is not intended to limit the present application. Any person skilled in the art can make some changes or modifications to the above disclosed technical content without departing from the scope of the technical solution of the present application, and any simple modification, equivalent change and modification of the above embodiment according to the technical essence of the present application are still within the scope of the technical solution of the present application.

Claims

1. A multi-modal perception based intelligent firefighting method, characterized in that, The implementation of the intelligent fire-fighting method based on multi-modal perception includes the following steps: Step S1: Obtain a global map and preset a patrol route and a patrol point, and the intelligent fire-fighting device moves according to the patrol route and reaches the patrol point through autonomous path planning to obtain environment perception data, wherein the environment perception data includes environment airtightness, asset value density, personnel risk factor and surface complexity; Step S2: Perform fire source identification, and when a fire source is identified, obtain fire scene perception data, wherein the fire scene perception data includes fire type, fire intensity index and estimated intervention time, and the fire source is located to obtain fire source positioning, and the fire source positioning is mapped to an intelligent fire-fighting device coordinate system; Step S3: The intelligent fire-fighting device reaches the fire source positioning, and sends a signal to a fire-fighting system to cut off power supply of the fire scene and start smoke exhaust; Step S4: Based on the environment perception data and the fire scene perception data, the best fire extinguishing medium is selected through multi-dimensional fire extinguishing medium adaptability selection; Step S401: Based on the environment perception data and the fire scene perception data, a fire scene feature vector is constructed and a medium ontology knowledge base is established; Step S402: Based on the fire scene feature vector and the medium ontology knowledge base, a dynamic effective efficiency index is obtained; Step S403: Based on the asset value density and the surface complexity, a residual impact index is obtained; Step S404: Based on the environment airtightness, the personnel risk factor, the fire intensity index and the dynamic effective efficiency index, a comprehensive survival safety index is obtained; Step S405: constructing a multi-objective optimization decision function based on the dynamic effective efficiency index, the residual impact index and the comprehensive survival safety index, and selecting the best fire extinguishing medium, the mathematical description of the multi-objective optimization decision function is wherein, is the score of the fire extinguishing medium, is the dynamic effective efficiency index, is the residual impact index, is the comprehensive survival safety index, is the cost index, is the maximum value of the task priority weight, is the task priority weight; all fire extinguishing media are traversed to obtain the fire extinguishing medium with the highest medium score as the best fire extinguishing medium; Step S5: The fire is extinguished by the best fire extinguishing medium, and after the fire is extinguished, the fire is confirmed to be extinguished through fire source identification, and the system returns to a standby state.

2. The multi-modal perception based intelligent firefighting method as claimed in claim 1, wherein, The step S401 specifically includes: A set of fire extinguishing media is preset, and each fire extinguishing medium is labeled with fire extinguishing medium characteristics, including fire extinguishing efficiency value, electrical conductivity risk factor, residual index, toxicity index and cost index; constructing the fire field feature vector based on the environment perception data and the fire field perception data, and establishing the medium ontology knowledge base for each fire extinguishing medium based on the fire extinguishing medium characteristics, the mathematical description of the fire field feature vector is wherein, is the fire field feature vector, is the six-dimensional vector of fire type conversion, is the environment tightness, is the asset value density, is the personnel risk factor, is the surface complexity, is the fire intensity index, is the expected intervention time, the medium ontology knowledge base contains the fire extinguishing medium characteristics and dynamic response function of the fire extinguishing medium , the dynamic response function describes the decay of the fire extinguishing medium efficiency from spraying to acting on the fire source.

3. The multi-modal perception based intelligent firefighting method as claimed in claim 2, wherein, The step S402 specifically includes: An environmental-extinguishing medium dynamic action term is obtained from the electrical conductivity risk factor and the environmental tightness, mathematically described as wherein is the electrical conductivity risk factor; The dynamic effective efficiency index is obtained based on the fire extinguishing efficiency value, the dynamic response function and the environment-extinguishing medium dynamic action term, and is mathematically described as wherein, is the dynamic effective efficiency index, is the fire extinguishing efficiency value, is the dynamic response function.

4. The multi-modal perception based intelligent firefighting method as claimed in claim 3, wherein, The step S403 specifically includes: A reversible recoverability is introduced, with a value range of [0, 1]; the residual impact index is obtained based on the reversible recoverability, and is mathematically described as wherein, is a residual impact index, is a residual index, is an adjustment coefficient, is a reversible recoverability.

5. The multi-modal perception based intelligent firefighting method as claimed in claim 4, wherein, The step S404 specifically includes: constructing a media toxicity risk term based on the toxicity index, the environmental containment, and the personnel risk factor wherein, is a toxicity index, is an environmental containment weight; constructing a fire extinguishing failure risk term based on the personnel risk factor, the fire intensity index, and the dynamic effective performance index ] The integrated survival safety index is obtained based on the medium toxicity risk term and the fire extinguishing failure risk term, and is mathematically described as wherein, is the integrated survival safety index.

6. A multi-modal perception based intelligent fire fighting system characterized in that, The system is applied to the intelligent fire-fighting method based on multi-modal perception as claimed in any one of claims 1-5, and includes an environment perception module, a fire source identification module, a fire extinguishing medium selection module and a fire situation elimination module; The environment perception module is used to obtain a global map and preset a patrol route and a patrol point, and the intelligent fire-fighting device moves according to the patrol route and reaches the patrol point through autonomous path planning to obtain environment perception data, wherein the environment perception data includes environment airtightness, asset value density, personnel risk factor and surface complexity; The fire source identification module is used to perform fire source identification, and when a fire source is identified, obtain fire scene perception data, wherein the fire scene perception data includes fire type, fire intensity index and estimated intervention time, and the fire source is located to obtain fire source positioning, and the fire source positioning is mapped to an intelligent fire-fighting device coordinate system; the intelligent fire-fighting device reaches the fire source positioning, and sends a signal to a fire-fighting system to cut off power supply of the fire scene and start smoke exhaust; The fire extinguishing medium selection module is configured to select an optimal fire extinguishing medium through multi-dimensional fire extinguishing medium adaptability selection based on the environment sensing data and the fire scene sensing data. The fire situation elimination module is configured to extinguish the fire through the optimal fire extinguishing medium, and after the fire is extinguished, confirm the fire extinguishing through fire source identification, and return to the standby state.

Images