Unmanned aerial vehicle cluster and management server thereof

Abstract

The invention relates to an unmanned aerial vehicle cluster and a management server thereof, and the management server serves as a management unit and is configured to send a collection angle and / or a collection coordinate to an unmanned aerial vehicle under the condition that angle information of a fire image in a three-dimensional virtual scene is lost, the at least two unmanned aerial vehicles collect fire image information at the same time in an angle complementation mode, so that images at different angles at the same time are obtained, fire parameter information in a three-dimensional virtual scene is perfected, and a real fire image is fitted in the three-dimensional scene. According to the invention, through fitting the real fire image, the spraying direction of the fire-fighting equipment is accurately controlled, so that the fire extinguishing completion efficiency of the fire-fighting equipment and the unmanned aerial vehicle cluster is improved.

Description

[0001] The original basis of this divisional application is the patent application with application number 202311796362.1, application date December 25, 2023, and invention name “A drone cluster management system and method based on accident emergency situations”. Technical Field

[0002] The present invention relates to the field of fire fighting technology, and in particular to a drone cluster and a management server thereof. Background Art

[0003] Drones are small in size, light in weight, and have flexible flight routes, so they are widely used in various emergency scenarios to assist in disaster relief or provide surveillance images.

[0004] In terms of emergency rescue, the advantages of drone swarms include: replacing traditional human observation methods, reducing casualties, and monitoring quickly and over a wide range. Based on the images transmitted by drone swarms, the spectral features in the fire can be analyzed and a fire recognition model can be constructed to identify fires over a large area.

[0005] When an emergency occurs, drone swarms can also quickly identify the distribution and location of people through images, providing an important basis for the formulation of rescue plans. Drones also undertake a variety of tasks in emergency scenarios, such as environmental monitoring, power grid inspection, oil pipeline inspection, and air sampling.

[0006] However, the drone swarm is uniformly controlled and commanded by a management unit, and its defects include: the management unit needs to process a large amount of data in real time. When the management unit crashes, a large number of drones will lose control and crash into surrounding buildings, causing damage to a large number of drones. Unlike daily scenarios, when an emergency occurs, the management unit needs to quickly process a large amount of information transmitted back by the drone swarm, which will cause the processing of some information to be delayed. In an emergency, especially a fire, the fire situation changes quickly, and appropriate control instructions need to be quickly generated to control the drones as the fire changes. At this time, the processing of complex data will undoubtedly affect the management unit's control over the drone group, and may even cause the management unit to crash.

[0007] The current management unit for drone swarm control only controls the drone swarm to collect images of the disaster site, and does not need to control the drone swarm to participate in firefighting operations, thus ignoring the management of emergency response to data processing and control of the drone swarm. The prior art does not yet have a management method that can control drone swarms to implement firefighting operations in industrial plants while reducing the amount of data transmission information of the management unit.

[0008] For example, the patent document with publication number CN116785611A discloses a 5G drone smart firefighting method based on edge computing, which uses edge computing terminals to control detection drones and fire-fighting drones in a drone group through 5G communication links. First, the detection drone is instructed to conduct close monitoring of the fire area to obtain fire environment information, thereby dividing the fire area into several sub-areas with different fire levels; then multiple fire-fighting drones are instructed to spray fire-fighting dry powder at fixed points in each sub-area, thereby achieving fire extinguishing treatment in each sub-area. In this patent, the computer's control of the drone is direct control. When the computer crashes or the calculation is delayed, the drone's behavior will be out of control accordingly.

[0009] The present invention hopes to provide a novel drone cluster management method and system, which can reduce the amount of data directly processed by the management unit, and even when the management unit crashes, the drone cluster can also land in an orderly manner to avoid loss of control and collision accidents.

[0010] In addition, on the one hand, there are differences in understanding among those skilled in the art; on the other hand, the applicant studied a large number of documents and patents when making the present invention, but due to space limitations, not all details and contents are listed in detail. However, this does not mean that the present invention does not have the characteristics of these prior arts. On the contrary, the present invention already has all the characteristics of the prior art, and the applicant reserves the right to add relevant prior art to the background technology. Summary of the invention

[0011] In the existing digital twin factories, the management unit must not only correct the fire parameter information in the three-dimensional simulation scene based on the fire image information and / or environmental information transmitted by the drone, but also adjust the fire fighting plan of the drone cluster according to the fire parameter information, and also solve the obstacles encountered by each drone. Obviously, the data processing volume of the management unit is huge. In the case of a large amount of data processing, the management unit may not be able to give priority to some emergency data, or may not be able to process it in time. In the case of rapid spread of fire, the calculation delay of some data of the management unit may reduce the fire fighting effect.

[0012] The prior art has already presented a technical solution for selecting or forming a corresponding drone cluster according to the cluster scheduling function requirements to improve the matching degree between the drone cluster and the task to be performed. For example, the patent document with the publication number CN108764652A discloses a drone cluster organization method and system adapted to the cluster waypoint task, wherein the drone cluster includes at least one master drone and multiple slave drones connected to the master drone in communication, and the drone cluster organization method includes: the central control system obtains the cluster waypoint task, the central control system selects a dispatchable drone cluster from each drone cluster according to the cluster scheduling function requirements, or the central control system selects multiple or all drones from multiple drone clusters according to the cluster scheduling function requirements, and temporarily forms a dispatchable drone cluster, the dispatchable drone cluster meets the cluster scheduling function requirements, and the central control system sends the cluster waypoint task to the master drone of the dispatchable drone cluster; wherein the cluster scheduling function requirements include the remaining power requirements of each drone, the drone failure requirements, and the cluster waypoint task requirements. However, in this technical solution, the drone clusters cannot exchange functions during the execution of the mission, and the specific functions they perform are determined by the pre-scheduling function requirement analysis. Under this setting, the drone clusters performing different tasks can only perform corresponding data processing and analysis processes through a unified central control system. On the one hand, it significantly increases the amount of data processing required by the central control system for data calculation, which is completely contrary to the goal of the present invention to reduce the data processing volume of the management unit to improve the response speed of handling accident emergencies; on the other hand, since each drone in this technical solution needs to transmit data back to the central control system for data processing, the drones under this setting cannot adjust the processing tasks in time according to specific environmental changes, which in turn causes delays in the handling of emergency accidents.

[0013] The present invention hopes to provide a management system and method that can reduce the data processing volume of the management unit, while also enabling drones to autonomously handle simple technical problems, thereby improving the fire-fighting effect of drone clusters.

[0014] In view of the deficiencies of the prior art, the present invention provides a drone cluster management system based on accident emergency situations, including a management unit, which establishes a communication connection with at least one drone. The management unit is configured to: correct the fire parameter information in the three-dimensional simulation scene based on the fire image information and / or environmental information sent by at least one drone, send the starting point, end point information and the overall fire fighting plan of the drone cluster's flight route to several first-level drones in the drone cluster based on the fire parameter information, and store the task allocation information of several second-level drones fed back by the first-level drone. In the case of changes in the fire parameter information, adjust the task allocation information of the second-level drone related to the change in the fire parameter and send it to the corresponding first-level drone, so that the second-level drone can adjust the fire fighting plan in time based on the control instructions of the first-level drone.

[0015] Compared with the above-mentioned prior art, the present invention can adjust the fire fighting scheme executed by the second-level drone according to the control instructions of the first-level drone. Based on the above-mentioned distinguishing technical features, the problems to be solved by the present invention may include: how to timely adjust the corresponding fire fighting scheme according to the changes in the fire parameter information to improve the response speed of the drone in performing the task. Specifically, the present invention sets the process of the drone performing the fire fighting task into two parts. First, the hierarchically arranged drones are dispatched to the task execution location in response to the fire information. During this flight, the drone is basically not disturbed by the fire information. Therefore, the amount of data that needs to be transmitted and processed between the management unit and different drones is relatively small, and it will not burden the data processing capacity of the management unit; secondly, in the process of the drone performing the fire fighting task, since the fire information changes all the time, it is necessary to collect the fire information at any time and adjust the fire fighting task of the drone according to the fire information. At this time, the data analysis and processing involved are significantly increased. If each drone interacts with the management unit for data, it will significantly increase its data processing burden, resulting in reduced efficiency of data analysis, and it is impossible to adjust the corresponding fire fighting scheme according to the fire information uploaded by each drone, which leads to the fire being unable to be controlled in time and causing it to spread. The present invention reduces the amount of data transmission and processing between the management unit and each drone by sending the starting point, end point and task to the drone by the management unit. When the firefighting plan needs to be adjusted, the first-level drone is used to calculate the specific firefighting plan and assign specific tasks, so that the first-level drone can share part of the calculation task. When the first-level drone encounters information that cannot be processed, the first-level drone can feedback to the management unit and request the management unit to process it, which will not delay the data processing in an emergency and can also ensure the normal operation of the drone cluster.

[0016] According to a preferred embodiment, during the autonomous flight of the first-level drone and the second-level drone from the starting point to the designated end point, the management unit sends the three-dimensional coordinate information and danger level of the building to be detected to the first-level drone based on the increase in the danger level of the building. In response to the first-level drone based on the three-dimensional coordinate information and danger level of the building, at least one second-level drone passing through or within the effective acquisition range of the building collects the image information and / or environmental parameter information of the building, and the management unit receives the image information and / or environmental parameter information of the building sent by the second-level drone. Here, collecting the image information and / or environmental parameter information of the building belongs to the task allocation information sent by the first-level drone to the second-level drone. Here, the first-level drone sends a work task that will not significantly increase its workload to the second-level drone passing through or within the effective acquisition range of the building, so as to achieve a better effect of obtaining the required information without occupying more second-level drones.

[0017] Compared with the above-mentioned prior art, the management unit of the present invention can instruct the drone to collect image information and / or environmental parameter information of the building where the danger level of the building changes according to the change of the danger level of the building. Based on the above-mentioned distinguishing technical features, the problems to be solved by the present invention may include: how to timely update the danger level of the building according to the actual progress of the fire. Specifically, as the spread of the fire changes, a building that is not dangerous at the beginning may become dangerous due to the ejection of nearby explosives, causing the danger level of the building to increase. Based on this situation, the present invention assigns a second-level drone to feedback the real-time image of the building as a supplement to the data based on the danger level of the building, reducing the phenomenon of prolonged image reception time caused by other more distant drones returning to the building, so that the management unit can quickly complete the information of the dangerous building and determine whether to generate a new fire protection plan.

[0018] According to a preferred embodiment, in response to the predicted impact of environmental information on the danger level of the building, the management unit sends priority fire extinguishing location information and fire extinguishing schemes to the first-level drone, and the first-level drone adjusts the fire extinguishing tasks of the second-level drone based on the priority fire extinguishing location information, the fire extinguishing scheme, the current power parameters and load types of the second-level drone. Here, adjusting the fire extinguishing tasks of the second-level drone means re-assigning tasks to the second-level drone based on the priority fire extinguishing location information, the fire extinguishing scheme, the current power parameters and load types of the second-level drone, that is, adjusting its task assignment information.

[0019] Compared with the above-mentioned prior art, the management unit of the present invention can adaptively adjust the fire-fighting task according to the predicted impact of environmental information on the danger level of the building. Based on the above-mentioned distinguishing technical features, the problems to be solved by the present invention may include: how to adjust the predetermined fire-fighting tasks according to the real-time changes of the fire to improve the fire-fighting efficiency. Specifically, with the changes in the danger level of the building and the influence of environmental information, such as the influence of wind direction and the influence of ambient temperature, the priority of fire extinguishing is different. Therefore, adjusting the fire-fighting tasks of the second-level drone can ensure that dangerous fire locations are extinguished first.

[0020] According to a preferred embodiment, based on the lack of angle information of the fire image in the three-dimensional simulation scene, the management unit sends the acquisition angle and / or acquisition coordinates to the drone to improve the fire parameter information in the three-dimensional simulation scene, and the management unit calculates the danger level of the building based on the danger source information of the building around the fire location and the distance from the fire location. The management unit can adjust the task allocation information of the first-level drone and the second-level drone according to the danger level of the building to reduce the number of damaged drones.

[0021] Since the drone is dynamically collecting fire image information, the collected fire image information belongs to two-dimensional image information, so there may be data missing when fitting the three-dimensional image information based on the two-dimensional image information collected at different angles and different times. The present invention directly commands the drone to collect fire image information according to the specified collection angle or height through the management unit, which is conducive to the management unit to quickly fit the real fire image in the three-dimensional scene.

[0022] According to a preferred embodiment, when the drone cluster is on standby, the management unit selects a drone as the first-level drone in the drone cluster based on the drone's power parameters and computing speed parameters. When the first-level drone is determined, the first-level drone broadcasts its identity to the surrounding second-level drones and establishes a communication connection with them. In response to the first-level drone and its communication connection with the second-level drone, the management unit generates a general firefighting plan based on the scale of the cluster formed by the first-level drone and the second-level drone. The general firefighting plan includes several task allocation information.

[0023] Compared with the above-mentioned prior art, the management unit of the present invention can adjust the specific classification of drones according to the current parameter information of different drones in the drone cluster, and generate corresponding fire fighting plans according to the adjusted drone cluster. Based on the above-mentioned distinguishing technical features, the problems to be solved by the present invention may include: how to determine the classification scheme of drones according to the specific situation when a fire occurs, so as to ensure the completion of the fire fighting task of the drone cluster. Specifically, the management unit preferably selects drones with sufficient power and fast speed calculation as the first-level drones, and allocates fire fighting plans according to the specific situation of the drone cluster, so as to avoid the situation where the drone cluster cannot complete the fire fighting plan, and also improve the completion efficiency of the drone cluster. In addition, the present invention generates the global behavior of the fire fighting task in a large number of local interactions through the interaction between the management unit and at least one first-level drone, and through the interaction between the first-level drone and several second-level drones, thereby greatly improving the data transmission efficiency, and can also greatly improve the reliability of data transmission, and avoid the problem of paralysis of the drone cluster control system caused by the overload of the calculation amount of the management unit.

[0024] According to a preferred embodiment, when the second-level drone is disconnected from the first-level drone or the second-level drone is unable to perform the fire-fighting task, the second-level drone actively changes to the first-level drone and feeds back the fire-fighting task information and its own parameter information to the management unit. The management unit changes its fire-fighting task based on the fire-fighting task information and the parameter information of the first-level drone, which is equivalent to adjusting the task allocation information. The fire-fighting task information here belongs to the assigned task. The present invention avoids the information delay phenomenon caused by the processing and feedback of all information by the first-level drone by allowing the first-level drone and the second-level drone to automatically switch identities based on needs. Information that requires urgent processing is processed by the management unit first, which is more conducive to the reasonable allocation of tasks in the drone cluster.

[0025] According to a preferred embodiment, when the current computational load of the first-level UAV is greater than the computational load threshold, the first-level UAV sends the information to be computed to the management unit and waits for its feedback information. This enables the first-level UAV to quickly determine the task allocation information for the second-level UAV, avoiding long waiting between the first-level UAV and the second-level UAV. This setting is also to avoid the phenomenon of delayed information feedback caused by the first-level UAV processing a large amount of data, and improve the efficiency of emergency firefighting of UAVs.

[0026] According to a preferred embodiment, the system also includes standby servers distributed in different areas. When the current computing amount of the first-level drone is greater than the computing amount threshold, the first-level drone connects to the standby server to wake up the server, and sends the information to be calculated to the management unit and waits for its feedback information. Using the server to help the first-level drone formulate a task allocation plan and form task allocation information can avoid delaying the timing of the second-level drone to perform emergency tasks.

[0027] The present invention sets up an additional server to process non-urgent data information for the first-level drone and the management unit, thereby reducing the time for delayed data feedback.

[0028] The present invention also provides a method for managing a drone cluster based on an accident emergency from a second aspect, the method comprising: correcting the fire parameter information in a three-dimensional simulation scene based on the fire image information and / or environmental information sent by at least one drone, sending the starting point, end point information and the overall fire fighting plan of the flight route of the drone cluster to several first-level drones in the drone cluster based on the fire parameter information, and storing the task allocation information of several second-level drones fed back by the first-level drones. In the case of changes in the fire parameter information, the task allocation information of the second-level drones related to the changes in the fire parameters is adjusted and sent to the corresponding first-level drones, so that the second-level drones can adjust the fire fighting plan in time based on the control instructions of the first-level drone.

[0029] The management method of the present invention has the advantages that the comprehensive data processing volume of the management unit is reduced, so that the management unit can give priority to processing urgent data information, thereby improving the fire extinguishing efficiency of the fire fighting plan of the drone.

[0030] According to a preferred embodiment, the method also includes: in response to the predicted impact of environmental information on the danger level of the building, sending priority fire extinguishing location information and fire extinguishing plans to the first-level UAV, and the first-level UAV adjusting the fire extinguishing task of the second-level UAV based on the priority fire extinguishing location information, the fire extinguishing plan, the current power parameters and load type of the second-level UAV, that is, the task allocation information.

[0031] The management method of the present invention arranges priority fire extinguishing location information and fire extinguishing plans according to the prediction of the building danger level, which is beneficial to avoid the expansion of the spread of fire and reduce the number of explosion accidents and casualties in the industrial park. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a simplified schematic diagram of module connection relationships of the drone cluster management system provided by the present invention;

[0033] Figure 2It is a schematic diagram of the drone identity conversion of the drone cluster management system provided by the present invention;

[0034] Figure 3 It is a flow chart of the drone cluster management method provided by the present invention.

[0035] Reference numerals list

[0036] 100: Management unit; 200: UAV; 201: First-level UAV; 202: Second-level UAV. DETAILED DESCRIPTION

[0037] The following is a detailed description with reference to the accompanying drawings.

[0038] In the prior art, the management methods of drone groups used in industrial plants generally include centralized management and decentralized management. The centralized management method means that the drones 200 are uniformly controlled and commanded by the management unit 100. The disadvantage of this is that a large amount of data information needs to be transmitted between the management unit 100 and the drones 200, and the management unit 100 needs to process the data information in a timely manner to cope with rapidly changing accident situations, which will cause the management unit 100 to delay the processing of some data information, thereby failing to respond quickly to some emergency situations.

[0039] Decentralized management refers to a management method that does not require the unified control and management of the drone ground station or the management unit 100, and the drone 200 performs tasks autonomously. The advantage of this is that less data information is transmitted between the management unit 100 and the drone 200. The disadvantage is that in an emergency situation where an accident occurs, the scene changes rapidly, and the drone 200 alone cannot meet the needs of the accident scene in an emergency.

[0040] For the digital twin factory, the management unit 100 needs to generate a three-dimensional simulation scene that is consistent with the physical factory. Industrial plant areas are different from ordinary residential areas, where a large number of flammable and explosive materials are placed. The distribution areas of these hazardous materials and the danger levels of buildings are information that the management unit 100 stores and can directly call. If the drone 200 performs decentralized autonomous management, then in the event of an emergency, the drone 200 will not be able to respond to emergencies in a timely manner because it cannot quickly determine the dangerous area.

[0041] Fire is a key protection situation in industrial parks, and it is also an accident situation that needs to be dealt with urgently in industrial parks. When a fire occurs, the drone 200 needs to fly according to a preset route and perform firefighting tasks. How to enable the drone 200 to reduce the amount of data transmission between the management unit 100 and the management unit 100, and reduce the processing of non-emergency data of the management unit 100, while enabling the drone 200 to obtain effective control instructions from the management unit 100 when necessary is the pain point of the current drone cluster performing firefighting tasks. The present invention hopes to change the traditional management method of the management unit 100 for drone cluster firefighting, and reduce the amount of data transmission between the drone 200 and the management unit 100 while ensuring that the drone cluster can effectively extinguish the fire.

[0042] The present invention can be a drone cluster management system and method based on accident emergency situations, or a drone cluster firefighting system and method. The present invention can also be a firefighting emergency system and method for a digital twin factory.

[0043] Management unit 100: can include only the management server of the digital twin factory, or can be formed by a combination of the drone ground station and the management server of the digital twin factory. Management unit 100 includes several servers, dedicated integrated chips and their combination. Management unit 100 is used to execute the coding program of the drone cluster management method of the present invention.

[0044] Drone swarm: refers to a formation consisting of several drones 200. When a fire occurs, drone swarms have the advantage of aerial firefighting, and can promptly carry out aerial firefighting with appropriate firefighting measures in the initial stage of the fire, so as to extinguish the fire or reduce the scope of the fire spread within the golden time of discovering the fire, and buy time for subsequent firefighters to extinguish the fire.

[0045] UAV 200: including detection UAV, heavy-load UAV, light-load UAV and other types of UAV.

[0046] The first-level drone 201 refers to a drone that is directly connected to the management unit 100 for communication.

[0047] The second-level drone 202 refers to a drone that is directly connected to the first-level drone 201 for communication.

[0048] like Figure 2 As shown, the first-level drone 201 and the second-level drone 202 can be converted to each other. The first-level drone 201 triggers a mode change instruction when the power is insufficient, the computing power is insufficient, it is damaged, or it is disconnected from the management unit 100 for more than a preset communication time threshold, and its own management module automatically changes to the second-level drone 202 based on the mode change instruction.

[0049] When the second-level drone 202 loses connection with the first-level drone 201 for more than a preset communication time threshold, the mode change instruction is triggered, and its own management module automatically changes to the first-level drone 201 based on the mode change instruction and sends the identity corresponding to the first-level drone 201 to the management unit 100. The management unit 100 confirms the identity of the first-level drone 201 based on the identity and establishes a communication connection relationship with the management unit 100.

[0050] Preferably, when the second-level drone 202 cannot obtain a timely response from the first-level drone 201 (exceeds the waiting threshold), the second-level drone 202 directly sends a request message with an identity identifier to the management unit 100, and the management unit 100 prioritizes processing the request message of the second-level drone 202 based on the identity identifier information.

[0051] Example 1

[0052] In view of the shortcomings of the prior art, the present invention provides a drone cluster management system based on accident emergency situations, such as Figure 1 As shown, it includes a management unit 100. The management unit 100 establishes a communication connection with at least one drone 200. Generally, several drones 200 are parked in drone ground stations or dedicated areas. There are a small number of drones 200 that conduct safety inspections on industrial parks on a daily basis to find fire points in time. Therefore, in general, the data processing volume of the management unit 100 of the digital twin factory is relatively small. When the drone 200 used for inspection finds a fire, the management unit 100 needs to generate a fire image in a three-dimensional simulation scene based on environmental information and fire image information, and generate or adjust the fire fighting plan of the drone cluster based on the development of the fire. In this case, the data processing volume of the management unit 100 will suddenly explode. When multiple types of calculations are carried out simultaneously, the calculation speed of the management unit 100 on the data will be reduced, and the requests of each drone 200 will not be fed back in time, thereby resulting in the inability to improve the efficiency of drone collaborative firefighting. Unlike ordinary disasters, a large number of flammable and explosive materials are placed in industrial parks. If the fire is not controlled in time and spreads to the vicinity of the warehouse where flammable and explosive materials are stored, it will cause a larger-scale fire and explosion. After a fire occurs, the ability of a person to move is obviously slower than that of the drone 200. If the management unit 100 controls the drone cluster in time to automatically extinguish the fire, it will undoubtedly play an important role before the arrival of firefighters. The accurate extinguishing of the fire by the drone cluster at the fire location is an effective means to prevent the spread of the fire range. Therefore, the management system and method of the present invention can play an important role in the safety management of industrial parks.

[0053] The management unit 100 of the present invention is configured to correct the fire parameter information in the three-dimensional simulation scene based on the fire image information and / or environmental information sent by at least one drone 200. Correcting the fire parameter information in the three-dimensional simulation scene without human intervention is of great significance. In the digital twin factory, the equipment in the three-dimensional simulation scene is consistent with the equipment in the physical factory. Therefore, only when the management unit 100 collects enough fire parameter information can it command the drone cluster to extinguish the fire with higher efficiency.

[0054] Specifically, the management unit 100 obtains parameter information such as the fire location, fire range, and fire height from the fire image information sent back by the drone 200, and extracts environmental information such as wind direction, wind speed, air humidity, and temperature that have an impact on the fire from the environmental information. The management unit 100 can predict the impact of the environment on the fire according to a preset algorithm, thereby determining the safety level of the buildings around the fire. By inputting the environmental information parameters and the fire parameter information into a pre-set fire change model, the direction of fire spread and the danger level of the surrounding buildings can be predicted. The fire change model can be formed by training a large amount of sample data and a neural network model.

[0055] Preferably, based on the lack of angle information of the fire image in the three-dimensional simulation scene, the management unit 100 sends the acquisition angle and / or acquisition coordinates to the drone 200 to improve the fire parameter information in the three-dimensional simulation scene. Since the drone 200 is dynamically acquiring fire image information, the acquired fire image information belongs to two-dimensional image information. Therefore, there may be a phenomenon of missing data when fitting the three-dimensional image information based on the two-dimensional image information acquired at different angles and at different times. The present invention directly commands the drone 200 to acquire fire image information according to the specified acquisition angle or height through the management unit 100, which is conducive to the management unit 100 quickly fitting a real fire image in the three-dimensional scene.

[0056] Preferably, the management unit 100 sends acquisition angles and / or acquisition coordinates to at least two drones 200, so that multiple drones 200 can simultaneously acquire fire image information at different angles. Existing drones 200 carry visual cameras, which have certain angle acquisition defects. Therefore, multiple drones 200 simultaneously acquire fire image information in a complementary manner, so that the management unit 100 obtains images at different angles at the same time, reducing the difficulty of the management unit 100 fitting a three-dimensional fire image. Preferably, the drone 200 can determine parameter information such as the fire location, fire range, and fire height based on spectral characteristics. The management unit 100 can also predict the direction in which the fire may spread through the smoke spread characteristics and environmental parameters in the fire image information.

[0057] Generally, without the participation and control of the management unit 100, the fire fighting equipment in the industrial plant can only be triggered and spray water based on environmental parameters. For example, the fire point is passively sprayed according to the ambient temperature or high-temperature objects detected by infrared. The disadvantage of this is that the fire environment cannot be actively eliminated, or the spraying angle is wrong, and the fire cannot be extinguished efficiently.

[0058] Preferably, in the present invention, when determining the fire parameter information, the management unit 100 adjusts the parameters of the fire fighting equipment at the fire scene, such as adjusting the spray angle and spray force of the water sprinkler, so that the fire fighting equipment can aim at the fire point and implement fire fighting. The management unit 100 can also start the corresponding fire fighting equipment in advance based on the direction of fire spread to reduce the ambient temperature, or form a fire isolation zone by controlling the spraying behavior of the fire fighting equipment to hinder the spread of the fire.

[0059] Preferably, the management unit 100 calculates the danger level of the building based on the danger source information of the buildings around the fire location and the distance from the fire location.

[0060] Preferably, the danger level of the building can be calculated and stored in advance by simulating the location of the fire point and the fire intensity. When a real fire occurs, it is only necessary to fine-tune the danger level of the building according to the environmental parameters and distance information.

[0061] The management unit 100 stores parameters such as the types of dangerous sources in the surrounding buildings, coordinate information, etc., so it can quickly calculate the danger level of the building based on the scope of fire spread, environmental information, and the distance between the fire and the building. For example, the safety level is calculated by superimposing the level parameters corresponding to various parameters. Since the calculation method is simple and mature, it will not be described in detail here. The calculation method is preset in the management unit 100.

[0062] Preferably, the management unit 100 adjusts the fire extinguishing array, fire extinguishing angle and fire extinguishing materials of the first-level UAV 201 and the second-level UAV 202 for coordinated fire extinguishing based on the fire parameter information and environmental information. The fire extinguishing materials include fire-fighting materials such as water, dry powder, fire extinguishing liquid, sand, etc.

[0063] Preferably, the management unit 100 adjusts the throwing angles and types of fire-fighting materials thrown by the first-level drone 201 and the second-level drone 202 based on the fire point characteristics, wind direction and wind force in the fire parameter information.

[0064] For example, according to the fire image in the three-dimensional simulation scene, the fire area is long and there are three fire points with large fires, which should be treated as a priority to form a fire isolation area. If the wind is small on that day, the drone 200 can be instructed to use fire extinguishing liquid or water to extinguish the fire. If the wind is strong on that day, the drone 200 can be instructed to extinguish the fire by throwing sandbags. Sandbags are heavy and not easily affected by wind, and are suitable for use when the wind is strong. The management unit 100 sends the location coordinates of the three fire points and the fire fighting plan of the drone cluster to throw sandbags in three groups to the first-level drone 201. The first-level drone 201 performs specific task allocation based on the number and type of the second drone 202 communicating with itself. Finally, three groups of drones (including the first-level drone 201 itself and the second-level drone 202) can be realized to throw sandbags to the three fire points. After the throwing behavior, the first-level drone 201 feedbacks the task completion signal to the management unit 100. Preferably, during the firefighting process, a number of detection drones 200 hover over the fire point in an autonomous flight mode to collect fire image information.

[0065] Specifically, when the drone cluster is on standby, the management unit 100 selects the drone 200 as the first-level drone 201 in the drone cluster based on the power parameters and calculation speed parameters of the drone 200. The management unit 100 preferentially selects the drone 200 with sufficient power and fast calculation speed as the first-level drone 201, and allocates the fire fighting plan according to the specific situation of the drone cluster, so as to avoid the situation that the drone cluster cannot complete the fire fighting plan, and also improve the completion efficiency of the drone cluster.

[0066] When a number of first-level drones 201 are identified, the first-level drone 201 broadcasts the identity to the surrounding second-level drones 202 and establishes a communication connection with them. This arrangement can avoid the situation where the communication confusion between the several drones 200 is avoided. In the case of not being designated as the first-level drone 201, multiple drones 200 are awakened from the standby state and establish a communication connection with the first-level drone 201 based on the broadcast identity, so as to be controlled and commanded by the first-level drone 201, which is conducive to the orderly implementation of fire-fighting by the several second-level drones 202.

[0067] In response to the first-level drone 201 and its communication connection with the second-level drone 202, the management unit 100 generates a general firefighting plan based on the scale of the cluster formed by the first-level drone 201 and the second-level drone 202. Specifically, the management unit 100 can set a more appropriate general firefighting plan only after it understands the number and types of drones 200 in the drone cluster. In the general firefighting plan, the number of drones required to perform a certain type of firefighting task, the type of firefighting task, the starting point and the moving point will be allocated, but no specific drones will be specified. This can reduce the amount of data processing by the management unit 100 for the drone cluster.

[0068] Preferably, when the fire parameter information is determined, the management unit 100 sends the starting point and end point information of the flight route of the drone cluster and the overall fire fighting plan to several first-level drones 201 in the drone cluster based on the fire parameter information.

[0069] Preferably, when there are three drone clusters, the management unit 100 generates differentiated fire fighting plans based on the types, numbers, and current power parameter information of the drones in the drone cluster and sends them to the corresponding first-level drones 201 respectively.

[0070] The types and sizes of drones in the drone cluster of the present invention are naturally formed by the first-level drone 201 through communication, which makes it easy to quickly form a group and fly to the indicated destination according to the current parking area advantage. This has the advantage that a firefighting plan with a large variety of firefighting tasks can be formed to avoid the singleness of firefighting tasks, so that the effect of integrated firefighting can be achieved. Since the identities of the first-level drone 201 and the second-level drone 202 in the present invention can be converted to each other, the drone 200 awakened from the standby state will automatically convert to the first-level drone 201 and receive the firefighting plan from the management unit 100 if it cannot communicate with the first-level drone 201, so there is no situation where some drones 200 cannot receive firefighting tasks.

[0071] The management unit 100 sends the start point, end point information and the overall firefighting plan to the first-level drone 201, so that the first-level drone 201 and the second-level drone 202 can autonomously determine the flight plan route and the details of the firefighting task, using the computing power of the dedicated integrated chip inside the drone 200. Preferably, both the first-level drone 201 and the second-level drone 202 of the present invention are provided with an algorithm capable of realizing autonomous flight.

[0072] Algorithms for achieving autonomous flight include, for example, genetic algorithms, evolutionary strategies, differential evolution, and particle swarm optimization algorithms, etc. Preferably, the use of the adaptive covariance matrix evolutionary strategy algorithm (CMA-ES) can also achieve autonomous flight of the UAV. The adaptive covariance matrix evolutionary strategy algorithm (CMA-ES) is a new evolutionary algorithm proposed by Nikolaus Hansen et al. It achieves the purpose of optimization by simulating the biological evolution process in nature. The results of multiple test functions show that the algorithm has the characteristics of good global performance and high optimization efficiency.

[0073] The management unit 100 stores the task allocation information of the first-level drone 201 for several second-level drones 202. The task allocation information refers to the specific work task information that the first-level drone or the second-level drone is instructed to perform. The task allocation information includes: specific work allocation information without specifying a specific drone. For example, ten second-level drones 202 carry ten bags of sand. The task information also includes: work task information when a specific drone is specified. For example, two second-level drones 202 numbered 10547 and 10385 each carry a bag of sand.

[0074] The advantage of this setting is that when the management unit 100 adjusts part of the fire-fighting tasks of the drone cluster based on changes in fire parameters, it does not need to adjust the fire-fighting tasks through the first-level drone 201, but directly specifies the corresponding second-level drone 202 through the stored task allocation information and controls it to modify the task allocation information, so as to reduce the information transmission process and improve the efficiency of information adjustment.

[0075] In the case of changes in fire parameter information, the task allocation information of the second-level UAV 202 related to the change in fire parameters is adjusted and sent to the corresponding first-level UAV 201 and / or second-level UAV 202. For example, it can be sent only to the first-level UAV 201, or it can be sent to the first-level UAV 201 and the second-level UAV 202 at the same time, so that the second-level UAV 202 can adjust the fire fighting plan in time based on the control instructions of the first-level UAV 201 or the management unit 100. The change of fire parameters is rapid, and the direct transmission of emergency information by the management unit 100 can enable the second-level UAV 202 to quickly adjust the fire fighting task.

[0076] The present invention reduces the amount of data transmission and processing between the management unit 100 and each drone by sending the starting point, end point and task to the drone by the management unit 100. When the firefighting plan needs to be adjusted, the specific firefighting plan is calculated and the specific task is assigned by the designated first-level drone 201, so that the first-level drone 201 can share part of the calculation task. When the first-level drone 201 encounters information that cannot be processed, the first-level drone 201 can feedback to the management unit 100 and request the management unit 100 to process it, which will not delay the data processing in an emergency and can also ensure the normal operation of the drone cluster.

[0077] Preferably, during the autonomous flight of the first-level drone 201 and the second-level drone 202 from the starting point to the designated end point, the management unit 100 sends the three-dimensional coordinate information and the danger level of the building to be detected to the first-level drone 201 based on the increased danger level of the building. In response to the first-level drone 201 collecting the image information and / or environmental parameter information of the building based on the three-dimensional coordinate information and the danger level of the building, at least one second-level drone 202 passing through or within the effective collection range of the building collects the image information and / or environmental parameter information of the building, and the management unit 100 receives the image information and / or environmental parameter information of the building sent by the second-level drone 202.

[0078] For example, when a building storing nitration waste spontaneously combusts due to temperature rise, its danger level rises rapidly and an explosion may occur at any time. At this time, the management unit 100 sends the identification information and coordinate information of the highest danger level of the building to the first-level drone 201, and designates two second-level drones 202 to fly at a specified altitude and circle the building to collect image information of the building at complementary angles at the same time. This makes it easier for the management unit 100 to determine whether the building needs firefighters to enter and extinguish the fire. If the temperature of the overall environment of the building is high, the management unit 100 can give priority to sending firefighting plans that can cooperate with each other to several first-level drones 201 to give priority to drone firefighting to reduce casualties among firefighters.

[0079] As the fire spreads, a building that was not dangerous at the beginning may become dangerous due to the ejection of nearby explosives, causing the building's danger level to rise. Based on this situation, the present invention assigns a second-level drone 202 to feed back real-time images of the building based on the building's danger level as a supplement to the data, reducing the phenomenon of extended image reception time caused by other distant drones 200 returning to the building, so that the management unit 100 can quickly improve the information of the dangerous building and determine whether to generate a new fire protection plan.

[0080] Preferably, based on the danger level and building height of the building, the management unit 100 sends the specified flight altitude and the safe planned route of the flyable area to the first-level drone 201, and the first-level drone 201 determines whether the second-level drone 202 flies to the end point autonomously or according to the safe planned route at the specified flight altitude and flyable area based on the current power parameters and load parameters of the second-level drone 202. This can avoid the situation where some second-level drones 202 with insufficient power lose connection or crash due to insufficient power.

[0081] Preferably, in response to the predicted impact of environmental information on the danger level of the building, the management unit 100 sends priority fire extinguishing location information and a fire extinguishing plan to the first-level UAV 201, and the first-level UAV 201 adjusts the fire extinguishing task of the second-level UAV 202 based on the priority fire extinguishing location information, the fire extinguishing plan, the current power parameters of the second-level UAV 202, and the load type.

[0082] With the change of the danger level of the building and the influence of environmental information, such as the influence of wind direction and ambient temperature, the priority of fire extinguishing is different. Therefore, adjusting the fire extinguishing task of the second-level drone 202 can ensure that the dangerous fire location is extinguished first.

[0083] Preferably, when the second-level drone 202 is disconnected from the first-level drone 201 or the second-level drone 202 is unable to perform the fire-fighting task, the second-level drone 202 actively changes to the first-level drone 201 and feeds back the fire-fighting task information and its own parameter information to the management unit 100, and the management unit 100 changes its fire-fighting task based on the fire-fighting task information and the parameter information of the first-level drone 201. The present invention avoids the information delay phenomenon caused by the processing and feedback of all information by the first-level drone 201 by making the first-level drone 201 and the second-level drone 202 automatically switch identities based on needs. For information that needs to be processed urgently, it is processed by the management unit 100 first, which is more conducive to the reasonable allocation of tasks of the drone cluster.

[0084] Preferably, when the current computational load of the first-level drone 201 is greater than the computational load threshold, the first-level drone 201 sends the information to be computed to the management unit 100 and waits for its feedback information. This arrangement is also to avoid the phenomenon of delayed information feedback caused by the first-level drone 201 processing a large amount of data, and to improve the efficiency of emergency firefighting of the drone cluster.

[0085] Preferably, the system also includes standby servers distributed in different areas. The server is generally in a standby state or processes other data in the industrial park. When the current computing amount of the first-level drone 201 is greater than the computing amount threshold, the first-level drone 201 connects to the standby server to wake up the server. For example, the first-level drone 201 sends a communication request message containing a confidentiality code to the server to request to establish a communication connection. After the communication connection is successful, the first-level drone 201 sends the information to be calculated to the management unit 100 and waits for its feedback information.

[0086] Alternatively, in the event of a fire, the management unit 100 wakes up the servers around the fire location and puts them into a waiting communication state so that they are ready to establish a communication relationship with the first-level drone 201 and share the task of data processing at any time. The present invention sets up additional servers to process non-urgent data information for the first-level drone 201 and the management unit 100, reducing the time for delayed data feedback.

[0087] This configuration can also reduce the amount of processing by the management unit 100 for non-urgent data.

[0088] Example 2

[0089] This embodiment is a further improvement on Embodiment 1, and the repeated contents will not be repeated here.

[0090] The present invention also provides a method for managing a drone cluster based on an emergency situation. Figure 3 As shown, the method includes:

[0091] S1: Correcting fire parameter information in a three-dimensional simulation scene based on fire image information and / or environmental information sent by at least one UAV 200.

[0092] S1.1: Based on the lack of angle information of the fire image in the three-dimensional simulation scene, the management unit 100 sends the acquisition angle and / or acquisition coordinates to the drone 200 to improve the fire parameter information in the three-dimensional simulation scene.

[0093] S2: Based on the fire parameter information, the starting point and end point information of the flight route of the drone cluster and the overall fire fighting plan are sent to several first-level drones 201 in the drone cluster.

[0094] S2.1: When the drone cluster is on standby, the management unit 100 selects the drone 200 as the first-level drone 201 in the drone cluster based on the drone's power parameters and calculation speed parameters. When the first-level drone 201 is determined, the first-level drone 201 broadcasts its identity to the surrounding second-level drones 202 and establishes a communication connection with them. In response to the first-level drone 201 and its communication connection with the second-level drone 202, the management unit 100 generates a general firefighting plan based on the scale of the cluster formed by the first-level drone 201 and the second-level drone 202.

[0095] S2.2: When the second-level drone 202 is disconnected from the first-level drone 201 or the second-level drone 202 is unable to perform the fire-fighting task, the second-level drone 202 actively changes to the first-level drone 201 and feeds back the fire-fighting task information and its own parameter information to the management unit 100.

[0096] The management unit 100 changes the fire extinguishing task based on the fire extinguishing task information and the parameter information of the first-level drone 201.

[0097] S3: storing the task allocation information of the first-level UAV 201 to the second-level UAVs 202,

[0098] S4: In the event of a change in fire parameter information, adjust the task allocation information of the second-level UAV 202 related to the fire parameter change and send it to the corresponding first-level UAV 201, so that the second-level UAV 202 can adjust the fire fighting plan in time based on the control instructions of the first-level UAV 201.

[0099] The management method of the present invention has the advantages that the comprehensive data processing volume of the management unit 100 is reduced, so that the management unit 100 can give priority to processing urgent data information, thereby improving the fire extinguishing efficiency of the fire fighting plan of the drone 200.

[0100] S4.1: The management unit 100 calculates the danger level of the building based on the danger source information of the buildings around the fire location and the distance from the fire location.

[0101] S4.2: In response to the predicted impact of environmental information on the danger level of the building, priority fire extinguishing location information and fire extinguishing plan are sent to the first-level UAV 201. The first-level UAV 201 adjusts the fire extinguishing task of the second-level UAV 202 based on the priority fire extinguishing location information, the fire extinguishing plan, the current power parameters and load type of the second-level UAV 202.

[0102] The management method of the present invention arranges priority fire extinguishing location information and fire extinguishing plans according to the prediction of the building danger level, which is beneficial to avoid the expansion of the spread of fire and reduce the number of explosion accidents and casualties in the industrial park.

[0103] S5: When the current computational load of the first-level UAV 201 is greater than the computational load threshold, the first-level UAV 201 sends the information to be computed to the management unit 100 and waits for its feedback information.

[0104] It should be noted that the above-mentioned specific embodiments are exemplary, and those skilled in the art can come up with various solutions inspired by the disclosure of the present invention, and these solutions also belong to the disclosure scope of the present invention and fall within the protection scope of the present invention. Those skilled in the art should understand that the present invention specification and its drawings are illustrative and do not constitute a limitation on the claims. The scope of protection of the present invention is defined by the claims and their equivalents. The present invention specification contains multiple inventive concepts, such as "preferably", "according to a preferred embodiment" or "optionally", which means that the corresponding paragraph discloses an independent concept, and the applicant reserves the right to file a divisional application based on each inventive concept.

Claims

1. A management server for a drone cluster, characterized in that: The management server serves as a management unit (100), which is configured as follows: In the case where angle information of a fire image in a three-dimensional virtual scene is missing, a collection angle and / or collection coordinates are sent to a drone (200), so that at least two drones (200) simultaneously collect fire image information in a complementary angle manner, thereby obtaining images at different angles at the same time, improving fire parameter information in the three-dimensional virtual scene, and fitting a real fire image in the three-dimensional scene.

2. The management server according to claim 1, characterized in that: The management server is also configured to: After determining the real fire parameter information, adjust the parameters of the fire fighting equipment at the fire scene, or, Predict the direction of fire spread, start corresponding fire-fighting equipment in advance based on the fire spread direction, and control the spraying behavior of the fire-fighting equipment to form a fire isolation zone to hinder the spread of the fire.

3. The management server according to claim 1 or 2, characterized in that: The management server is also configured to: Adjusting the fire extinguishing array, fire extinguishing angle and fire extinguishing materials for the coordinated fire extinguishing of the first-level UAV (201) and the second-level UAV (202) based on the fire parameter information and the environmental information; The first-level drone (201) refers to a drone that is directly connected to the management unit (100) for communication; The second-level drone (202) refers to a drone that is directly connected to the first-level drone (201) for communication.

4. The management server according to claim 3, characterized in that: The management server is also configured to: The throwing angle and throwing type of fire-fighting materials thrown by the first-level drone (201) and the second-level drone (202) are adjusted based on the characteristics of the ignition point, the wind direction and the wind force in the fire parameter information.

5. The management server according to claim 3, characterized in that: The management server is also configured to: During the process of firefighting by the first-level drone (201) and the second-level drone (202), fire image information collected by a number of detection drones (200) hovering over the fire point in an autonomous flight mode is received.

6. A drone swarm based on a digital twin factory, characterized in that: It includes a first-level drone (201) and a second-level drone (202), The first-level drone (201) refers to a drone that is directly connected to the management unit (100) for communication; The second-level drone (202) refers to a drone that is directly connected to the first-level drone (201) for communication; In the case where angle information of a fire image in a three-dimensional virtual scene is missing, the first-level drone (201) receives a collection angle and / or a collection coordinate, so that the first-level drone (201) and / or the second-level drone (202) simultaneously collect fire image information in a complementary angle manner, so that the management unit (100) obtains images at different angles at the same time, improves fire parameter information in the three-dimensional virtual scene, and fits a real fire image in the three-dimensional scene.

7. The drone cluster according to claim 6, characterized in that: When the fire parameter information is determined, the management unit (100) sends the start and end point information of the flight route of the drone cluster and the overall fire fighting plan to a plurality of first-level drones (201) in the drone cluster based on the real fire parameter information; The first-level drone (201) and the second-level drone (202) autonomously determine the flight plan route and details of the fire-fighting task, wherein the first-level drone (201) and the second-level drone (202) are provided with algorithms capable of realizing autonomous flight.

8. The drone cluster according to claim 6 or 7, characterized in that: During the process of the first-level drone (201) and the second-level drone (202) autonomously flying from a starting point to a designated end point, the management unit (100) sends the three-dimensional coordinate information and the danger level of the building to be detected to the first-level drone (201) based on the increase in the danger level of the building; In response to the three-dimensional coordinate information and danger level of the building by the first-level drone (201), at least one second-level drone (202) passing through or within the effective collection range of the building collects image information and / or environmental parameter information of the building, and the second-level drone (202) sends the image information and / or environmental parameter information of the building to the management unit (100).

9. The drone cluster according to any one of claims 6 to 8, characterized in that: In response to the information of the designated flight altitude and the safe planned route of the flyable area sent by the management unit (100), the first-level drone (201) determines based on the current power parameters and load parameters of the second-level drone (202): The second-level drone (202) autonomously flies to the destination at a specified flight altitude and a flyable area, or the second-level drone (202) flies to the destination according to a safe planned route.

10. The drone cluster according to claim 9, characterized in that: In the case of disconnection with the first-level drone (201) or inability to perform a fire-fighting task, the second-level drone (202) actively changes to the first-level drone (201) and feeds back fire-fighting task information and its own parameter information to the management unit (100), and the management unit (100) changes its fire-fighting task based on the fire-fighting task information and the current parameter information of the first-level drone (201).

Images