Intelligent fire-fighting unmanned aerial vehicle cluster system

Abstract

When the intelligent fire-fighting unmanned aerial vehicle cluster system works, the method comprises the following steps that 1, an intelligent fire-fighting unmanned aerial vehicle cluster detects a fire scene and searches a water source near the fire scene, and an optimal route is searched between the water source and the fire scene to lay a pressurizing water pump and a fire hose; 2, the intelligent fire-fighting unmanned aerial vehicle cluster controls a fire-fighting nozzle to conduct fire extinguishing operation, and the positions of a pressurizing water pump and a fire hose are adjusted according to fire behavior changes of a fire scene; and 3, the intelligent fire-fighting unmanned aerial vehicle cluster performs handover control on a fire-fighting nozzle, a pressurizing water pump and a fire-fighting hose in turn, and fuel oil is supplemented to the pressurizing water pump until fire extinguishing operation is finished. A fire-fighting water supply channel can be laid between a water source and a fire scene at an ultra-fast speed, the fire extinguishing operation speed is high, manual intervention is not needed, and nearly infinite fire-fighting water can be continuously provided.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and more particularly to fire rescue, drones and artificial intelligence technologies, specifically to an intelligent fire-fighting drone swarm system. Background Technology

[0002] Fires of all kinds cause countless personal and property losses worldwide every year. Common firefighting methods include cooling, asphyxiation, isolation, and suppression, using equipment such as fire water tanks, fire hydrants, fire sandboxes, portable fire extinguishers, fire trucks, and firefighting aircraft.

[0003] Intelligent firefighting drones, as an emerging technology application in the field of emergency rescue, have made significant progress in recent years. For example, CN119987431A discloses a high-order distributed firefighting drone task allocation method and system. Its technological development is mainly reflected in the following aspects: Current intelligent firefighting drones have integrated a variety of advanced functions, such as thermal imaging, high-definition video transmission, high-altitude fire extinguishing bomb dropping, and on-site reconnaissance. These functions enable drones to respond quickly in complex environments, providing valuable information to firefighters and reducing the risk of casualties; technological advancements have significantly improved the flight time and payload capacity of intelligent firefighting drones. Some new drones can carry sufficient fire extinguishing agents for extended firefighting operations. For example, Walkera's "Motorcycle-Linked High-Rise Firefighting and Rescue System" supports ultra-high-altitude operations up to 170 meters and can achieve two hours of uninterrupted firefighting. For example, Zhuoyi Intelligent's drone system can cover various tasks such as high-rise building firefighting and forest fire patrol.

[0004] Despite significant progress in fire and rescue technology in recent years, some problems still urgently need to be addressed in practical applications: 1. Water Supply Difficulties. Water resources are scarce in remote areas, and traditional water supply methods are insufficient to meet demand. For example, in virgin forests, technologies such as drone reconnaissance and water pump relays are needed to overcome water supply challenges. While drones can overcome terrain obstacles for long-distance water supply, their power limitations prevent them from carrying large amounts of water, and the round trips to refuel are extremely time-consuming, leaving them helpless in the face of large fires. Even large-capacity fire trucks or firefighting aircraft have extremely limited water capacity. Each fire hose sprays water at a rate of 1.5 meters per second. Medium-sized fire trucks carry 5-8 tons of water, and large fire trucks carry 12-30 tons. In actual fire scenes, due to pressure reduction, the water load typically lasts only about 30-40 minutes for continuous firefighting operations. Therefore, the water capacity of fire trucks is often insufficient for large fires. Firefighting helicopters drop about 4 tons of water per mission, and large firefighting aircraft carry about 6 tons per mission, far less than the capacity of large fire trucks, requiring repeated trips between water sources and the fire scene.

[0005] 2. Rapid fire spread. Forest fires are often accompanied by strong winds, which propel the fire rapidly. Densely vegetated areas are prone to three-dimensional burning, making it difficult for rescue personnel to approach. Furthermore, the stability and reliability of drones in adverse weather conditions need improvement; drones also suffer from poor accuracy in delivering extinguishing agents and limited effectiveness, as they are hesitant to fly too close to the fire, hindering the optimization of extinguishing precision and effectiveness. Firefighting aircraft typically fly at altitudes of around 60 meters, making it difficult to guarantee extinguishing accuracy. For example, drones carrying water or foam extinguishing agents may struggle to accurately spray the fire from high-rise buildings.

[0006] 3. Complex terrain. Mountainous terrain is rugged and difficult to access, making it hard for fire trucks and large equipment to reach, especially in virgin forest areas where steep slopes and difficult roads hinder conventional firefighting methods. In complex urban environments and high-rise building fires, firefighters often have to enter the scene to extinguish the fire, which also poses a significant threat to their personal safety. The autonomous flight capabilities of drones in complex terrain areas still need improvement. For example, drones may not be able to make autonomous judgments and fly when entering building interiors or facing complex obstacles.

[0007] 4. Challenges in Special Scenarios. Lightning strikes can ignite multiple fires under extreme weather conditions, requiring firefighters to address challenges such as difficulty approaching fires on cliff edges and treacherous pine forests. Fires involving new energy sources, such as lithium batteries, can experience high reignition rates due to thermal runaway, necessitating large amounts of water and specialized equipment for extinguishing. Fires in historical buildings are structurally complex with abundant flammable materials, and limited fire access routes further complicate rescue efforts. The type and method of extinguishing agents must be optimized based on the fire type. Data processing and real-time performance are also insufficient. While drones can transmit data in real time, the speed of data processing and analysis in complex fire scenes needs improvement. For example, AI algorithms may be affected by environmental factors when identifying fire points and trapped individuals, leading to an increased misjudgment rate. Summary of the Invention

[0008] To address the technical problems mentioned in the background section, the present invention proposes an intelligent firefighting drone that combines fire extinguishing materials, drones, and artificial intelligence technology to provide an intelligent firefighting and rescue platform, thereby further improving fire rescue efficiency.

[0009] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A smart firefighting drone swarm system includes smart firefighting drones and a fire command center. The smart firefighting drones are equipped with grappling hooks for gripping pressurized water pumps, fire hoses, and fire nozzles. When the system is in operation, it includes the following steps: Step 1: The intelligent fire-fighting drone swarm reconnoiters the fire scene and searches for water sources near the fire scene, and searches for the optimal route between the water source and the fire scene to lay pressurized water pumps and fire hoses; Step 2: The intelligent fire-fighting drone swarm controls the fire nozzles to carry out fire extinguishing operations, and adjusts the position of the booster pump and fire hose according to the changes in the fire situation at the fire scene; Step 3: The intelligent fire-fighting drone swarm takes turns controlling the fire nozzles, booster pumps, and fire hoses, and replenishes fuel to the booster pumps until the fire-fighting operation is completed.

[0010] Step 1 includes the following sub-steps: Sub-step 1-1: The intelligent fire-fighting drone cluster receives a fire alarm and goes to the fire scene to investigate and determine the fire scene information and surrounding water source information; Fire scene information includes the geographical location, area, height, distribution of ignition points, residential situation, distribution of combustibles, wind direction, wind force, and estimates of fire intensity and development trend. Information on surrounding water sources includes the geographical location of the water source, the route from the fire site, its area, total water volume, height, and environmental conditions. Sub-steps 1-2: Based on the reconnaissance of the fire scene and surrounding water source information, the intelligent fire-fighting drone cluster determines how many fire water supply channels need to be laid, calculates the type and quantity of pressurized water pumps, fire hoses or fire extinguishing materials to be carried, and calculates the number of intelligent fire-fighting drones to be dispatched. Sub-steps 1-3 involve the intelligent firefighting drone swarm using artificial intelligence algorithms to search for the optimal route and firefighting plan between the water source and the fire scene; Sub-steps 1-4: The intelligent fire-fighting drone swarm lays pressurized water pumps and fire hoses along the optimal route; Step 1-1 includes the following steps: Sub-step 1-1-1: The intelligent fire-fighting drone cluster receives a fire alarm, obtains the approximate geographical location of the fire alarm, dispatches several intelligent fire-fighting drones, activates the GPS module, and goes to the geographical location to reconnoiter and determine the fire scene information and surrounding water source information. Sub-step 1-1-2: After arriving at the geographical location, the dispatched intelligent fire-fighting drones use artificial intelligence algorithms to automatically divide the patrol area and turn on infrared cameras to conduct full-area patrol of the entire fire scene. In sub-step 1-1-3, several intelligent firefighting drones are dispatched. After arriving at the geographical location, they use artificial intelligence algorithms to automatically divide the patrol area and conduct a full-area patrol of the surrounding water sources. During the patrol, they turn on the infrared cameras to continuously take infrared photos. Furthermore, they use target recognition algorithms to identify targets in the infrared photos and calculate the geographical location of the water source, the route from the fire scene, the area, the total water volume, the height, and the environmental conditions of the water source based on the GPS module positioning data at the corresponding time.

[0011] Step 1-1-2 includes the following steps: Sub-step 1-1-2-1: The intelligent fire-fighting drone cluster divides the entire fire scene into patrol areas of equal number and similar size based on the number of dispatched drones, and uses a random algorithm to assign several intelligent fire-fighting drones to patrol different patrol areas. Sub-step 1-1-2-2: During the inspection, the intelligent computer continuously captures infrared photos of the fire scene using infrared cameras, and uses microphones and speakers to collect sounds from the fire scene. Furthermore, an adaptive mean filtering algorithm is used to reduce noise in the infrared photos or videos of the fire scene acquired by the infrared cameras. The formula is: ; in It is a pixel The domain pixels, The number of pixels in the window's neighborhood. For window neighborhood, It is the grayscale value of the pixels in the infrared photos or videos of the fire scene; Sub-steps 1-1-2-3 involve the intelligent computer segmenting infrared images or video frames from the fire scene. Threshold segmentation of the color space is performed based on pixel hue (H), saturation (S), and brightness (V), using the following formula: ; Sub-steps 1-1-2-4 involve the intelligent computer filtering the sound signals collected by the microphone and speaker, extracting features, and extracting combustion sound features and distress call features related to the ignition point. The formula is as follows: ; in, This represents the energy of the m-th filter bank. Indicates the number of filters. These are the MFCC coefficient numbers of the sound signals collected by the microphone and speaker; Sub-step 1-1-2-5: The intelligent computer uses a threshold detection algorithm to calculate areas with high temperatures in the infrared photos of the fire scene, such as those greater than 100℃, and marks them as ignition points. Furthermore, threshold edge detection is performed on the denoised image to locate the ignition point. The formula for calculating the center of the ignition point is as follows: ; in, It is a comprehensive window coefficient used to calculate the gray value of the center pixel in a weighted manner, for example, to determine whether it is greater than 100°C; Sub-steps 1-1-2-6: The intelligent computer uses GPS module positioning data at the corresponding time, combined with the characteristics, propagation direction, and intensity of the sound signal, to assist in determining the location of key points within the fire area. Furthermore, calculate the geographical location, area, height, and distribution of fire points at the fire scene, including all ignition points. Sub-step 1-1-2-7: If the intelligent computer detects an area with a high temperature at the edge of the infrared photo or video of the fire scene, i.e. the ignition point, it expands the patrol range in that direction until no area with a high temperature is detected at the edge of the infrared photo of the fire scene taken during the patrol. Sub-step 1-1-2-8: The intelligent computer uses a target recognition algorithm to identify targets in the infrared photos of the fire scene, calculates the living conditions of people, the distribution of combustibles, wind direction, and wind force; further, based on the target identification and living conditions of people at the fire scene, it calculates whether it is a hazardous chemical warehouse or production area, or whether it is an explosive material warehouse or production area. Sub-steps 1-1-2-9: The intelligent computer estimates the fire spread speed and development trend based on the geographical location, area, height, distribution of ignition points, distribution of combustibles, wind direction, and wind force of the fire scene. Furthermore, the rate of fire spread is estimated based on the properties of the combustible material, the building structure, and the ventilation conditions at the fire scene, using the following formula: ; in, To the speed of the fire's spread, The constant is related to the properties of the combustible material. The wind speed is [value], and the area of ​​the fire scene is [area]. The height is ; Sub-step 1-1-2-10: Based on the estimated fire spread rate, the intelligent computer estimates the fire development trend and calculates the time required for the fire to reach the critical location. The key locations mentioned are densely populated residential areas, areas with high population density, areas containing flammable, explosive, or hazardous chemicals, or other important areas requiring attention, such as bridges, tunnels, schools, and military zones; the formula for the fire development trend is: ; in, It is the distance from the ignition point to the critical location. This indicates the speed at which the fire spreads.

[0012] Steps 1-2 include the following steps: Sub-step 1-2-1: The intelligent fire-fighting drone swarm estimates the time required to lay fire-fighting water supply channels between the fire site and the water source based on the reconnaissance information of the fire scene and surrounding water sources. It also estimates the fire intensity and development trend after the estimated time for laying the fire-fighting water supply channels, based on sub-step 1-1-2, and calculates the water consumption and the number of fire-fighting water supply channels required. Further, the fire-fighting water supply channel refers to all fire hoses, booster pumps, and fire nozzles continuously laid from the water source to the fire scene; one fire-fighting water supply channel can independently complete one fire-fighting water supply operation. Furthermore, smaller fires require only a smaller water consumption and one fire-fighting water supply channel, while larger fires require a larger water consumption and multiple fire-fighting water supply channels. Furthermore, the effective coverage area of ​​one fire water supply channel or fire extinguishing material is... The calculation formula is: ; in, It is the altitude at which intelligent firefighting drones perform their missions. It refers to the spray angle of the fire hose or extinguishing material. Sub-step 1-2-2: Based on the reconnaissance information of the fire scene and surrounding water sources, the intelligent fire-fighting drone swarm estimates the total length required to lay a fire water supply channel between the fire scene and the water source. Based on this total length, it calculates the actual required length of fire hoses, reserving a 20%-30% margin. Preferably, before a fire occurs, the fire hoses are assembled in 100-meter sections to reduce assembly time in case of a fire. Further, based on the actual required length of the fire hoses, it calculates the type and quantity of booster pumps to be carried. Preferably, each 10-meter section... A 0-meter fire hose is equipped with a booster pump; further, based on the target identification in sub-step 1-1-2, including the distribution of ignition points, combustible material distribution, hazardous chemical warehouses or production areas, and explosive material warehouses or production areas, the type and quantity of fire extinguishing materials to be carried are calculated; firefighters work together to assemble the fire hoses and booster pumps on the ground according to the required quantity and length, and load the required type and quantity of fire extinguishing materials onto the corresponding intelligent fire-fighting drones. Multiple fire water supply channels are appropriately separated, and one fire nozzle is equipped at the end of each fire water supply channel. Furthermore, based on the properties of the obtained combustible material, the type of extinguishing material to be carried is determined, and the total demand for extinguishing material is estimated, using the following formula: ; in, The standard for extinguishing material dosage per unit area is given, and the fire scene area is... , It is the density of the combustible material. It is the extinguishing agent efficiency coefficient; Sub-steps 1-2-3: The intelligent fire-fighting drone cluster calculates the number of intelligent fire-fighting drones to be dispatched. Preferably, 1-2 intelligent fire-fighting drones are equipped for every 100 meters of fire hose, 1-2 intelligent fire-fighting drones are equipped for every booster pump, 1 intelligent fire-fighting drone is equipped for every set of fire extinguishing materials, and 1-2 intelligent fire-fighting drones are equipped for every fire nozzle. Furthermore, the formula for estimating the number of intelligent firefighting drones that need to be dispatched is: ; in, This is the safety redundancy factor; the fire scene area is... One fire water supply channel or fire extinguishing material effectively covers an area of , This indicates rounding up to the nearest integer.

[0013] In steps 1-3, the following sub-steps are used: Sub-step 1-3-1: The intelligent firefighting drone swarm, through its respective intelligent computers and remote communication modules, forms a distributed edge computing system. It shares the geographical locations of water sources and the fire scene, as well as the current geographical locations of each intelligent firefighting drone. It then models the firefighting operation task, encoding flight routes into feasible solutions. The locations from the fire command center, firefighting materials, water sources, to the fire scene are divided into grids and fixed coordinates. The grid set of the firefighting operation area is then used... To indicate, the first Each grid is used To indicate, Maximum number of grid cells in the firefighting operation area; intelligent firefighting drones Firefighting operation grid set accessed via flight path To indicate; Sub-steps 1-3-2 are for coordinating the dispatch of... Task allocation among intelligent firefighting drones: The intelligent firefighting drone swarm employs an improved K-means algorithm, which divides the set of flight paths for the firefighting operation area into... Each cluster serves as a set of firefighting operations, i.e. This includes transporting fire extinguishing materials, transporting fire hoses and booster pumps, operating fire nozzles, and connecting the water source and the fire scene using fire hoses and booster pumps; the fire extinguishing operations of each cluster are independent of each other, without overlap or intersection, and must follow specific constraints: ; This refers to two different intelligent firefighting drones; Sub-step 1-3-3, in the... A collection of firefighting operations by intelligent firefighting drones In this context, the intelligent firefighting drone swarm defines an extended set. This set is composed of elements indexed by... Indicates the first Firefighting operation clusters in grids Collection of fire scene locations For each intelligent firefighting drone's firefighting operation cluster Construct a firefighting operation plan It consists of the locations of the fire scene. and edge set Composition; edge set It includes a collection of information from the fire command center, fire extinguishing materials, water sources to the location of the fire scene. All possible flight routes , , and This indicates the grid locations along the flight path, including fire command centers, fire extinguishing materials, water sources, and other fire scene locations. yes China is different The location of the fire scene; Sub-steps 1-3-4, firefighting operation diagrams for each intelligent firefighting drone. In the middle, each edge , Each corresponds to a non-negative weight. This weight indicates the location of the intelligent firefighting drone within the two grids at the fire scene. and The distance between the flight paths is calculated using the following formula: - ; Sub-steps 1-3-5: Setting up the intelligent firefighting drone cluster, ensuring that the intelligent firefighting drone n maintains a fixed altitude during flight. and constant flight speed; furthermore, in intelligent firefighting drone firefighting operations grouping In the context of the communication link between the intelligent firefighting drone n and the fire command center, the channel power gain can be expressed as: , The power gain that the channel can provide at a distance of 1 meter from the reference point; when the fire command center is operating at constant power For data transmission, its data transmission rate on the link can be expressed as: ; Here, B represents the bandwidth that can be used during communication; This refers to the power consumed by the intelligent firefighting drone when transmitting data; This is the power gain that the channel can provide at a distance of 1 meter from the reference point; This represents the power of the noise in the channel; Sub-steps 1-3-6: Setting up the intelligent firefighting drone It will travel at a fixed speed during flight. The formula for the propulsion power generated during flight is: ; in, The power generated by the outer contour of the blade, and This represents the power generated by the induction effect; This refers to the speed at the tip of the rotor blades. It is the average speed sensed by the rotor while it is hovering; It is the ratio between fuselage drag and overall drag. Represents the density of air. This reflects the robustness of the rotor, while Y represents the area occupied by the rotor disk; Sub-steps 1-3-7, when the intelligent firefighting drone In the selected route grid During flight, intelligent computers calculate the energy consumption generated.

[0014] ; Among them, intelligent firefighting drones Flight time ; For intelligent firefighting drones Fixed flight speed; Sub-step 1-3-8: To improve firefighting efficiency, the "firefighting information age" indicator is introduced to measure the firefighting timeliness of the intelligent firefighting drone swarm. This measures the total time elapsed from when the intelligent firefighting drone swarm begins capturing infrared photos or videos of the fire scene to when it arrives at the scene and begins firefighting operations. This includes the time spent capturing infrared photos or videos, analyzing those photos or videos, route planning, and flight time. The firefighting information age accurately represents the total time from the discovery of the fire scene to the current firefighting operation, thus providing a more precise description of firefighting efficiency. Furthermore, the average information age of all locations within the intelligent firefighting drone swarm is taken to further evaluate the overall firefighting efficiency of the system. The formula for calculating the average information age of these locations is: ,

[0015] in, This refers to a cluster of intelligent firefighting drones for firefighting operations. The total number of fire scene locations included in the data; It is used to represent intelligent firefighting drones. From the firefighting operation location Fly to On the way, from the moment we left the fire command center, we arrived A parameter representing the number of all locations visited before the current location; by sub-process For example, we can get ; Sub-steps 1-3-9: The intelligent firefighting drone swarm needs to achieve a balance between the average age of firefighting information and the energy consumption of the intelligent firefighting drones. The optimal route selection formula for the rational planning of the intelligent firefighting drones' flight paths is as follows: ; in, It is a binary variable whose value can only be 0 or 1, representing the intelligent firefighting drone. Have you selected a route? , Specifically, when This means that intelligent firefighting drones We did indeed travel this route; at the same time, Representative of intelligent firefighting drones Access the total number of fire scene locations, and group all such binary variables together and label them as follows. ; Sub-steps 1-3-10 are to ensure the fire extinguishing operation diagram. Each fire scene location in the system has the same in-degree, and this value is set to 1. The intelligent fire-fighting drone swarm applies an in-degree constraint to the route selection formula: ; Sub-step 1-3-11, to ensure the diagram Each fire scene location in the system has the same out-degree, and this value is set to 1. The intelligent fire-fighting drone swarm applies out-degree constraints to the route selection formula: ; Sub-steps 1-3-12 are to ensure that the intelligent firefighting drone... Do not select route , hour hour, When the intelligent firefighting drone chooses this route... hour Sub-journey constraints are applied to the route selection formula: , ,

[0016] In sub-step 1-3-13, the intelligent firefighting drone swarm checks whether the solution results meet the predetermined error threshold or solution time limit. If not, it returns to sub-step 1-3-1 to continue calculation until the optimal feasible solution, i.e., an optimal flight route, is obtained. The optimal flight route This refers to the intelligent firefighting drone swarm departing from its current location and arriving at its flight path. Upon reaching the destination, they laid one or more fire water supply channels between the water source and the fire scene, using the fire hoses, booster pumps, fire nozzles, and other necessary fire extinguishing materials they carried.

[0017] Sub-steps 1-3-14: The intelligent firefighting drone swarm each follows its optimal flight path. Fly to the fire scene.

[0018] In steps 1-4, the intelligent fire-fighting drone swarm uses grappling hooks to load one or more fire water supply channels that have been assembled in sub-step 1-2-2, including fire hoses, booster pumps, fire nozzles, and the necessary fire extinguishing materials, according to their division of labor. After all loading is completed and confirmed to be correct, they take off simultaneously and follow the optimal route to lay booster pumps and fire hoses between the water source and the fire scene.

[0019] Step 2 includes the following sub-steps: Sub-step 2-1: The intelligent fire-fighting drone cluster calculates the ignition point information and the optimal fire extinguishing plan based on infrared photos, videos or sounds of the fire scene, and calculates the optimal spraying method of the fire nozzles or the release method of the fire extinguishing materials. Sub-step 2-2: The intelligent fire-fighting drone cluster activates all pressurized water pumps and controls fire nozzles to spray water for fire extinguishing operations, or releases fire extinguishing materials for fire extinguishing operations, according to the optimal fire extinguishing plan. Sub-steps 2-3: The intelligent fire-fighting drone cluster adjusts the positions of the booster pumps and fire hoses according to changes in the fire situation at the fire scene; Sub-step 2-1 includes the following steps: Sub-step 2-1-1: The intelligent fire-fighting drone cluster activates its infrared cameras to continuously capture infrared photos, videos, or sounds of the fire scene, calculates the ignition point information and distribution, and uses artificial intelligence algorithms to rate the hazard level of each ignition point. Furthermore, it predicts the development trend of the fire at each ignition point. If it is likely to develop into a larger-scale fire, cause greater danger, or be more difficult to control in the short term, it is considered high-risk; otherwise, it is considered low-risk. Sub-step 2-1-2: Based on the distribution of fire points and the hazard rating of each fire point, the intelligent fire-fighting drone cluster calculates the optimal fire extinguishing plan according to the principle of extinguishing all fire points in the shortest time. Based on the location of the fire point and the fire scene environment, the intelligent computer calculates the optimal hovering position for controlling fire nozzles or intelligent fire-fighting drones carrying fire extinguishing materials, i.e., the optimal fire extinguishing plan, so that it can cover the fire point and effectively extinguish the fire. The formula for the optimal hovering position of the drone is: ; in, q represents the location of the fire points within the fire scene, and q represents the number of fire points. Sub-step 2-1-3: Based on the location of one or more fire water supply channels, the intelligent fire-fighting drone cluster calculates the optimal spraying method of the fire nozzles or the release method of the fire extinguishing materials according to the optimal fire extinguishing plan; furthermore, high-risk fire points should be extinguished first, low-risk fire points can be extinguished second best, and fire points where water spraying is ineffective should be considered for releasing other fire extinguishing materials. Furthermore, the intelligent computer calculates the optimal hovering height of the intelligent firefighting drone based on the effective range of the fire nozzles or extinguishing materials, using the following formula: ; in, The height of the ignition point, The minimum safe flight altitude for intelligent firefighting drones, The effective range height of the fire extinguishing equipment; Furthermore, the intelligent computer uses the relative position and altitude of the intelligent firefighting drone to the fire point. The intelligent computer calculates the optimal spray angle for fire-fighting water or extinguishing materials using the following formula: .

[0020] Sub-step 2-2 includes the following steps: Sub-step 2-2-1: The intelligent fire-fighting drone cluster sends a remote start command to all pressurized water pumps through the short-range communication module. After receiving the remote start command, the pressurized water pumps start running. Sub-step 2-2-2: The intelligent fire-fighting drone cluster controls the fire nozzles to spray water for fire-fighting operations according to the optimal fire-fighting plan; furthermore, since the fire nozzles have a large spray force, 2-3 fire-fighting drones can be dispatched to control the fire nozzles and the fire hoses connected to their rear ends, so as to better control the effective range height and the optimal spray angle. Furthermore, based on the relative position and altitude of the intelligent firefighting drone to the fire point, the intelligent computer calculates the flow rate of fire-fighting water or fire extinguishing materials using the following formula: ; in This refers to the amount of fire water or fire extinguishing materials required per unit area. Area of ​​the ignition point; Sub-step 2-2-3: The intelligent fire-fighting drone cluster checks for any fire points where water spraying is ineffective. If any fire is found, other fire-fighting materials should be considered for fire-fighting operations, including fire sand, carbon dioxide extinguishing agent, solid dry ice, chemical flame-retardant materials, window-breaking fire extinguishing grenade launchers and fire extinguishing grenades, ball throwers, barrel throwers, or other customized fire-fighting equipment. Sub-step 2-3-1: The intelligent fire-fighting drone cluster activates infrared cameras to continuously capture infrared photos of the fire scene, uses artificial intelligence algorithms to calculate the changes in fire intensity and the distribution of ignition points, and assigns a hazard rating to each ignition point. On the one hand, the fire in the vicinity may be getting smaller and smaller, and water spraying may not be able to reach the ignition points further away. On the other hand, the fire in the vicinity may be getting bigger and bigger, endangering the pressurized water pump, fire hoses, and intelligent fire-fighting drones. Sub-step 2-3-2: The intelligent fire-fighting drone cluster recalculates the optimal fire extinguishing plan based on the distribution of fire points; Furthermore, referring to sub-steps 1-3, the intelligent fire-fighting drone swarm updates the distribution of fire points and re-searches for the optimal route and optimal fire-fighting plan between the water source and the fire scene; In sub-step 2-3-3, the intelligent fire-fighting drone cluster adjusts the positions of the booster pump and fire hose according to the recalculated optimal fire-fighting plan.

[0021] Step 3 includes the following steps: Sub-step 3-1: The intelligent fire-fighting drone cluster checks the remaining power of each drone. The drone with sufficient power takes over the fire-fighting drone with insufficient power to continue the fire-fighting operation, and controls the fire nozzles, pressurized water pumps and fire hoses. The drone with insufficient power searches for the optimal route to return to the charging station. Sub-step 3-2: The intelligent fire-fighting drone cluster checks the remaining fuel of the booster pump. If the fuel is insufficient, it carries the corresponding grade of fuel to replenish the booster pump. Sub-step 3-3: The intelligent fire-fighting drone cluster checks the remaining quantity of fire extinguishing materials. If the fire extinguishing materials are insufficient, it carries the corresponding quantity of fire extinguishing materials to replenish the fire scene. Sub-steps 3-4: The intelligent fire-fighting drone cluster uses infrared cameras to check if there are any remaining fire points at the fire scene. If not, the fire-fighting operation is considered to be over; otherwise, return to step 2 to continue the fire-fighting operation. Step 3-1 includes the following steps: Sub-step 3-1-1: The intelligent fire-fighting drones check their remaining battery power and share the remaining battery power information with the intelligent fire-fighting drone cluster through the remote communication module; In sub-step 3-1-2, the intelligent fire-fighting drone cluster checks for drones with insufficient remaining power, such as those with less than 20% remaining power. It measures the location of the drone's GPS module and broadcasts its remaining power and location to the entire network. Further, temporarily idle fire-fighting drones with sufficient power go to the location and take over the fire-fighting operation from the drones with insufficient power, controlling the fire nozzles, pressurized water pumps, and fire hoses. If the fire-fighting drone with insufficient power is carrying a fire water tank or fire extinguishing equipment box and fire extinguishing materials, then the temporarily idle fire-fighting drone with sufficient power carries the same type and quantity of fire water tank or fire extinguishing equipment box and fire extinguishing materials to the location and take over the fire-fighting operation from the drones with insufficient power. Sub-step 3-1-3: After the fire-fighting drone with sufficient power arrives and takes over the corresponding fire-fighting operation, the fire-fighting drone with insufficient power will search for the optimal return route and return to charge along the optimal route. Sub-step 3-1-3-1: The intelligent computer of the fire-fighting drone, which is running low on power, calculates the Euclidean distance from the fire scene to the fire command center. The formula is: ; in, and These are the coordinates of the intelligent firefighting drone at the fire scene and the fire command center (300), respectively. Sub-step 3-1-3-2: The intelligent computer of the fire-fighting drone with insufficient power obtains the remaining power of the intelligent fire-fighting drone. The system also performs dynamic return-to-home battery threshold calculations on the intelligent firefighting drone, determining whether the remaining battery power is sufficient for a successful return. The formula is as follows: ; in, This indicates the safety margin threshold, which defaults to 15%. Indicates the power consumption coefficient per unit distance; If the remaining battery power of the intelligent firefighting drone is sufficient to return to base, proceed to sub-step 3-1-3-3; otherwise, if it cannot return to base, it will remain in place and maintain communication, but will no longer undertake any firefighting operations. Sub-step 3-1-3-3: The intelligent computer of the fire-fighting drone with insufficient power plans the return route of the drone based on the remaining power using a route planning algorithm. The route cost function is: ; in, Indicates cumulative energy consumption. Represents the cost of heuristic distance. , These are weighting coefficients, and they have... + ; Sub-step 3-1-3-4: The intelligent computer of the firefighting drone with insufficient power simultaneously applies power consumption constraints to the above functions, as shown in the formula: ; in Indicates the first The power function of the route segment, Indicates the first Flight speed of the segment route, Indicates the first The length of the route segment Indicates the safety margin threshold; In sub-step 3-1-3-5, the intelligent computer of the firefighting drone with insufficient power checks whether the solution result meets the predetermined error threshold or solution time limit. If not, it returns to sub-step 3-1-3-1 to continue calculation until the optimal feasible solution, i.e., an optimal flight route, is obtained. ; Sub-step 3-1-3-6: The intelligent firefighting drone with insufficient power follows the optimal flight path solved by the intelligent computer. return.

[0022] Step 3-2 includes the following steps: Sub-step 3-2-1: During the firefighting operation, the pressurized water pump continuously monitors the remaining fuel level and reports it to the nearest intelligent firefighting drone via a short-range communication module. The intelligent firefighting drone measures its current geographical location via a GPS module. Sub-step 3-2-2: When the pressurized water pump detects that the remaining fuel level is lower than the threshold (e.g., 10%), it will alarm. The nearest intelligent fire-fighting drone will report the geographical location measured by the GPS module, and the nearest intelligent fire-fighting drone carrying the corresponding fuel grade will go to the location of the pressurized water pump to replenish it. In sub-step 3-2-3, the intelligent fire-fighting drone carrying the corresponding grade of fuel aligns the fuel nozzle of the fire water tank or fire extinguisher box with the refueling port of the pressurized water pump, and issues a command to the pressurized water pump to open the refueling port via the short-range communication module; further, after receiving the command, the pressurized water pump opens the refueling port, the intelligent fire-fighting drone carrying the corresponding grade of fuel opens the solenoid valve, and fuel is added to the pressurized water pump through the fuel nozzle; when the pressurized water pump detects that the remaining fuel level is full, it commands the intelligent fire-fighting drone to close the solenoid valve and the fuel nozzle via the short-range communication module, and then the pressurized water pump closes the refueling port, and the refueling is completed; Step 3-3 includes the following steps: Sub-step 3-3-1: The intelligent fire-fighting drone measures the remaining quantity of fire extinguishing materials based on the pressure sensor of the fire water tank or fire extinguishing equipment box, and broadcasts the geographical location measured by the GPS module of the intelligent fire-fighting drone and the remaining quantity of fire extinguishing materials to the entire network. Sub-step 3-3-2: If the pressure sensor value of the fire water tank or fire extinguishing equipment box is lower than a certain threshold, it is determined that the fire extinguishing material is insufficient. The intelligent fire-fighting drone cluster will then arrange for a temporarily idle fire-fighting drone with sufficient power to carry the same model and quantity of fire water tanks or fire extinguishing equipment boxes and fire extinguishing materials to go to the location and take over the fire-fighting operation from the fire-fighting drone with insufficient fire extinguishing materials. Sub-step 3-3-3: After the fire-fighting drone with sufficient fire-fighting materials arrives and takes over the corresponding fire-fighting operation, the fire-fighting drone with insufficient fire-fighting materials will search for the optimal return route on its own and return to charge along the optimal route. In sections 3-4, the following steps are included: In sub-steps 3-4-1 and 1-1-2, the intelligent fire-fighting drones responsible for reconnaissance use artificial intelligence algorithms to automatically divide the patrol area during the fire-fighting operation, conduct a full-area patrol of the entire fire scene, and continuously take infrared photos of the fire scene using infrared cameras during the patrol. Furthermore, the artificial intelligence algorithm is used to calculate the areas with high temperatures in the infrared photos of the fire scene. If there are no areas with high temperatures, such as those greater than 100°C, it is determined that there are no residual fire points, and the intelligent fire-fighting drone cluster determines that the fire-fighting operation has ended. If there are, it returns to step 2 to continue the fire-fighting operation. Sub-step 3-4-2: If the intelligent fire-fighting drone cluster determines that the fire-fighting operation has ended, it sends a remote shutdown command to all pressurized water pumps through the short-range communication module. After receiving the remote shutdown command, the pressurized water pumps stop running. The intelligent fire-fighting drone cluster uses its grappling hooks to carry fire hoses, pressurized water pumps, and fire nozzles to drain excess water and select the optimal route to return. Step 3-4-2-1: The intelligent firefighting drone cluster receives the firefighting operation completion command confirmed by the fire command center, drains excess water, and the intelligent computer obtains its current precise coordinates. And determine the coordinates of the fire command center. The flight paths between the two are encoded into feasible solutions; Furthermore, referring to sub-steps 1-3, the intelligent fire-fighting drone swarm searches for the optimal route between the water source and the fire scene to drain excess water, and uses grappling hooks to retrieve fire hoses, booster pumps, and fire nozzles; furthermore, it uses grappling hooks to retrieve intelligent fire-fighting drones that cannot return successfully in sub-steps 3-1-3-2. Step 3-4-2-2: The intelligent computer builds a 3D map model and records the 3D coordinates of each grid cell. The system uses artificial intelligence algorithms to calculate the optimal return flight route; preferably, the shortest path algorithm is used. Furthermore, the intelligent firefighting drone swarm, referring to sub-step 3-1-3, searches for the optimal return route from the fire scene to the fire command center; Steps 3-4-2-3: The intelligent firefighting drone swarm returns along the optimal route to recharge, perform maintenance, and reload various fire extinguishing materials.

[0023] Compared with the prior art, the present invention has the following technical effects: 1) This invention provides a nearly unlimited supply of fire-fighting water. It can rapidly establish fire-fighting water supply channels between water sources and the fire scene, enabling fast firefighting operations without human intervention. The invention utilizes an intelligent fire-fighting drone swarm system to search for water sources near the fire scene (one or more), and the drone swarm replenishes fuel to booster pumps. The water supply far exceeds that of existing large fire trucks and large fire-fighting aircraft, theoretically providing a nearly unlimited supply of fire-fighting water. By using artificial intelligence algorithms, the drones can autonomously assess obstacles and plan optimal flight routes while ensuring data transmission efficiency, flight time, and energy consumption, thereby improving fire-fighting efficiency.

[0024] 2) This invention offers near-fastest fire extinguishing speed currently available in industrial applications, significantly improving firefighting efficiency and enabling immediate fire suppression. The invention utilizes an intelligent firefighting drone swarm system to search for water sources near the fire scene and then lays the optimal route between the water source and the fire scene, deploying pressurized water pumps and fire hoses without any human intervention. The intelligent firefighting drone swarm used in this invention boasts high flight speed and strong obstacle-crossing capabilities, virtually ignoring all outdoor obstacles, thus facilitating rapid fire control. When facing forests, rivers, deep pits, cliffs, and large obstacles, the speed at which this invention lays pressurized water pumps and fire hoses and provides rapid water supply far exceeds that of firefighters and fire trucks, and even surpasses that of firefighting aircraft. This invention optimizes the calculation of ignition point information. Through advanced image processing and sound analysis algorithms, it more accurately calculates ignition point information, including the location of the ignition point, the burning area, and the flame height, allowing for the development of the optimal firefighting plan on-site. This significantly improves the positioning accuracy of drones in complex environments. The intelligent firefighting drone operation, through optimized algorithms and system design, utilizes an improved K-means algorithm based on information age, effectively addressing the key issue of low firefighting efficiency in intelligent firefighting drone technology.

[0025] 3) This invention possesses strong obstacle-crossing capabilities, easily overcoming obstacles that firefighters and fire trucks cannot, and theoretically, it can overcome almost all outdoor obstacles. This invention is entirely implemented by a cluster of intelligent firefighting drones, carrying fire nozzles, pressurized water pumps, and fire hoses or extinguishing materials. Its flight altitude ensures the invention's powerful ability to overcome terrain obstacles. The system utilizes artificial intelligence algorithms to improve data processing and analysis speed based on the actual conditions of the fire scene, calculating the optimal firefighting plan in real time, including determining the number of fire water supply channels to be laid, and the optimal position, altitude, and spraying method of the drones. This will effectively solve the current problems of accurate deployment, firefighting effect, and efficiency of drones in complex fire scenarios. Based on the forest fire in Yajiang, Sichuan on March 15, 2024, the total length of the two fire water supply channels (irrigation lines) is approximately 4 kilometers, and each hose needs to be laid approximately 2 kilometers long on average. If two groups of intelligent firefighting drones of this invention are used, with each group laying one line, based on an average drone flight speed of 10 meters per second, it would take 200 seconds (approximately 3.33 minutes) to lay two fire water supply channels (irrigation lines), far faster than the 2 hours taken by firefighters at the time. If the average drone flight speed is estimated at 20 meters per second, it would take 100 seconds (approximately 1.67 minutes) to lay two fire water supply channels (irrigation lines).

[0026] 4) This invention can address challenges in special scenarios and enhance collaborative combat capabilities. The invention utilizes an intelligent firefighting drone swarm system with intelligent algorithms to achieve efficient collaborative operations between drones and ground firefighting equipment, forming an integrated firefighting system. Some firefighting drones can be quickly replaced or equipped with dry powder or liquid extinguishing agent spraying devices or other extinguishing materials as needed. This allows for precise firefighting in initial fires or high-risk areas (such as lithium battery fires, high-rise buildings, fireworks workshops) or special ignition points (such as explosive materials, chemical plants, and chemical warehouses), reducing the direct exposure risk to firefighters. This will compensate for the shortcomings of traditional firefighting equipment in high-rise building fires, and enable drones to search for water sources in forest and grassland areas where water is scarce, promoting the widespread application of intelligent firefighting drone technology. Furthermore, the application scenarios of this invention's intelligent firefighting drones are constantly expanding, not only applicable to urban fires but also playing a crucial role in complex scenarios such as forest fires, chemical fires, and high-rise building fires. Attached Figure Description

[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a system structure diagram of the present invention; Figure 2 This is a flowchart of the working method of the present invention; Figure 3 This is a schematic diagram of the operation of one fire water supply channel of the present invention. Detailed Implementation

[0028] The system structure diagram of the present invention is as follows: Figure 1 As shown, the intelligent fire-fighting drone swarm system includes an intelligent fire-fighting drone 100, other intelligent fire-fighting drones 200, a fire command center 300, a pressurized water pump 400, a fire hose 500, and a fire nozzle 600.

[0029] The intelligent fire-fighting drone 100 is an unmanned rotorcraft helicopter, comprising an intelligent computer 101, a fire water tank or fire extinguishing equipment box 102, an infrared camera 103, a microphone and speaker 104, a long-range communication module 105, a short-range communication module 106, a grappling hook 107, and a GPS module 108. This constitutes the main body of the intelligent fire-fighting drone. It can be equipped with a fire water tank or fire extinguishing equipment box 102 and an intelligent computer 101. The fire water tank or fire extinguishing equipment box 102 is used to load fire extinguishing materials for firefighting. The intelligent computer 101 runs artificial intelligence algorithms, and the long-range communication module 105 allows communication with the fire command center 300 or other intelligent fire-fighting drones 200. The short-range communication module 106 allows communication with a pressurized water pump 400. The grappling hook 107 is used to grip the pressurized water pump 400, fire hose 500, and fire nozzle 600. The intelligent firefighting drone 100 provides unmanned aerial vehicle equipment for this invention. Preferably, it is a heavy-load, long-endurance, wind-resistant drone with a wind resistance level of 8, an unloaded flight time of more than 60 minutes, a fully loaded flight time of more than 40 minutes, a payload of more than 100KG, an IP56 protection level, a cruising speed of 0-20 m / s, and supports hovering, waypoint planning, autonomous flight, automatic return, loss of control protection, and precise positioning. Further, it can accept remote control signals for remote control. Preferably, it is equipped with a GPS locator, which is a high-precision, low-power four-mode satellite positioning module, namely, Beidou + GPS + Galileo + GLONASS single-mode, dual-mode, and multi-mode operation, and can be switched between them by commands. The operating voltage is 3.0V~3.5V, supports SBAS, QZSS, and A-GNSS assisted positioning, with a capture sensitivity of -147dB, a tracking sensitivity of -163dB, a positioning accuracy of less than 3 meters, and supports power supply to an active antenna. The operating temperature is -40℃~85℃.

[0030] The intelligent computer 101, installed inside the unmanned rotorcraft helicopter 101, can control the flight speed, altitude, and direction of the intelligent fire-fighting drone, the loading and unloading of the fire water tank or fire extinguishing equipment box 102, the gripping or releasing of the grappling hook 107, and the implementation of fire-fighting operations. Furthermore, it supports an infrared camera 103, a microphone, and a speaker 104, capable of capturing infrared photos or videos of the fire scene, listening to sounds at the fire scene, and playing sounds to the fire scene. Furthermore, it can run artificial intelligence algorithms to send infrared photos, videos, or sounds from the fire scene to the fire command center 300, or receive fire-fighting operation instructions from the fire command center 300. The intelligent computer 101 provides a computing platform for this invention; preferably, it is an intelligent accelerator card equipped with an ultra-thin AE7100 chip.

[0031] The fire water tank or fire extinguishing equipment box 102 is installed below the unmanned rotor helicopter 101. It can be repeatedly filled with fire water and sprayed, or repeatedly filled with various fire extinguishing materials and released. Furthermore, the invention provides various fire extinguishing materials, and different types of fire extinguishing materials can be selected according to the fire extinguishing operation, including fire water tanks and water, fire sandboxes and sand, carbon dioxide extinguishing agents, solid dry ice, chemical flame retardant materials, window-breaking fire extinguishing grenade launchers and fire extinguishing grenades, ball throwers, barrel throwers, or other customized fire extinguishing equipment. Furthermore, according to the fire extinguishing needs, the pressurized water pump 400 can be loaded with fuel, and a solenoid valve and fuel nozzle can be installed to add fuel to the pressurized water pump 400 and extend the fire extinguishing operation time. Furthermore, the solenoid valve can be started or stopped by the intelligent computer 101 to control the flow of fuel out of the fuel nozzle or stop the flow. The fuel nozzle needs to be compatible with the fuel filling port of the pressurized water pump 400.

[0032] The infrared camera 103 is an infrared-enabled camera that provides depth images and target detection for both day and night. Preferably, it is an infrared high-definition camera. Furthermore, the camera possesses intelligent surveillance capabilities, including area intrusion detection, boundary crossing detection, area entry detection, area exit detection, object leaving detection, object taking detection, loitering detection, parking detection, person / person gathering detection, and rapid movement detection. The microphone and speaker 104 provide voice input and output functions for this invention. Preferably, the microphone is an externally polarized condenser microphone with a built-in 22mm gold-plated large diaphragm pickup head and an equivalent noise level of 15dB, ensuring high sound reproduction and detail. The remote communication module 105 provides real-time wireless broadband communication for this invention, enabling communication with the fire command center 300 or other intelligent firefighting drones 200. Preferably, the transmission distance can reach over 10km in open ground environments, 300-1000m in obstructed environments (depending on the obstruction environment), and over 30km from air to ground. Furthermore, the frequency range includes L-band, S-band, and U-band; carrier bandwidth is 5 / 10 / 20MHz, flexibly configurable; transmission system is COFDM; modulation methods include BPSK / OPSK / 16QAM / 64QAM (adaptive); transmission capability is a broadband peak rate of 120Mbps | narrowband peak rate of 600Kbps; transmit power is 2W; transmission distance is >30km (line-of-sight); supports Wi-Fi; receiver sensitivity is -103dBm@2.5MHz; video input supports IP network video input and Wi-Fi video access (HDMI / AV requires customization); voice intercom is optional with a microphone or headset; network capacity is ≥128 nodes; network hop count is >10 hops; network access time is 5 seconds after system startup; network topology supports decentralized networks, star networks, chain networks, mesh networks, etc.; encryption method is AES128 / AES256; power supply is DC. 15-36V, power consumption ≤30W, antenna interface SMA connector x2, TTL serial port J30J, power port J30J, Ethernet interface J30J, device size ≤125*81*40mm, device weight ≤400g, protection level IP65, operating temperature -30℃ to +65℃.

[0033] The short-range communication module 106 provides wireless short-range communication for this invention. Preferably, it is a multi-protocol communication module supporting Bluetooth, Zigbee, Thread, Proprietary, and Wi-Fi, with a maximum flash memory of 3200 KB / RAM of 512 KB and an output power range of -20 to 19.5 dBm. Furthermore, it can communicate with the pressurized water pump 400 to obtain the remaining fuel level, send a refueling command, and open the refueling port of the pressurized water pump 400 to add fuel. The hook 107 is preferably an electronically controlled steel robotic gripper with a load capacity of not less than 300KG. It can open or close according to the instructions of the intelligent computer 101, thereby fixing or releasing the pressurized water pump 400, fire hose 500, and fire nozzle 600, and even grabbing intelligent fire-fighting drones that cannot return. Preferably, there is one or two hooks. The minimum operating temperature is -30℃, the maximum operating temperature is 90℃, the minimum power supply voltage is 1V, the maximum power supply voltage is 9.5V, the width is 3.6mm, the height is 1.3mm, and the length is greater than the diameter of the fire hose 500 and the fire nozzle 600 or greater than 100mm, that is, greater than the diameter of the largest model fire hose 500 by 80mm.

[0034] The GPS module (108) provides precise positioning for this invention. Preferably, it is a high-precision, low-power quad-mode satellite positioning module, i.e., single-mode, dual-mode, and multi-mode operation of BeiDou + GPS + Galileo + GLONASS, which can be switched between via commands. It operates at 3.0V~3.5V, supports SBAS, QZSS, and A-GNSS assisted positioning, has a capture sensitivity of -147dB, a tracking sensitivity of -163dB, a positioning accuracy of less than 3 meters, supports power supply to active antennas, and operates at a temperature of -40℃~85℃. The fire command center 300 is a computer controlled by fire command personnel. It can communicate with the intelligent computer 101 on the intelligent fire-fighting drone, send fire-fighting operation commands to the intelligent fire-fighting drone, modify or adjust the fire-fighting operation plan of the intelligent fire-fighting drone cluster, and receive fire scene information observed by the intelligent fire-fighting drone, including infrared photos, videos, or audio of the fire scene. Furthermore, it provides a computing and storage center for this invention, preferably multiple rack-mounted servers, 2U, with Hygon 7380 CPUs.

[0035] The pressurized water pump 400 provides a fire water supply channel (irrigation line) or boosts the pressure of fire hoses to ensure long-distance water transmission. Preferably, it is a diesel-fueled fire pump set, requiring no external power supply, with a small size to reduce the floor space required for outdoor use. It does not require disassembly of inlet and outlet pipes and can be connected to the fire hose 500 for conveying clean water without solid particles and liquids with physical and chemical properties similar to water. It is mainly used for fire hydrant spraying and pressure stabilization. Furthermore, the flow rate of the conveyed liquid ranges from 0.42 to 333 L / s, the pressure ranges from 0.8 to 2.25 MPa, the power ranges from 0.18 to 450 KW, the diameter ranges from 15 to 400 mm, and the weight does not exceed 100 KG. Furthermore, the system is equipped with a remaining fuel level detection sensor and a short-range communication module 106, which can report the remaining fuel level to the nearest intelligent fire-fighting drone. When the remaining fuel level is below a threshold (e.g., 10%), an alarm is triggered, and the refueling port is opened or closed according to instructions to allow the intelligent fire-fighting drone to refuel. Furthermore, the short-range communication module 106 can receive remote start / stop commands from the intelligent fire-fighting drone, thereby executing the corresponding start or stop commands. Preferably, one booster pump 400 is installed for every 100 meters (500mm) of fire hose.

[0036] The fire hose 500 provides a fire water supply channel (water delivery line) for this invention, and is preferably made of polymer materials such as polyurethane. Furthermore, both ends of the fire hose have metal connectors, which can be connected to another fire hose to extend the distance, or to a fire nozzle 600 to increase the liquid spray pressure, or to a booster pump 400 to pressurize and deliver water. The fire nozzle 600 provides a water gun nozzle and water source spray, increasing the fire water spray pressure. Preferably, it is made of copper alloy or stainless steel, with an inner wall roughness Ra≤12.5μm, and can be connected to the fire hose 500. Furthermore, the conventional fire nozzle 600 diameter specifications are 13mm, 16mm, and 19mm, corresponding to different flow rate requirements.

[0037] like Figure 2 The diagram shown is a flowchart of the working method of the present invention. An intelligent firefighting drone swarm system is characterized by comprising the following steps: Step 1: The intelligent firefighting drone swarm reconnoiters the fire scene and searches for water sources near the fire scene, then searches for the optimal route between the water source and the fire scene to lay booster pumps and fire hoses; For example... Figure 3 The diagram shown is a schematic diagram of the operation of one fire water supply channel of the present invention.

[0038] Sub-step 1-1: The intelligent fire-fighting drone cluster receives a fire alarm and goes to the fire scene to reconnoiter and determine the fire scene information and surrounding water source information; the fire scene information includes the geographical location, area, height, distribution of ignition points, resident situation, distribution of combustibles, wind direction, wind force, and estimates the fire intensity and development trend; the surrounding water source information includes the geographical location of the water source, the route from the fire scene, area, total water volume, height, and water source environmental conditions; Sub-step 1-1-1: The intelligent fire-fighting drone cluster receives a fire alarm, obtains the approximate geographical location of the fire alarm, dispatches several intelligent fire-fighting drones, activates the GPS module 108, and goes to the geographical location to reconnoiter and determine the fire scene information and surrounding water source information. Sub-step 1-1-2: After arriving at the geographical location, the dispatched intelligent fire-fighting drones use artificial intelligence algorithms to automatically divide the patrol area and turn on the infrared camera 103 to conduct a full-area patrol of the entire fire scene. Sub-step 1-1-2-1: The intelligent fire-fighting drone cluster divides the entire fire scene into patrol areas of equal number and similar size based on the number of dispatched drones, and uses a random algorithm to assign several intelligent fire-fighting drones to patrol different patrol areas. Sub-step 1-1-2-2: During the inspection, the intelligent computer 101 activates the infrared camera 103 to continuously capture infrared photos of the fire scene, and activates the microphone and speaker 104 to collect sounds from the fire scene. Furthermore, an adaptive mean filtering algorithm is used to denoise the infrared photos or videos of the fire scene acquired by the infrared camera 103. The formula is: ; in It is a pixel The domain pixels, The number of pixels in the window's neighborhood. For window neighborhood, It is the grayscale value of the pixels in the infrared photos or videos of the fire scene; Sub-steps 1-1-2-3: The intelligent computer 101 segments the infrared image or video frame of the fire scene, performing threshold segmentation of the color space based on pixel hue H, saturation S, and brightness value V, using the following formula: ; Sub-steps 1-1-2-4: The intelligent computer 101 filters the sound signals collected by the microphone and speaker 104, performs feature extraction, and extracts the combustion sound features and distress call features related to the ignition point. The formula is: ; in, This represents the energy of the m-th filter bank. Indicates the number of filters. These are the MFCC coefficient numbers of the sound signals collected by the microphone and speaker 104; Sub-step 1-1-2-5: Intelligent computer 101 uses a threshold detection algorithm to calculate areas with high temperatures in the infrared photos of the fire scene, such as those greater than 100℃, and marks them as ignition points. Furthermore, threshold edge detection is performed on the denoised image to locate the ignition point. The formula for calculating the center of the ignition point is as follows: ; in, It is a comprehensive window coefficient used to calculate the gray value of the center pixel in a weighted manner, for example, to determine whether it is greater than 100°C; Sub-steps 1-1-2-6: The intelligent computer 101, based on the GPS module 108 positioning data at the corresponding time, and combined with the characteristics, propagation direction, and intensity of the sound signal, assists in determining the location of key points within the fire area. Furthermore, calculate the geographical location, area, height, and distribution of fire points at the fire scene, including all ignition points. Sub-step 1-1-2-7: If the intelligent computer 101 detects an area with a high temperature at the edge of the infrared photo or video of the fire scene, i.e. the ignition point, it expands the patrol range in that direction until no area with a high temperature is detected at the edge of the infrared photo of the fire scene taken during the patrol. Sub-step 1-1-2-8: Intelligent computer 101 uses a target recognition algorithm to identify targets in infrared photos of the fire scene, calculates the living conditions of people, the distribution of combustibles, wind direction, and wind force; further, based on the target identification and living conditions of people at the fire scene, it calculates whether it is a hazardous chemical warehouse or production area, or whether it is an explosive material warehouse or production area. Sub-steps 1-1-2-9: Intelligent computer 101 estimates the fire spread speed and development trend based on the geographical location, area, height, distribution of ignition points, distribution of combustibles, wind direction, and wind force of the fire scene. Furthermore, the rate of fire spread is estimated based on the properties of the combustible material, the building structure, and the ventilation conditions at the fire scene, using the following formula: ; in, To the speed of the fire's spread, The constant is related to the properties of the combustible material. The wind speed is [value], and the area of ​​the fire scene is [area]. The height is ; Sub-step 1-1-2-10: Based on the estimated fire spread rate, the intelligent computer 101 estimates the fire development trend and calculates the time required for the fire to develop to the critical location. The key locations mentioned refer to densely populated residential areas, areas with high population density, areas containing flammable, explosive, or hazardous chemicals, or other important areas requiring attention, such as bridges, tunnels, schools, and military zones. The formula for predicting the fire's development trend is: ; in, It is the distance from the ignition point to the critical location. The speed at which the fire spreads; In sub-step 1-1-3, several intelligent fire-fighting drones are dispatched. After arriving at the geographical location, they use artificial intelligence algorithms to automatically divide the patrol area and conduct a full-area patrol of the surrounding water sources. During the patrol, the infrared cameras 103 are turned on to continuously take infrared photos. Further, the target recognition algorithm is used to identify targets in the infrared photos. Based on the GPS module 108 positioning data at the corresponding time, the geographical location of the water source, the route from the fire scene, the area, the total water volume, the height, and the environmental conditions of the water source are calculated. Furthermore, the intelligent fire-fighting drone swarm is divided into patrol areas according to sub-step 1-1-2-1, and a random algorithm is used to assign several intelligent fire-fighting drones to different patrol areas for patrol. Sub-steps 1-2: Based on the reconnaissance of the fire scene and surrounding water source information, the intelligent fire-fighting drone cluster determines how many fire water supply channels need to be laid, calculates the type and quantity of pressurized water pumps, fire hoses or fire extinguishing materials to be carried, and calculates the number of intelligent fire-fighting drones to be dispatched. Sub-step 1-2-1: The intelligent fire-fighting drone swarm estimates the time required to lay fire-fighting water supply channels between the fire site and the water source based on the reconnaissance information of the fire scene and surrounding water sources. It also estimates the fire intensity and development trend after the estimated time for laying the fire-fighting water supply channels, based on sub-step 1-1-2, and calculates the water consumption and the number of fire-fighting water supply channels required. Further, the fire-fighting water supply channel refers to all fire hoses, booster pumps, and fire nozzles continuously laid from the water source to the fire scene; one fire-fighting water supply channel can independently complete one fire-fighting water supply operation. Furthermore, smaller fires require only a smaller water consumption and one fire-fighting water supply channel, while larger fires require a larger water consumption and multiple fire-fighting water supply channels. Furthermore, the effective coverage area of ​​one fire water supply channel or fire extinguishing material is... The calculation formula is: ; in, It is the altitude at which intelligent firefighting drones perform their missions. It refers to the spray angle of the fire hose or extinguishing material. Sub-step 1-2-2: Based on the reconnaissance information of the fire scene and surrounding water sources, the intelligent fire-fighting drone swarm estimates the total length required to lay a fire water supply channel between the fire scene and the water source. Based on this total length, it calculates the actual required length of fire hose (500mm), reserving a 20%-30% margin. Preferably, before a fire occurs, the fire hose (500mm) is assembled in 100-meter sections to reduce assembly time in case of a fire. Further, based on the actual required length of the fire hose (500mm), it calculates the type and quantity of the required booster pumps (400mm). Preferably, each 100-meter section... A fire hose of 500 is equipped with a booster pump of 400. Further, based on the target identification in sub-step 1-1-2, including the distribution of ignition points, the distribution of combustibles, hazardous chemical warehouses or production areas, and explosive material warehouses or production areas, the type and quantity of fire extinguishing materials to be carried are calculated. Firefighters work together to assemble the fire hose of 500 and the booster pump of 400 on the ground according to the required quantity and length. The type and quantity of fire extinguishing materials to be carried are loaded onto the corresponding intelligent fire-fighting drone. Multiple fire water supply channels are appropriately separated, and one fire nozzle of 600 is equipped at the end of each fire water supply channel. Furthermore, based on the properties of the obtained combustible material, the type of extinguishing material to be carried is determined, and the total demand for extinguishing material is estimated, using the following formula: ; in, The standard for extinguishing material dosage per unit area is given, and the fire scene area is... , It is the density of the combustible material. It is the extinguishing agent efficiency coefficient; Sub-steps 1-2-3: The intelligent fire-fighting drone cluster calculates the number of intelligent fire-fighting drones to be dispatched. Preferably, 1-2 intelligent fire-fighting drones are equipped for every 100 meters of fire hose (500), 1-2 intelligent fire-fighting drones are equipped for every 400 booster pump, 1 intelligent fire-fighting drone is equipped for every set of fire extinguishing materials, and 1-2 intelligent fire-fighting drones are equipped for every 600 fire nozzles. Furthermore, the formula for estimating the number of intelligent firefighting drones that need to be dispatched is: ; in, This is the safety redundancy factor; the fire scene area is... One fire water supply channel or fire extinguishing material effectively covers an area of , Indicates rounding up; Sub-steps 1-3 involve the intelligent firefighting drone swarm using artificial intelligence algorithms to search for the optimal route and firefighting plan between the water source and the fire scene; Sub-step 1-3-1: The intelligent firefighting drone cluster, through its respective intelligent computer 101 and remote communication module 105, forms a distributed edge computing system. It shares the geographical locations of water sources and the fire scene, as well as the current geographical locations of each intelligent firefighting drone. It then models the firefighting operation task, encodes the flight path into a feasible solution, and divides the locations from the fire command center 300, firefighting materials, water sources to the fire scene into grids and fixed coordinates. The grid set of the firefighting operation area is then used... To indicate, the first Each grid is used To indicate, This represents the maximum number of grid cells in the firefighting operation area. (Intelligent firefighting drones) Firefighting operation grid set accessed via flight path To express.

[0039] Sub-steps 1-3-2 are for coordinating the dispatch of... Task allocation among intelligent firefighting drones: The intelligent firefighting drone swarm employs an improved K-means algorithm, which divides the set of flight paths for the firefighting operation area into... Each cluster serves as a set of firefighting operations, i.e. This includes transporting fire extinguishing materials, transporting 500 fire hoses and 400 booster pumps, operating 600 fire nozzles, and connecting the water source and the fire scene using the 500 fire hoses and 400 booster pumps. The fire extinguishing operations of each cluster are independent of each other, without overlap or intersection, and must adhere to specific constraints. ; This refers to two different intelligent firefighting drones.

[0040] Sub-step 1-3-3, in the... A collection of firefighting operations by intelligent firefighting drones In this context, the intelligent firefighting drone swarm defines an extended set. This set is composed of elements indexed by... Indicates the first Firefighting operation clusters in grids Collection of fire scene locations For each intelligent firefighting drone's firefighting operation cluster Construct a firefighting operation plan It consists of the locations of the fire scene. and edge set Composition. Edge set It includes information from the fire command center (300), fire extinguishing materials, water sources, to the location of the fire scene. All possible flight routes , , and This indicates the grid locations along the flight path, including the fire command center 300, fire extinguishing materials, water sources, and other fire scene locations. yes China is different The location of the fire scene.

[0041] Sub-steps 1-3-4, firefighting operation diagrams for each intelligent firefighting drone. In the middle, each edge , Each corresponds to a non-negative weight. This weight indicates the location of the intelligent firefighting drone within the two grids at the fire scene. and The distance between the flight paths is calculated using the following formula: - ; Sub-steps 1-3-5: Setting up the intelligent firefighting drone cluster, ensuring that the intelligent firefighting drone n maintains a fixed altitude during flight. And a constant flight speed. Furthermore, in intelligent firefighting drone firefighting operations... In the context of the communication link between the intelligent firefighting drone n and the fire command center 300, the channel power gain can be expressed as: , The power gain that the channel can provide at a distance of 1 meter from the reference point; when the fire command center is operating at constant power For data transmission, its data transmission rate on the link can be expressed as: ; Here, B represents the bandwidth that can be used during communication; This refers to the power consumed by the intelligent firefighting drone when transmitting data; This is the power gain that the channel can provide at a distance of 1 meter from the reference point; This represents the power of the noise in the channel.

[0042] Sub-steps 1-3-6: Setting up the intelligent firefighting drone It will travel at a fixed speed during flight. The formula for the propulsion power generated during flight is: ; in, The power generated by the outer contour of the blade, and This represents the power generated by the induction effect. This refers to the speed at the tip of the rotor blades. It is the average speed sensed by the rotor while it is hovering. It is the ratio between fuselage drag and overall drag. Represents the density of air. This reflects the robustness of the rotor, while Y represents the area occupied by the rotor disk; Sub-steps 1-3-7, when the intelligent firefighting drone In the selected route grid During flight, the intelligent computer 101 calculates the flight energy consumption it generates.

[0043] ; Among them, intelligent firefighting drones Flight time ; For intelligent firefighting drones Fixed flight speed.

[0044] Sub-step 1-3-8: To improve firefighting efficiency, the "firefighting information age" indicator is introduced to measure the firefighting timeliness of the intelligent firefighting drone swarm. This measures the total time elapsed from when the intelligent firefighting drone swarm begins capturing infrared photos or videos of the fire scene to when it arrives at the scene and begins firefighting operations. This includes the time spent capturing infrared photos or videos, analyzing those photos or videos, route planning, and flight time. The firefighting information age accurately represents the total time from the discovery of the fire scene to the current firefighting operation, thus providing a more precise description of firefighting efficiency. Furthermore, the average information age of all locations within the intelligent firefighting drone swarm is taken to further evaluate the overall firefighting efficiency of the system. The formula for calculating the average information age of these locations is: ,

[0045] in, This refers to a cluster of intelligent firefighting drones for firefighting operations. The total number of fire scene locations included in the data. It is used to represent intelligent firefighting drones. From the firefighting operation location Fly to On the way, from the moment we left the fire command center 300, we arrived A parameter representing the number of all locations visited before the current location. (In sub-routes) For example, we can get ; Sub-steps 1-3-9: The intelligent firefighting drone swarm needs to achieve a balance between the average age of firefighting information and the energy consumption of the intelligent firefighting drones. The optimal route selection formula for the rational planning of the intelligent firefighting drones' flight paths is as follows: ; in, It is a binary variable whose value can only be 0 or 1, representing the intelligent firefighting drone. Have you selected a route? , Specifically, when This means that intelligent firefighting drones We did indeed travel along this route. At the same time, Representative of intelligent firefighting drones Access the total number of fire scene locations, and group all such binary variables together and label them as follows. .

[0046] Sub-steps 1-3-10 are to ensure the fire extinguishing operation diagram. Each fire scene location in the system has the same in-degree, and this value is set to 1. The intelligent fire-fighting drone swarm applies an in-degree constraint to the route selection formula: ; Sub-step 1-3-11, to ensure the diagram Each fire scene location in the system has the same out-degree, and this value is set to 1. The intelligent fire-fighting drone swarm applies out-degree constraints to the route selection formula: ; Sub-steps 1-3-12 are to ensure that the intelligent firefighting drone... Do not select route , hour hour, When the intelligent firefighting drone chooses this route... hour Sub-journey constraints are applied to the route selection formula: , ,

[0047] In sub-step 1-3-13, the intelligent firefighting drone swarm checks whether the solution results meet the predetermined error threshold or solution time limit. If not, it returns to sub-step 1-3-1 to continue calculation until the optimal feasible solution, i.e., an optimal flight route, is obtained. The optimal flight route This refers to the intelligent firefighting drone swarm departing from its current location and arriving at its flight path. Upon reaching the destination, the team laid one or more fire water supply channels between the water source and the fire scene, using the 500 fire hoses, 400 booster pumps, 600 fire nozzles, and other necessary fire extinguishing materials they carried.

[0048] Sub-steps 1-3-14: The intelligent firefighting drone swarm each follows its optimal flight path. Fly to the fire scene; In sub-steps 1-4, the intelligent fire-fighting drone swarm uses grappling hooks 107 to load one or more fire water supply channels that have been assembled in sub-steps 1-2-2, including fire hoses 500, booster pumps 400, fire nozzles 600, and the necessary fire extinguishing materials, according to their assigned tasks. After all loading is completed and confirmed to be correct, they take off simultaneously and follow the optimal route to lay booster pumps and fire hoses between the water source and the fire scene.

[0049] Step 2: The intelligent fire-fighting drone cluster controls the fire nozzles to carry out fire extinguishing operations, and adjusts the position of the booster pump and fire hose according to the changes in the fire situation at the fire scene; Sub-step 2-1: The intelligent fire-fighting drone cluster calculates the ignition point information and the optimal fire extinguishing plan based on infrared photos, videos or sounds of the fire scene, and calculates the optimal spraying method of the fire nozzles or the release method of the fire extinguishing materials. Sub-step 2-1-1: The intelligent fire-fighting drone cluster activates infrared cameras 103 to continuously capture infrared photos, videos, or sounds of the fire scene, calculates the ignition point information and distribution, and uses artificial intelligence algorithms to rate the hazard level of each ignition point; further, it predicts the development trend of the fire at each ignition point. If it is likely to develop into a larger-scale fire, cause greater danger, or be more difficult to control in the short term, it is considered high-risk; otherwise, it is considered low-risk.

[0050] Sub-step 2-1-2: Based on the distribution of fire points and the hazard rating of each fire point, the intelligent fire-fighting drone cluster calculates the optimal fire extinguishing plan according to the principle of extinguishing all fire points in the shortest time. Based on the location of the fire point and the fire scene environment, the intelligent computer 101 calculates the optimal hovering position for controlling fire nozzles or intelligent fire-fighting drones carrying fire extinguishing materials, i.e., the optimal fire extinguishing plan, so that it can cover the fire point and effectively extinguish the fire. The formula for the optimal hovering position of the drone is: ; in, q represents the location of the fire points within the fire scene, and q represents the number of fire points. In sub-steps 2-1-3, the intelligent fire-fighting drone cluster calculates the optimal spraying method of the fire nozzles or the release method of the fire extinguishing material according to the location of one or more fire water supply channels and the optimal fire extinguishing plan. Furthermore, high-risk fire points should be extinguished first, low-risk fire points can be extinguished second best, and fire points where water spraying is ineffective should be considered for releasing other fire extinguishing materials.

[0051] Furthermore, the intelligent computer 101 calculates the optimal hovering height of the intelligent fire-fighting drone based on the effective range of the fire nozzle or extinguishing material, using the following formula: ; in, The height of the ignition point, The minimum safe flight altitude for intelligent firefighting drones, The effective range height of the fire extinguishing equipment; Furthermore, the intelligent computer 101 determines the relative position and altitude of the intelligent firefighting drone to the fire point. The intelligent computer 101 calculates the optimal spray angle for fire-fighting water or extinguishing materials using the following formula: ; Sub-step 2-2: The intelligent fire-fighting drone cluster activates all pressurized water pumps and controls fire nozzles to spray water for fire extinguishing operations, or releases fire extinguishing materials for fire extinguishing operations, according to the optimal fire extinguishing plan. Sub-step 2-2-1: The intelligent fire-fighting drone cluster sends a remote start command to all pressurized water pumps 400 through the short-range communication module 106. After receiving the remote start command, the pressurized water pumps 400 start running. Sub-step 2-2-2: The intelligent fire-fighting drone cluster controls the fire nozzles 600 to spray water for fire-fighting operations according to the optimal fire-fighting plan; furthermore, since the fire nozzles 600 have a large spray force, 2-3 fire-fighting drones can be dispatched to control the fire nozzles 600 and the fire hoses 500 connected to their rear ends, so as to better control the effective range height and the optimal spray angle. Furthermore, based on the relative position and altitude of the intelligent firefighting drone to the fire point, the intelligent computer 101 calculates the flow rate of fire-fighting water or fire extinguishing materials using the following formula: ; in This refers to the amount of fire water or fire extinguishing materials required per unit area. Area of ​​the ignition point; Sub-step 2-2-3: The intelligent fire-fighting drone cluster checks for any fire points where water spraying is ineffective. If any fire is found, other fire-fighting materials should be considered for fire-fighting operations, including fire sand, carbon dioxide extinguishing agent, solid dry ice, chemical flame-retardant materials, window-breaking fire extinguishing grenade launchers and fire extinguishing grenades, ball throwers, barrel throwers, or other customized fire-fighting equipment. Sub-steps 2-3 involve the intelligent firefighting drone cluster adjusting the positions of the booster pumps and fire hoses based on changes in the fire situation at the scene.

[0052] Sub-step 2-3-1: The intelligent fire-fighting drone cluster activates infrared camera 103 to continuously capture infrared photos of the fire scene, uses artificial intelligence algorithms to calculate the changes in fire intensity and the distribution of ignition points, and assigns a hazard rating to each ignition point. On the one hand, the fire in the vicinity may be getting smaller and smaller, and water spraying may not be able to reach the ignition points further away. On the other hand, the fire in the vicinity may be getting bigger and bigger, endangering the pressurized water pump 400, fire hose 500, and intelligent fire-fighting drones. Sub-step 2-3-2: The intelligent fire-fighting drone cluster recalculates the optimal fire extinguishing plan based on the distribution of fire points; Furthermore, referring to sub-steps 1-3, the intelligent fire-fighting drone swarm updates the distribution of fire points and re-searches for the optimal route and optimal fire-fighting plan between the water source and the fire scene; In sub-step 2-3-3, the intelligent fire-fighting drone cluster adjusts the positions of the booster pump and fire hose according to the recalculated optimal fire-fighting plan.

[0053] Step 3: The intelligent fire-fighting drone swarm takes turns controlling the fire nozzles, booster pumps, and fire hoses, and replenishes fuel to the booster pumps until the fire-fighting operation is completed.

[0054] Sub-step 3-1: The intelligent fire-fighting drone cluster checks the remaining power of each drone. The drone with sufficient power takes over the fire-fighting drone with insufficient power to continue the fire-fighting operation, and controls the fire nozzles, pressurized water pumps and fire hoses. The drone with insufficient power searches for the optimal route to return to the charging station. Sub-step 3-1-1: The intelligent fire-fighting drones check their remaining battery power and share the remaining battery power information with the intelligent fire-fighting drone cluster through the remote communication module 105; In sub-step 3-1-2, the intelligent fire-fighting drone cluster checks for intelligent fire-fighting drones with insufficient remaining power, for example, less than 20% remaining power. It measures the location of the drone using its GPS module 108 and broadcasts its remaining power and location to the entire network. Further, temporarily idle fire-fighting drones with sufficient power go to the location and take over the fire-fighting operation from the drone with insufficient power, controlling the fire nozzles, pressurized water pumps, and fire hoses. If the fire-fighting drone with insufficient power is carrying a fire water tank or fire extinguishing equipment box 102 and fire extinguishing materials, then the temporarily idle fire-fighting drone with sufficient power carries the same type and quantity of fire water tanks or fire extinguishing equipment boxes 102 and fire extinguishing materials to the location and take over the fire-fighting operation from the drone with insufficient power. Sub-step 3-1-3: After the fire-fighting drone with sufficient power arrives and takes over the corresponding fire-fighting operation, the fire-fighting drone with insufficient power will search for the optimal return route and return to charge along the optimal route. Sub-step 3-1-3-1: The intelligent computer 101 of the fire-fighting drone, which is running low on power, calculates the Euclidean distance of the intelligent fire-fighting drone from the fire scene to the fire command center (300 meters away). The formula is: ; in, and These are the coordinates of the intelligent firefighting drone at two locations: the fire scene and the fire command center (300 units away). Sub-step 3-1-3-2: The intelligent computer 101 of the fire-fighting drone with insufficient power obtains the remaining power of the intelligent fire-fighting drone. The system also performs dynamic return-to-home battery threshold calculations on the intelligent firefighting drone, determining whether the remaining battery power is sufficient for a successful return. The formula is as follows: ; in, This indicates the safety margin threshold, which defaults to 15%. Indicates the power consumption coefficient per unit distance; If the remaining battery power of the intelligent firefighting drone is sufficient to return to base, proceed to sub-step 3-1-3-3; otherwise, if it cannot return to base, it will remain in place and maintain communication, but will no longer undertake any firefighting operations. Sub-step 3-1-3-3: The intelligent computer 101 of the fire-fighting drone, with insufficient power, plans the return route of the intelligent fire-fighting drone based on the remaining power using a route planning algorithm. The route cost function is: ; in, Indicates cumulative energy consumption. Represents the cost of heuristic distance. , These are weighting coefficients, and they have... + ; Sub-step 3-1-3-4: The intelligent computer 101 of the firefighting drone, which has insufficient power, simultaneously applies power consumption constraints to the above functions, as shown in the formula: ; in Indicates the first The power function of the route segment, Indicates the first Flight speed of the segment route, Indicates the first The length of the route segment Indicates the safety margin threshold; In sub-step 3-1-3-5, the intelligent computer 101 of the fire-fighting drone, which has insufficient power, checks whether the solution result meets the predetermined error threshold or solution time limit. If not, it returns to sub-step 3-1-3-1 to continue calculation until the optimal feasible solution, i.e., an optimal flight route, is obtained. ; Sub-step 3-1-3-6: The intelligent firefighting drone with insufficient power follows the optimal flight path solved by the intelligent computer 101. return; Sub-step 3-2: The intelligent fire-fighting drone cluster checks the remaining fuel of the booster pump. If the fuel is insufficient, it carries the corresponding grade of fuel to replenish the booster pump. Sub-step 3-2-1: During fire-fighting operations, the booster pump 400 continuously monitors the remaining fuel level and reports it to the nearest intelligent fire-fighting drone via the short-range communication module 106. The intelligent fire-fighting drone measures its current geographical location via the GPS module 108. Sub-step 3-2-2: When the booster pump 400 detects that the remaining fuel level is below a threshold (e.g., 10%), it alarms. The nearest intelligent fire-fighting drone reports its geographical location measured by the GPS module 108, and the nearest adjacent intelligent fire-fighting drone carrying the corresponding grade of fuel travels to the location of the booster pump 400 to replenish it. Sub-step 3-2-3: The intelligent fire-fighting drone carrying the corresponding grade of fuel aligns the fuel nozzle of the fire water tank or fire extinguisher box 102 with the fuel filler port of the booster pump 400 and communicates with the nearest intelligent fire-fighting drone via the short-range communication module 106. The short-range communication module 106 issues a command to the pressurized water pump 400 to open the refueling port; further, after receiving the command, the pressurized water pump 400 opens the refueling port, and the intelligent fire-fighting drone carrying the corresponding grade of fuel opens the solenoid valve, and the fuel is added to the pressurized water pump 400 through the fuel nozzle; when the pressurized water pump 400 detects that the remaining fuel level is full, it commands the intelligent fire-fighting drone to close the solenoid valve and fuel nozzle through the short-range communication module 106, and then the pressurized water pump 400 closes the refueling port, and the refueling ends; sub-step 3-3, the intelligent fire-fighting drone cluster checks the remaining quantity of fire extinguishing materials. If the fire extinguishing materials are insufficient, it carries the corresponding quantity of fire extinguishing materials to replenish the fire scene; sub-step 3-3-1, the intelligent fire-fighting drone measures the remaining quantity of fire extinguishing materials according to the pressure sensor of the fire water tank or fire extinguishing equipment box 102, and broadcasts the geographical location measured by the intelligent fire-fighting drone's GPS module 108 and the remaining quantity of fire extinguishing materials to the entire network; Sub-step 3-3-2: If the pressure sensor value of the fire water tank or fire extinguishing equipment box 102 is lower than a certain threshold, it is determined that the fire extinguishing material is insufficient. The intelligent fire-fighting drone cluster then arranges a temporarily idle fire-fighting drone with sufficient power to carry the same model and quantity of fire water tanks or fire extinguishing equipment boxes 102 and fire extinguishing materials to the location and take over the fire-fighting operation from the drone with insufficient fire extinguishing materials. Sub-step 3-3-3: After the fire-fighting drone with sufficient fire extinguishing materials arrives and takes over the corresponding fire-fighting operation, the fire-fighting drone with insufficient fire extinguishing materials searches for the optimal return route and returns to charge along the optimal route. In sub-steps 3-4, the intelligent fire-fighting drone cluster uses infrared cameras to check if there are any remaining ignition points at the fire scene. If not, the fire-fighting operation is considered complete; otherwise, return to step 2 to continue the fire-fighting operation.

[0055] In sub-steps 3-4-1 and 1-1-2, the intelligent fire-fighting drones responsible for reconnaissance use artificial intelligence algorithms to automatically divide patrol areas during firefighting operations, conducting full-area coverage patrols of the entire fire scene. During the patrol, the infrared camera 103 continuously captures infrared photos of the fire scene. Furthermore, the artificial intelligence algorithm calculates areas with high temperatures in the infrared photos of the fire scene. If there are no areas with high temperatures, such as above 100°C, it is determined that there are no residual fire points, and the intelligent fire-fighting drone cluster determines that the firefighting operation is over. If there are, it returns to step 2 to continue the firefighting operation. In sub-step 3-4-2, if the intelligent fire-fighting drone cluster determines that the firefighting operation is over, it sends a remote shutdown command to all pressurized water pumps 400 through the short-range communication module 106. After receiving the remote shutdown command, the pressurized water pumps 400 stop operating. The intelligent fire-fighting drone cluster uses the grappling hook 107 to carry the fire hose 500, pressurized water pumps 400, and fire nozzles 600 to drain excess water and selects the optimal route to return. Step 3-4-2-1: The intelligent firefighting drone cluster receives the firefighting operation completion command confirmed by the fire command center 300, drains excess water, and the intelligent computer 101 obtains its current precise coordinates. And determine the coordinates of the fire command center. The flight paths between the two are encoded into feasible solutions. Further, referring to sub-steps 1-3, the intelligent fire-fighting drone swarm re-searches for the optimal route between the water source and the fire scene to drain excess water, and uses grappling hook 107 to retrieve fire hose 500, booster pump 400, and fire nozzle 600. Further, grappling hook 107 is used to retrieve any intelligent fire-fighting drones that cannot return successfully in sub-step 3-1-3-2. In step 3-4-2-2, the intelligent computer 101 establishes a three-dimensional map model and records the three-dimensional coordinates of each grid cell. The system uses an artificial intelligence algorithm to calculate the optimal return flight route; preferably, the shortest path algorithm is used; further, the intelligent fire-fighting drone cluster refers to sub-step 3-1-3 to search for the optimal return route from the fire scene to the fire command center 300; in step 3-4-2-3, the intelligent fire-fighting drone cluster returns along the optimal route, recharges, performs maintenance, and reloads various fire-fighting materials.

Claims

1. An intelligent firefighting drone swarm system, characterized in that, The system includes an intelligent fire-fighting drone and a fire command center (300); the intelligent fire-fighting drone is equipped with a grappling hook (107) for gripping a pressurized water pump (400), a fire hose (500), and a fire nozzle (600); when the system is in operation, it includes the following steps: Step 1: The intelligent fire-fighting drone swarm reconnoiters the fire scene and searches for water sources near the fire scene, and searches for the optimal route between the water source and the fire scene to lay pressurized water pumps and fire hoses; Step 2: The intelligent fire-fighting drone swarm controls the fire nozzles to carry out fire extinguishing operations, and adjusts the position of the booster pump and fire hose according to the changes in the fire situation at the fire scene; Step 3: The intelligent fire-fighting drone swarm takes turns controlling the fire nozzles, booster pumps, and fire hoses, and replenishes fuel to the booster pumps until the fire-fighting operation is completed.

2. The method according to claim 1, characterized in that, Step 1 includes the following sub-steps: Sub-step 1-1: The intelligent fire-fighting drone cluster receives a fire alarm and goes to the fire scene to investigate and determine the fire scene information and surrounding water source information; Fire scene information includes the geographical location, area, height, distribution of ignition points, occupancy status, distribution of combustibles, wind direction, wind force, and estimates of fire intensity and development trend. Information on surrounding water sources includes the geographical location of the water source, the route from the fire site, its area, total water volume, height, and environmental conditions. Sub-steps 1-2: Based on the reconnaissance of the fire scene and surrounding water source information, the intelligent fire-fighting drone cluster determines how many fire water supply channels need to be laid, calculates the type and quantity of pressurized water pumps, fire hoses or fire extinguishing materials to be carried, and calculates the number of intelligent fire-fighting drones to be dispatched. Sub-steps 1-3 involve the intelligent firefighting drone swarm using artificial intelligence algorithms to search for the optimal route and firefighting plan between the water source and the fire scene; Sub-steps 1-4: The intelligent fire-fighting drone swarm lays pressurized water pumps and fire hoses along the optimal route; Step 1-1 includes the following steps: Sub-step 1-1-1: The intelligent fire-fighting drone cluster receives a fire alarm, obtains the approximate geographical location of the fire alarm, dispatches several intelligent fire-fighting drones, activates the GPS module (108), and goes to the geographical location to reconnoiter and determine the fire scene information and surrounding water source information. Sub-step 1-1-2: Several intelligent fire-fighting drones are dispatched. After arriving at the geographical location, they use artificial intelligence algorithms to automatically divide the patrol area and turn on the infrared camera (103) to conduct a full-area patrol of the entire fire scene. Sub-step 1-1-3: Several intelligent fire-fighting drones are dispatched. After arriving at the geographical location, they use artificial intelligence algorithms to automatically divide the patrol area, conduct full-area coverage patrol of the surrounding water sources, and continuously take infrared photos by turning on the infrared camera (103) during the patrol. Furthermore, a target recognition algorithm is used to identify targets in the captured infrared photos. Based on the GPS module (108) positioning data at the corresponding time, the geographical location of the water source, the route from the fire site, the area, the total water volume, the height, and the environmental conditions of the water source are calculated.

3. The method according to claim 2, characterized in that, Step 1-1-2 includes the following steps: Sub-step 1-1-2-1: The intelligent fire-fighting drone cluster divides the entire fire scene into patrol areas of equal number and similar size based on the number of dispatched drones, and uses a random algorithm to assign several intelligent fire-fighting drones to patrol different patrol areas. Sub-step 1-1-2-2: During the inspection, the intelligent computer (101) activates the infrared camera (103) to continuously capture infrared photos of the fire scene, and activates the microphone and speaker (104) to collect sounds from the fire scene. Furthermore, an adaptive mean filtering algorithm is used to denoise the infrared photos or videos of the fire scene acquired by the infrared camera (103). The formula is as follows: ； in It is a pixel The domain pixels, The number of pixels in the window's neighborhood. For window neighborhood, It is the grayscale value of the pixels in the infrared photos or videos of the fire scene; Sub-steps 1-1-2-3: The intelligent computer (101) segments the infrared image or video frame of the fire scene, performing threshold segmentation of the color space based on pixel hue H, saturation S, and brightness value V, using the following formula: ； Sub-step 1-1-2-4: The intelligent computer (101) filters the sound signals collected by the microphone and speaker (104), performs feature extraction, and extracts the combustion sound features and distress call features related to the ignition point. The formula is: ； in, This represents the energy of the m-th filter bank. Indicates the number of filters. It is the MFCC coefficient number of the sound signal collected by the microphone and speaker (104); Sub-step 1-1-2-5: The intelligent computer (101) uses a threshold detection algorithm to calculate the areas with higher temperatures in the infrared photos of the fire scene, and then marks them as ignition points; Furthermore, threshold edge detection is performed on the denoised image to locate the ignition point. The formula for calculating the center of the ignition point is as follows: ； in, It is the comprehensive window coefficient, used to calculate the gray value of the center pixel in a weighted calculation; Sub-step 1-1-2-6: The intelligent computer (101) uses GPS module (108) positioning data at the corresponding time, combined with the characteristics, propagation direction and intensity of the sound signal, to assist in determining the location of key points within the fire area. Furthermore, calculate the geographical location, area, height, and distribution of fire points at the fire scene, including all ignition points. Sub-step 1-1-2-7, if the intelligent computer (101) detects an area with a high temperature at the edge of the infrared photo or video of the fire scene, i.e. the ignition point, it expands the patrol range in that direction until no area with a high temperature is detected at the edge of the infrared photo of the fire scene taken during the patrol. Sub-step 1-1-2-8: The intelligent computer (101) uses a target recognition algorithm to identify targets in the infrared photos of the fire scene, calculates the living conditions of people, the distribution of combustibles, wind direction, and wind force; further, based on the target recognition and living conditions of people at the fire scene, it calculates whether it is a hazardous chemical warehouse or production area, or whether it is an explosive warehouse or production area. Sub-step 1-1-2-9, Intelligent computer (101) estimates the fire spread speed and development trend based on the geographical location, area, height, distribution of ignition points, distribution of combustibles, wind direction, and wind force of the fire scene. Furthermore, the rate of fire spread is estimated based on the properties of the combustible material, the building structure, and the ventilation conditions at the fire scene, using the following formula: ； in, To the speed of the fire's spread, The constant is related to the properties of the combustible material. The wind speed is [value], and the area of ​​the fire scene is [area]. The height is ; Sub-step 1-1-2-10: The intelligent computer (101) estimates the fire development trend based on the estimated fire spread rate and calculates the time required for the fire to develop to the critical location. The key locations mentioned are densely populated residential areas, areas with high population density, areas containing flammable, explosive, or hazardous chemicals, or other important areas requiring attention, such as bridges, tunnels, schools, and military zones; the formula for the fire development trend is: ； in, It is the distance from the ignition point to the critical location. This indicates the speed at which the fire spreads.

4. The method according to claim 2, characterized in that, Steps 1-2 include the following steps: Sub-step 1-2-1: The intelligent fire-fighting drone cluster estimates the time required to lay fire-fighting water supply channels between the fire site and the water source based on the reconnaissance information of the fire site and the surrounding water source. Based on sub-step 1-1-2, it estimates the fire intensity and development trend after the time required to lay the fire-fighting water supply channels, and calculates and determines the water consumption and the number of fire-fighting water supply channels to be laid. Further, the fire-fighting water supply channel refers to all fire hoses, pressurized water pumps and fire nozzles continuously laid from the water source to the fire site. One fire-fighting water supply channel can independently complete one fire-fighting water supply operation. Smaller fires require less water and only one fire water supply channel; larger fires require more water and multiple fire water supply channels. Furthermore, the effective coverage area of ​​one fire water supply channel or fire extinguishing material is... The calculation formula is: ； in, It is the altitude at which intelligent firefighting drones perform their missions. It refers to the spray angle of fire hoses or fire extinguishing materials; Sub-step 1-2-2: Based on the information of the fire scene and the surrounding water source, the intelligent fire-fighting drone cluster estimates the total length required to lay the fire water supply channel between the fire scene and the water source. Based on this total length, it calculates the actual required length of the fire hose (500) and reserves a certain margin. Before a fire occurs, fire hoses (500) are assembled at regular intervals to reduce assembly time in case of a fire. Furthermore, based on the actual required length of fire hoses (500), the type and quantity of booster pumps (400) to be carried are calculated. A booster pump (400) is installed at every certain distance of fire hose (500); Based on the target identification in sub-step 1-1-2, including the distribution of ignition points, combustible material distribution, hazardous chemical warehouses or production areas, and explosive material warehouses or production areas, calculate the type and quantity of fire extinguishing materials to be carried; firefighters work together to assemble fire hoses (500) and booster pumps (400) on the ground according to the required quantity and length, load the required type and quantity of fire extinguishing materials onto the corresponding intelligent fire-fighting drone, appropriately separate multiple fire water supply channels, and equip one fire nozzle (600) at the end of each fire water supply channel. The type of extinguishing material to be carried is determined based on the properties of the combustible material, and the total demand for extinguishing material is estimated using the following formula: ； in, The standard for extinguishing material dosage per unit area is given, and the fire scene area is... , It is the density of the combustible material. It is the extinguishing agent efficiency coefficient; Sub-steps 1-2-3: The intelligent fire-fighting drone cluster calculates the number of intelligent fire-fighting drones that need to be dispatched. For every certain distance of fire hose (500), 1-2 intelligent fire-fighting drones are equipped; for every booster pump (400), 1-2 intelligent fire-fighting drones are equipped; for every set of fire extinguishing materials, 1 intelligent fire-fighting drone is equipped; and for every fire nozzle (600), 1-2 intelligent fire-fighting drones are equipped. Furthermore, the formula for estimating the number of intelligent firefighting drones that need to be dispatched is: ； in, This is the safety redundancy factor; the fire scene area is... One fire water supply channel or fire extinguishing material effectively covers an area of , This indicates rounding up to the nearest integer.

5. The method according to claim 2, characterized in that, In steps 1-3, the following sub-steps are used: Sub-step 1-3-1: The intelligent fire-fighting drone cluster forms a distributed edge computing system through its respective intelligent computer (101) and remote communication module (105), sharing the geographical locations of water sources and fire scenes, as well as the current geographical locations of each intelligent fire-fighting drone. It also models the fire-fighting operation task, encodes the flight path into a feasible solution, and divides the locations from the fire command center (300), fire-fighting materials, water sources to the fire scene into grids and fixed coordinates. The grid set of the fire-fighting operation area is then used... To indicate, the first Each grid is used To indicate, Maximum number of grid cells in the firefighting operation area; intelligent firefighting drones Firefighting operation grid set accessed via flight path To indicate; Sub-steps 1-3-2 are for coordinating the dispatch of... Task allocation among intelligent firefighting drones: The intelligent firefighting drone swarm employs an improved K-means algorithm, which divides the set of flight paths for the firefighting operation area into... Each cluster serves as a set of firefighting operations, i.e. This includes transporting fire extinguishing materials, transporting fire hoses (500) and booster pumps (400), operating fire nozzles (600), and connecting the water source and the fire scene using fire hoses (500) and booster pumps (400); the fire extinguishing operations of each cluster are independent of each other, without any overlap or intersection, and must follow specific constraints: ； This refers to two different intelligent firefighting drones; Sub-step 1-3-3, in the... A collection of firefighting operations by intelligent firefighting drones In this context, the intelligent firefighting drone swarm defines an extended set. This set is composed of elements indexed by... Indicates the first Firefighting operation clusters in individual grids Collection of fire scene locations For each intelligent firefighting drone's firefighting operation cluster Construct a firefighting operation plan It consists of the locations of the fire scene. and edge set Composition; edge set It includes a collection of information from the fire command center (300), fire extinguishing materials, water sources to the location of the fire scene. All possible flight routes , , and This indicates the grid locations along the flight path, including the fire command center (300), fire extinguishing materials, water sources, and other fire scene locations. yes China is different The location of the fire scene; Sub-steps 1-3-4, firefighting operation diagrams for each intelligent firefighting drone. In the middle, each edge , Each corresponds to a non-negative weight. This weight indicates the location of the intelligent firefighting drone within the two grids at the fire scene. and The distance between the flight paths is calculated using the following formula: - ; Sub-steps 1-3-5: Setting up the intelligent firefighting drone cluster, ensuring that the intelligent firefighting drone n maintains a fixed altitude during flight. and constant flight speed; furthermore, in intelligent firefighting drone firefighting operations grouping In the communication link between the intelligent firefighting drone n and the fire command center (300), the channel power gain can be expressed as: , The power gain that the channel can provide at a distance of 1 meter from the reference point; when the fire command center is operating at constant power For data transmission, its data transmission rate on the link can be expressed as: ； Here, B represents the bandwidth that can be used during communication; This refers to the power consumed by the intelligent firefighting drone when transmitting data; This is the power gain that the channel can provide at a distance of 1 meter from the reference point; This represents the power of the noise in the channel; Sub-steps 1-3-6: Setting up the intelligent firefighting drone It will travel at a fixed speed during flight. The formula for the propulsion power generated during flight is: ； in, The power generated by the outer contour of the blade, and This represents the power generated by the induction effect; This refers to the speed at the tip of the rotor blades. It is the average speed sensed by the rotor while it is hovering; It is the ratio between fuselage drag and overall drag. Represents the density of air. This reflects the robustness of the rotor, while Y represents the area occupied by the rotor disk; Sub-steps 1-3-7, when the intelligent firefighting drone... In the selected route grid During flight, the intelligent computer (101) calculates the flight energy consumption it generates. ； Among them, intelligent firefighting drones Flight time ; For intelligent firefighting drones Fixed flight speed; Sub-step 1-3-8: To improve firefighting efficiency, the "firefighting information age" indicator is introduced to measure the firefighting timeliness of the intelligent firefighting drone swarm. This measures the total time elapsed from when the intelligent firefighting drone swarm begins capturing infrared photos or videos of the fire scene to when it arrives at the scene and begins firefighting operations. This includes the time spent capturing infrared photos or videos, analyzing those photos or videos, route planning, and flight time. The firefighting information age accurately represents the total time from the discovery of the fire scene to the current firefighting operation, thus providing a more precise description of firefighting efficiency. Furthermore, the average information age of all locations within the intelligent firefighting drone swarm is taken to further evaluate the overall firefighting efficiency of the system. The formula for calculating the average information age of these locations is: ， ； in, This refers to a cluster of intelligent firefighting drones for firefighting operations. The total number of fire scene locations included in the data; It is used to represent intelligent firefighting drones. From the firefighting operation location Fly to On the way, from the moment of leaving the fire command center (300), to arriving A parameter representing the number of all locations visited before the current location; by sub-process For example, we can get ; Sub-steps 1-3-9: The intelligent firefighting drone swarm needs to achieve a balance between the average age of firefighting information and the energy consumption of the intelligent firefighting drones. The optimal route selection formula for the rational planning of the intelligent firefighting drones' flight paths is as follows: ； in, It is a binary variable whose value can only be 0 or 1, representing the intelligent firefighting drone. Have you selected a route? , Specifically, when This means that intelligent firefighting drones We did indeed travel this route; at the same time, Representative of intelligent firefighting drones Access the total number of fire scene locations, and group all such binary variables together and label them as follows. ; Sub-steps 1-3-10 are to ensure the fire extinguishing operation diagram. Each fire scene location in the system has the same in-degree, and this value is set to 1. The intelligent fire-fighting drone swarm applies an in-degree constraint to the route selection formula: ; Sub-step 1-3-11, to ensure the diagram Each fire scene location in the system has the same out-degree, and this value is set to 1. The intelligent fire-fighting drone swarm applies out-degree constraints to the route selection formula: ; Sub-steps 1-3-12 are to ensure that the intelligent firefighting drone... Do not select route , hour hour, When the intelligent firefighting drone chooses this route... hour Sub-journey constraints are applied to the route selection formula: ， , ； In sub-step 1-3-13, the intelligent firefighting drone swarm checks whether the solution results meet the predetermined error threshold or solution time limit. If not, it returns to sub-step 1-3-1 to continue calculation until the optimal feasible solution, i.e., an optimal flight route, is obtained. The optimal flight route This refers to the intelligent firefighting drone swarm departing from its current location and arriving at its flight path. After reaching the destination, the fire hoses (500), booster pumps (400), fire nozzles (600), and fire extinguishing materials were laid between the water source and the fire scene to create one or more fire water supply channels. Sub-steps 1-3-14: The intelligent firefighting drone swarm each follows its optimal flight path. Fly to the fire scene.

6. The method according to claim 2, characterized in that, In steps 1-4, the intelligent fire-fighting drone swarm uses grappling hooks (107) to load one or more fire water supply channels that have been assembled in sub-step 1-2-2, including fire hoses (500), booster pumps (400), fire nozzles (600), and the fire extinguishing materials to be carried, according to their division of labor. After all loading is completed and confirmed to be correct, they take off at the same time and lay booster pumps and fire hoses between the water source and the fire scene according to the optimal route.

7. The method according to claim 1, characterized in that, Step 2 includes the following sub-steps: Sub-step 2-1: The intelligent fire-fighting drone cluster calculates the ignition point information and the optimal fire extinguishing plan based on infrared photos, videos or sounds of the fire scene, and calculates the optimal spraying method of the fire nozzles or the release method of the fire extinguishing materials. Sub-step 2-2: The intelligent fire-fighting drone cluster activates all pressurized water pumps and controls fire nozzles to spray water for fire extinguishing operations, or releases fire extinguishing materials for fire extinguishing operations, according to the optimal fire extinguishing plan. Sub-steps 2-3: The intelligent fire-fighting drone cluster adjusts the positions of the booster pumps and fire hoses according to changes in the fire situation at the fire scene; Sub-step 2-1 includes the following steps: Sub-step 2-1-1: The intelligent fire-fighting drone cluster activates the infrared camera (103) to continuously capture infrared photos, videos or sounds of the fire scene, calculates the ignition point information and the distribution of ignition points, and uses artificial intelligence algorithms to rate the hazard level of each ignition point; further, it predicts the development trend of the fire at the ignition point. If it is likely to develop into a larger scale or cause greater danger or be more difficult to control in the short term, it is considered high-risk, and vice versa. Sub-step 2-1-2: Based on the distribution of fire points and the hazard rating of each fire point, the intelligent fire-fighting drone cluster calculates the optimal fire extinguishing plan according to the principle of extinguishing all fire points in the shortest time. Based on the location of the fire point and the fire scene environment, the intelligent computer (101) calculates the optimal hovering position for controlling the fire nozzles or the intelligent fire-fighting drone carrying fire extinguishing materials, i.e., the optimal fire extinguishing plan, so that it can cover the fire point in the fire scene and carry out effective fire extinguishing. The formula for the optimal hovering position of the drone is: ； in, q represents the location of the fire points within the fire scene, and q represents the number of fire points. Sub-step 2-1-3: Based on the location of one or more fire water supply channels, the intelligent fire-fighting drone cluster calculates the optimal spraying method of the fire nozzles or the release method of the fire extinguishing materials according to the optimal fire extinguishing plan; furthermore, high-risk fire points should be extinguished first, low-risk fire points can be extinguished second best, and fire points where water spraying is ineffective should be considered for releasing other fire extinguishing materials. The intelligent computer (101) calculates the optimal hovering height of the intelligent fire-fighting drone based on the effective range of the fire nozzle or extinguishing material, using the following formula: ； in, Height of the ignition point The minimum safe flight altitude for intelligent firefighting drones, The effective range height of the fire extinguishing equipment; Furthermore, the intelligent computer (101) determines the relative position and altitude of the intelligent firefighting drone to the fire point. The intelligent computer (101) calculates the optimal spray angle for fire-fighting water or extinguishing materials using the following formula: 。 8. The method according to claim 7, characterized in that, Sub-step 2-2 includes the following steps: Sub-step 2-2-1: The intelligent fire-fighting drone cluster sends a remote start command to all pressurized water pumps (400) through the short-range communication module (106). After receiving the remote start command, the pressurized water pumps (400) start running. Sub-step 2-2-2: The intelligent fire-fighting drone cluster controls the fire nozzles (600) to spray water for fire-fighting operations according to the optimal fire-fighting plan; furthermore, the fire nozzles (600) have a large spray force, and 2-3 fire-fighting drones can be dispatched to control the fire nozzles (600) and the fire hoses (500) connected to their rear ends, so as to better control the effective range height and the optimal spray angle; Based on the relative position and altitude of the intelligent fire-fighting drone to the fire point, the intelligent computer (101) calculates the flow rate of fire-fighting water or fire-extinguishing materials using the following formula: ； in This refers to the amount of fire water or fire extinguishing materials required per unit area. Area of ​​the ignition point; Sub-step 2-2-3: The intelligent fire-fighting drone cluster checks for any fire points where water spraying is ineffective. If any fire is found, other fire-fighting materials should be considered for fire-fighting operations, including fire sand, carbon dioxide extinguishing agent, solid dry ice, chemical flame-retardant materials, window-breaking fire extinguishing grenade launchers and fire extinguishing grenades, ball throwers, barrel throwers, or other customized fire-fighting equipment. Sub-step 2-3-1: The intelligent fire-fighting drone cluster activates the infrared camera (103) to continuously take infrared photos of the fire scene, uses artificial intelligence algorithms to calculate the changes in the fire intensity and the distribution of ignition points, and assigns a hazard rating to each ignition point; on the one hand, the fire intensity in the vicinity may become smaller and smaller, and the water spraying may no longer be able to reach the ignition points in the distance; on the other hand, the fire intensity in the vicinity may become larger and larger, endangering the pressurized water pump (400), fire hose (500), and intelligent fire-fighting drone; Sub-step 2-3-2: The intelligent fire-fighting drone cluster recalculates the optimal fire extinguishing plan based on the distribution of fire points; Furthermore, referring to sub-steps 1-3, the intelligent fire-fighting drone swarm updates the distribution of fire points and re-searches for the optimal route and optimal fire-fighting plan between the water source and the fire scene; In sub-step 2-3-3, the intelligent fire-fighting drone cluster adjusts the positions of the booster pump and fire hose according to the recalculated optimal fire-fighting plan.

9. The method according to any one of claims 1 to 8, characterized in that, Step 3 includes the following steps: Sub-step 3-1: The intelligent fire-fighting drone cluster checks the remaining power of each drone. The drone with sufficient power takes over the fire-fighting drone with insufficient power to continue the fire-fighting operation, and controls the fire nozzles, pressurized water pumps and fire hoses. The drone with insufficient power searches for the optimal route to return to the charging station. Sub-step 3-2: The intelligent fire-fighting drone cluster checks the remaining fuel of the booster pump. If the fuel is insufficient, it carries the corresponding grade of fuel to replenish the booster pump. Sub-step 3-3: The intelligent fire-fighting drone cluster checks the remaining quantity of fire extinguishing materials. If the fire extinguishing materials are insufficient, it carries the corresponding quantity of fire extinguishing materials to replenish the fire scene. Sub-steps 3-4: The intelligent fire-fighting drone cluster uses infrared cameras to check if there are any remaining fire points at the fire scene. If not, the fire-fighting operation is considered to be over; otherwise, return to step 2 to continue the fire-fighting operation. Step 3-1 includes the following steps: Sub-step 3-1-1: The intelligent fire-fighting drones check their remaining power and share the remaining power information with the intelligent fire-fighting drone cluster through the remote communication module (105); In sub-step 3-1-2, the intelligent fire-fighting drone cluster checks the intelligent fire-fighting drones with insufficient remaining power, for example, less than 20% remaining power, measures the location of their GPS module (108), and broadcasts their remaining power and location to the entire network; further, the fire-fighting drones with sufficient power that are temporarily idle go to the location and take over the fire-fighting drones with insufficient power to continue the fire-fighting operation, and control the fire nozzles, pressurized water pumps and fire hoses; if the fire-fighting drones with insufficient power are carrying fire water tanks or fire extinguishing equipment boxes (102) and fire extinguishing materials, then the fire-fighting drones with sufficient power that are temporarily idle go to the location and take over the fire-fighting drones with insufficient power to continue the fire-fighting operation, carrying the same type and quantity of fire water tanks or fire extinguishing equipment boxes (102) and fire extinguishing materials. Sub-step 3-1-3: After the fire-fighting drone with sufficient power arrives and takes over the corresponding fire-fighting operation, the fire-fighting drone with insufficient power will search for the optimal return route and return to charge along the optimal route. Sub-step 3-1-3-1: The intelligent computer (101) of the fire-fighting drone, which is low on power, calculates the Euclidean distance from the fire scene to the fire command center (300) using the following formula: ； in, and These are the coordinates of the intelligent firefighting drone at the fire scene and the fire command center (300), respectively. Sub-step 3-1-3-2: The intelligent computer (101) of the fire-fighting drone with insufficient power obtains the remaining power of the intelligent fire-fighting drone. The system also performs dynamic return-to-home battery threshold calculations on the intelligent firefighting drone, determining whether the remaining battery power is sufficient for a successful return. The formula is as follows: ； in, Indicates the safety margin threshold. Indicates the power consumption coefficient per unit distance; If the remaining battery power of the intelligent firefighting drone is sufficient to return to base, proceed to sub-step 3-1-3-3; otherwise, if it cannot return to base, it will remain in place and maintain communication, but will no longer undertake any firefighting operations. Sub-step 3-1-3-3: The intelligent computer (101) of the fire-fighting drone with insufficient power plans the return route of the intelligent fire-fighting drone based on the remaining power using a route planning algorithm. The route cost function is: ； in, Indicates cumulative energy consumption. Represents the cost of heuristic distance. , These are weighting coefficients, and they have... + ; Sub-step 3-1-3-4: The intelligent computer (101) of the fire-fighting drone with insufficient power simultaneously applies power consumption constraints to the above functions, as shown in the formula: ； in Indicates the first The power function of the route segment Indicates the first Flight speed of the segment route, Indicates the first The length of the route segment Indicates the safety margin threshold; In sub-step 3-1-3-5, the intelligent computer (101) of the fire-fighting drone with insufficient power checks whether the solution result meets the predetermined error threshold or solution time limit. If not, it returns to sub-step 3-1-3-1 to continue calculation until the optimal feasible solution, i.e., an optimal flight route, is obtained. ; Sub-step 3-1-3-6: The intelligent firefighting drone with insufficient power follows the optimal flight path solved by the intelligent computer (101). return.

10. The method according to claim 9, characterized in that, Step 3-2 includes the following steps: In sub-step 3-2-1, during the firefighting operation, the pressurized water pump (400) continuously monitors the remaining fuel level and reports it to the nearest intelligent firefighting drone through the short-range communication module (106). The intelligent firefighting drone measures the current geographical location through the GPS module (108). Sub-step 3-2-2: When the pressurized water pump (400) detects that the remaining fuel level is lower than the threshold (e.g., 10%), it will alarm. The nearest intelligent fire-fighting drone will report the geographical location measured by the GPS module (108) and carry the nearest intelligent fire-fighting drone with the corresponding fuel grade to the location of the pressurized water pump (400) to replenish it. In sub-step 3-2-3, the intelligent fire-fighting drone carrying the corresponding grade of fuel aligns the fuel nozzle of the fire water tank or fire extinguisher box (102) with the fuel filler port of the pressurized water pump (400), and issues an instruction to the pressurized water pump (400) to open the fuel filler port through the short-range communication module (106); further, after receiving the instruction, the pressurized water pump (400) opens the fuel filler port, the intelligent fire-fighting drone carrying the corresponding grade of fuel opens the solenoid valve, and the fuel is added to the pressurized water pump (400) through the fuel nozzle; when the pressurized water pump (400) detects that the remaining fuel level is full, it commands the intelligent fire-fighting drone to close the solenoid valve and the fuel nozzle through the short-range communication module (106), and then the pressurized water pump (400) closes the fuel filler port, and the refueling ends; Step 3-3 includes the following steps: Sub-step 3-3-1: The intelligent fire-fighting drone measures the remaining quantity of fire extinguishing materials based on the pressure sensor of the fire water tank or fire extinguishing equipment box (102), and broadcasts the geographical location and remaining quantity of fire extinguishing materials measured by the GPS module (108) of the intelligent fire-fighting drone to the entire network; Sub-step 3-3-2: If the pressure sensor value of the fire water tank or fire extinguishing equipment box (102) is lower than a certain threshold, it is determined that the fire extinguishing material is insufficient. The intelligent fire-fighting drone cluster will arrange for a temporarily idle fire-fighting drone with sufficient power to carry the same model and quantity of fire water tanks or fire extinguishing equipment boxes (102) and fire extinguishing materials to go to the location and take over the fire-fighting drone with insufficient fire extinguishing materials to continue the fire-fighting operation. Sub-step 3-3-3: After the fire-fighting drone with sufficient fire-fighting materials arrives and takes over the corresponding fire-fighting operation, the fire-fighting drone with insufficient fire-fighting materials will search for the optimal return route on its own and return to charge along the optimal route. In sections 3-4, the following steps are included: In sub-steps 3-4-1 and 1-1-2, the intelligent fire-fighting drones responsible for reconnaissance use artificial intelligence algorithms to automatically divide the patrol area during the fire-fighting operation, conduct full-area coverage patrol of the entire fire scene, and continuously take infrared photos of the fire scene by turning on the infrared camera (103) during the patrol; further, the artificial intelligence algorithm is used to calculate the areas with high temperature in the infrared photos of the fire scene. If there are no areas with high temperature, such as greater than 100℃, it is determined that there are no residual fire points, and the intelligent fire-fighting drone cluster determines that the fire-fighting operation is over. If there are, it returns to step 2 to continue the fire-fighting operation. Sub-step 3-4-2: If the intelligent fire-fighting drone cluster determines that the fire-fighting operation has ended, it sends a remote stop command to all pressurized water pumps (400) through the short-range communication module (106). After receiving the remote stop command, the pressurized water pumps (400) stop running. The intelligent fire-fighting drone cluster uses the grappling hook (107) to carry the fire hose (500), pressurized water pumps (400), and fire nozzles (600) to drain excess water and select the optimal route to return. Step 3-4-2-1: The intelligent fire-fighting drone cluster receives the fire-fighting operation completion instruction confirmed by the fire command center (300), drains excess water, and the intelligent computer (101) obtains its current precise coordinates. And determine the coordinates of the fire command center. The flight paths between the two are encoded into feasible solutions; Furthermore, referring to sub-steps 1-3, the intelligent fire-fighting drone swarm searches for the optimal route between the water source and the fire scene to drain excess water, and uses grappling hooks (107) to retrieve fire hoses (500), booster pumps (400), and fire nozzles (600); furthermore, grappling hooks (107) are used to retrieve the intelligent fire-fighting drones that cannot return successfully in sub-steps 3-1-3-2; Step 3-4-2-2: The intelligent computer (101) establishes a three-dimensional map model and records the three-dimensional coordinates of each grid cell. The system uses artificial intelligence algorithms to calculate the optimal return flight route; preferably, the shortest path algorithm is used. Furthermore, the intelligent fire-fighting drone swarm refers to sub-step 3-1-3 to search for the optimal return route from the fire scene to the fire command center (300); Steps 3-4-2-3: The intelligent firefighting drone swarm returns along the optimal route to recharge, perform maintenance, and reload various fire extinguishing materials.

Images