Military training system based on artificial intelligence and big data

Abstract

The invention discloses a military training system based on artificial intelligence and big data, relates to the technical field of unmanned aerial vehicle automation, and solves the problems that an unmanned aerial vehicle lacks the capability of automatically executing tasks and cannot be adjusted in time when encountering obstacles in autonomous flight with lost signals. Comprising a training modeling module, a simulation training module, an obstacle perception module, a real training module and a target identification module, the training modeling module is used for converting three-dimensional geographic data of an initial area into a three-dimensional model, and the simulation training module is used for carrying out simulation training on an unmanned aerial vehicle model. The obstacle sensing module is used for identifying and avoiding obstacles in a flight path in the flight process of the unmanned aerial vehicle, the path planning module is used for updating the flight path of the unmanned aerial vehicle according to a path updating instruction and obstacle coordinates, and the real training module is used for adjusting the load of the unmanned aerial vehicle. And the target identification module is used for the unmanned aerial vehicle to automatically identify the identification target in the to-be-launched area. The intelligent training of the military unmanned aerial vehicle is realized based on simulation training.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle automation, and in particular is a military training system based on artificial intelligence and big data. Background Art

[0002] UAV automated training in military training systems is the process of using UAV simulation systems and related equipment to automatically perform various military training tasks in a virtual environment. This training method uses preset programs and algorithms to enable UAVs to autonomously complete tasks such as takeoff, flight, reconnaissance, and strike, thereby effectively improving training efficiency and safety and reducing training costs and risks. It can help trainers become familiar with UAV operations, improve combat skills, and enhance their ability to cope with complex battlefield environments.

[0003] In the existing technology, drones are often controlled by drone pilots and lack the ability to automatically perform tasks, resulting in the inability of drones to make timely adjustments when encountering obstacles during autonomous flight due to signal loss. In addition, existing load-bearing drones lack actual data to connect with the environment, resulting in weak environmental perception capabilities of drones. To this end, the present invention proposes a military training system based on artificial intelligence and big data. Summary of the invention

[0004] The purpose of the present invention is to propose a military training system based on artificial intelligence and big data to solve the problems raised in the above background technology.

[0005] The technical problems to be solved by the present invention are: How to achieve intelligent training of military drones based on simulation training.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: A military training system based on artificial intelligence and big data, comprising: A database module, used to store the building names of all buildings in the initial area, the road names of all roads and the longitude and latitude of the area center; A data acquisition module is used to collect three-dimensional geographic data of the initial area and send it to the training modeling module; A training modeling module, used for converting the three-dimensional geographic data of the initial area into a three-dimensional model, and sending the three-dimensional model of the initial area to the path planning module; The path planning module is used to plan the flight path of the UAV from the initial coordinates to the initial area, and send the flight path of the UAV model to the simulation training module, obstacle perception module, real training module and target recognition module; The database module is also used to store the allowed flight data of the drone model; The simulation training module is used to perform simulation training on the UAV model, obtain the simulated flight data of the UAV model and send it to the real training module; The obstacle perception module is used to identify and avoid obstacles in the flight path during the flight of the drone, generate path update instructions or obtain obstacle coordinates and send them to the path planning module; The path planning module is used to update the flight path of the drone based on the path update instructions and obstacle coordinates; The data acquisition module is also used to collect the real-time wind speed and real-time temperature of the initial area and send them to the real training module; The real training module is used to adjust the load of the UAV and generate simulation training instructions and send them to the simulation training module; The data acquisition module is also used to collect the first detection target and the second detection target in the area to be delivered and send them to the target recognition module; The target recognition module is used for the drone to automatically identify the target in the area to be delivered, and obtain the material delivery instruction and send it to the material delivery module; The material delivery module is used to deliver materials to the center of the first detection target or the edge of the target delivery area.

[0007] Furthermore, the initial region is a circular region formed by taking the region center of the initial region as the circle point and the first length as the radius; The three-dimensional geographic data of the initial area includes road width, building height, building area and building external structure in the initial area.

[0008] Furthermore, the working process of the training modeling module is as follows: The latitude and longitude of the center of the initial area are taken as the origin O of the three-dimensional rectangular coordinate system, the due east direction is taken as the positive direction of the X axis, the due north direction is taken as the positive direction of the Y axis, and the direction perpendicular to the ground is taken as the positive direction of the Z axis, to establish a three-dimensional rectangular coordinate system OXYZ; Preprocessing the three-dimensional geographic data of the initial area, and then converting the preprocessed three-dimensional geographic data into three-dimensional grid data; The preprocessing steps are as follows: denoising the three-dimensional geographic data of the initial area and filling the blank values ​​in the denoised three-dimensional geographic data; gridding the preprocessed three-dimensional geographic data to obtain the three-dimensional grid data of the initial area; The three-dimensional grid data of the initial area are constructed one by one in the three-dimensional rectangular coordinate system, and then the building names of all buildings in the initial area are matched with the buildings in the three-dimensional rectangular coordinate system, and the road names of all roads are matched with the roads in the three-dimensional rectangular coordinate system to construct a three-dimensional model of the initial area.

[0009] Furthermore, the planning process of the path planning module is as follows: Constructing a UAV model in the three-dimensional model of the initial area, obtaining the initial coordinates of the UAV model in the three-dimensional rectangular coordinate system, and setting the target coordinates of the UAV model; Among them, the initial coordinates of the drone model belong to the initial area; Construct the flight path of the drone model; Set the load of the drone model to zero, and then send flight instructions to the drone model, and the drone model flies according to the flight path; Among them, the flight instruction is to command the drone model to fly from the initial coordinates along the flight path to the target coordinates, and automatically return along the flight path after staying at the target coordinates for a preset time.

[0010] Furthermore, the allowable flight data includes the maximum allowable take-off temperature, the minimum allowable take-off temperature, the standard allowable take-off wind speed, the maximum allowable take-off wind speed and the minimum allowable take-off wind speed of the UAV model. The minimum allowable take-off temperature is used as the left endpoint and the maximum allowable take-off temperature is used as the right endpoint to obtain the standard allowable take-off temperature range of the UAV model. Similarly, the standard allowable take-off wind speed range can be obtained.

[0011] Furthermore, the training process of the simulation training module is as follows: Step S1, setting the load capacity of the drone model to zero, performing a first test on the drone model in the standard allowable takeoff temperature range and when the wind speed is zero, obtaining the remaining power of the drone model when returning home, if the remaining power of the drone model when returning home is less than the power threshold or it cannot return home, then stop the test and determine that the drone model training has failed; if the remaining power of the drone model when returning home is greater than or equal to the power threshold, then increase the fixed load capacity upward until the remaining power of the drone model when returning home is zero, and record the corresponding load capacity as the ideal maximum load capacity; wherein, the middle value of the standard allowable takeoff temperature range is taken for the first test; Step S2, obtaining the maximum allowable take-off temperature, the minimum allowable take-off temperature, the standard allowable take-off wind speed, the maximum allowable take-off wind speed and the minimum allowable take-off wind speed of the drone model, setting the temperature corresponding to the three-dimensional model of the initial area to the minimum allowable take-off temperature, setting the wind speed corresponding to the three-dimensional model of the initial area to the minimum allowable take-off wind speed, setting the load capacity of the drone model to zero, and performing a second test on the drone model; Step S3: If the remaining power of the drone model when returning is less than the power threshold or it cannot return, the test is stopped and the drone model training is determined to have failed.

[0012] Furthermore, the training process of the simulation training module also includes: Step S4, if the remaining power of the drone model when returning is greater than or equal to the power threshold, the temperature and wind speed corresponding to the three-dimensional model of the initial area are set unchanged, and the load of the drone model is increased upward with a fixed load until the ideal maximum load is reached. When the test shows that the remaining power of the drone model when returning is less than the power threshold or it cannot return, the corresponding load is recorded as the maximum load of the drone model; After obtaining the maximum load capacity of the drone at the current temperature and wind speed, the maximum load capacity of the drone model at fixed temperature and different wind speeds is tested with wind speed as the variable and temperature as the quantity. Similarly, taking temperature as the variable and wind speed as the quantity, the maximum load capacity of the drone model at different temperatures and fixed wind speeds was tested; In this way, the maximum load capacity of the UAV model under different wind speeds and temperatures is obtained; Step S5, taking the ideal maximum load capacity of the drone model and the maximum load capacity of the drone model at different wind speeds and different temperatures as simulated flight data of the drone model.

[0013] Furthermore, the working process of the obstacle perception module is as follows: Construct the obstacle perception area of ​​the drone, specifically: The geometric center of the drone is taken as the center of the circle, the maximum sensing length of the sensing sensor in the drone is taken as the radius, and the constructed spherical area is taken as the obstacle sensing area of ​​the drone. Set the drone's load to zero and send a flight command to the drone. When the drone flies along the flight path, the obstacle sensing area issues an alarm, the drone stops flying, and an obstacle model is drawn through the obstacle sensing area. If the obstacle model is an obstacle that the drone cannot pass through, the drone will return immediately and generate a path update instruction; If the obstacle model is an obstacle that the UAV can pass through, the UAV records the obstacle coordinates corresponding to the obstacle model.

[0014] Furthermore, the working process of the real training module is as follows: Get simulated flight data of the drone model and configure the payload for the drone based on the real-time wind speed and temperature in the initial area; A fixed percentage of the load in the simulated flight data of the UAV model is used as the initial load of the UAV, and the remaining power of the UAV after returning along the flight path with the initial load is obtained; If the remaining power of the drone is greater than zero, the simulated flight data of the drone model is determined to be accurate, and training continues; If the remaining power of the drone is zero or it cannot return, it is determined that the simulated flight data of the drone model is inaccurate, and an optimized simulation training instruction is generated; Increase upward with a fixed load, record the actual load corresponding to the time when the remaining power of the UAV is zero when returning, and compare it with the maximum load corresponding to the current wind speed and temperature. If the actual load is greater than or equal to the maximum load, the simulated flight data of the UAV model is determined to be accurate, and the simulated flight data of the UAV model is saved to the database; if the actual load is less than the maximum load, the simulated flight data of the UAV model is determined to be inaccurate, and an optimized simulation training instruction is generated.

[0015] Furthermore, the working process of the target recognition module is as follows: The area to be placed is constructed with the center of the initial area as the dot and the second length as the radius; When the UAV reaches the edge of the area to be dropped along the flight path, target recognition is initiated; When the UAV detects the first detection target through target recognition, the UAV flies to the top of the center of the first detection target and generates a material delivery instruction; The first detection target is a special pattern or a special shaped object on the ground in the area to be dropped. The special pattern or special object is a pattern or object that has been previously input into the image recognition model of the drone for training. The drone can perform image recognition locally. When the UAV fails to detect the first detection target through target recognition, it starts to detect the second detection target. If the second detection target is detected, a target delivery area is constructed with the second detection target as the center and the third length as the radius. At the same time, the UAV obtains the delivery area edge of the target delivery area and flies to just above the edge of the delivery area to generate a material delivery instruction. If the second detection target is not detected, the drone will return immediately; Specifically, the second detection target is a person wearing special colors or an object with special colors in the area to be delivered. The drone captures the special colors in the area to be delivered and compares them with the local preset colors. If the local preset color is the same as the special color, the person with the special color or the object with the special color is determined as the second detection target; If the local preset color is different from the special color, no action is performed; The first length is greater than the second length, the second length is greater than the third length, and the third length is greater than zero.

[0016] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. The present invention first converts the three-dimensional geographic data of the initial area into a three-dimensional model through a training modeling module, and sends the three-dimensional model of the initial area to a path planning module. The path planning module plans the flight path of the UAV from the initial coordinates to the initial area, and sends the flight path of the UAV model to a simulation training module, an obstacle perception module, a real training module and a target recognition module. Then, the UAV model is simulated and trained through the simulation training module to obtain simulated flight data of the UAV model, and the simulated flight data of the UAV model is sent to the real training module. At the same time, the obstacle perception module is used to identify and avoid obstacles in the flight path during the flight of the UAV. During actual training, the load of the UAV is adjusted through the real training module, and a simulation training instruction is generated and sent to the simulation training module. The present invention combines simulation training to realize intelligent training of military UAVs; 2. The present invention also collects the first detection target and the second detection target in the area to be delivered, and automatically identifies the identification target in the area to be delivered to the drone through the target identification module, obtains the material delivery instruction and sends it to the material delivery module, and uses the material delivery module to deliver the materials to the center of the first detection target or the edge of the target delivery area. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0018] Figure 1 is the overall system block diagram of the present invention; Figure 2 A top view of the three-dimensional model corresponding to the initial area in the present invention; Figure 3 This is an example diagram of temperature-maximum load capacity of the drone model in the present invention; Figure 4 This is an example diagram of wind speed-maximum load capacity of the drone model in the present invention; Figure 5 A top view of the obstacle sensing area of ​​the drone of the present invention; Figure 6 This is an example diagram of the first detection target and the second detection target in the area to be deployed. DETAILED DESCRIPTION

[0019] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Example 1: Please refer to Figure 1-Figure 6As shown, the technical solution provided by the present invention is: a military training system based on artificial intelligence and big data, which is mainly used to enable drones to complete tasks strictly according to preset instructions in areas with weak signals, including a data acquisition module, a training modeling module, a path planning module, a simulation training module, an obstacle perception module, a real training module, a target recognition module, a material delivery module and a database module, wherein the database module is connected to the training modeling module, the simulation training module and the target recognition module; The database module is used to store the building names of all buildings in the initial area, the road names of all roads and the longitude and latitude of the area center; It should be specifically stated that the initial region is a circular region formed by taking the region center of the initial region as the circle point and the first length as the radius; The data acquisition module is used to collect three-dimensional geographic data of the initial area and send it to the training modeling module; wherein the three-dimensional geographic data of the initial area includes road width, building height, building area and building external structure in the initial area.

[0021] The training modeling module is used to convert the three-dimensional geographic data of the initial area into a three-dimensional model; like Figure 2 As shown, in this embodiment, the working process of the training modeling module is as follows: The latitude and longitude of the center of the initial area are taken as the origin O of the three-dimensional rectangular coordinate system, the due east direction is taken as the positive direction of the X axis, the due north direction is taken as the positive direction of the Y axis, and the direction perpendicular to the ground is taken as the positive direction of the Z axis, to establish a three-dimensional rectangular coordinate system OXYZ; Preprocessing the three-dimensional geographic data of the initial area, and then converting the preprocessed three-dimensional geographic data into three-dimensional grid data; The preprocessing steps are as follows: denoising the three-dimensional geographic data of the initial area and filling the blank values ​​in the denoised three-dimensional geographic data; gridding the preprocessed three-dimensional geographic data to obtain the three-dimensional grid data of the initial area; The three-dimensional grid data of the initial area are constructed one by one in the three-dimensional rectangular coordinate system, and then the building names of all buildings in the initial area are matched with the buildings in the three-dimensional rectangular coordinate system, and the road names of all roads are matched with the roads in the three-dimensional rectangular coordinate system, so as to construct a three-dimensional model of the initial area; It should be specifically noted that the subsequent simulation training of the UAV was conducted in the three-dimensional model of the initial area, not in the real environment; The training modeling module sends the three-dimensional model of the initial area to the path planning module.

[0022] In this embodiment, the path planning module is used to plan the flight path of the UAV from the initial coordinates to the initial area; In this embodiment, the planning process of the path planning module is as follows: Constructing a UAV model in the three-dimensional model of the initial area, obtaining the initial coordinates of the UAV model in the three-dimensional rectangular coordinate system, and setting the target coordinates of the UAV model; Among them, the initial coordinates of the drone model belong to the initial area; In a specific implementation, the target coordinates of the drone model may be the coordinates of any point in the initial area. In this embodiment, the target coordinates of the drone model are the coordinates corresponding to the center of the initial area. Constructing a flight path of the UAV model; wherein, constructing the flight path of the UAV model is an existing technology, and usually uses an A-Star algorithm, a fast random tree algorithm, or an ant colony algorithm, etc.; Set the load of the drone model to zero, and then send flight instructions to the drone model, and the drone model flies according to the flight path; Among them, the flight instruction is to command the drone model to fly from the initial coordinates along the flight path to the target coordinates, and automatically return along the flight path after staying at the target coordinates for a preset time.

[0023] The path planning module sends the flight path of the drone model to the simulation training module, the obstacle perception module, the real training module and the target recognition module.

[0024] The database module is also used to store the allowed flight data of the drone model, the allowed flight data is the maximum allowed take-off temperature, the minimum allowed take-off temperature, the standard allowed take-off wind speed, the maximum allowed take-off wind speed and the minimum allowed take-off wind speed of the drone model, with the minimum allowed take-off temperature as the left endpoint and the maximum allowed take-off temperature as the right endpoint, so as to obtain the standard allowed take-off temperature range of the drone model, and similarly, the standard allowed take-off wind speed range can be obtained; The simulation training module is used to perform simulation training on the UAV model to obtain simulated flight data of the UAV model; like Figure 2-Figure 4 As shown, in this embodiment, the training process of the simulation training module is as follows: Step S1, setting the load capacity of the drone model to zero, performing a first test on the drone model in the standard allowable takeoff temperature range and when the wind speed is zero, obtaining the remaining power of the drone model when returning home, if the remaining power of the drone model when returning home is less than the power threshold or it cannot return home, then stop the test and determine that the drone model training has failed; if the remaining power of the drone model when returning home is greater than or equal to the power threshold, then increase the fixed load capacity upward until the remaining power of the drone model when returning home is zero, and record the corresponding load capacity as the ideal maximum load capacity; wherein, the middle value of the standard allowable takeoff temperature range is taken for the first test; Step S2, obtaining the maximum allowable take-off temperature, the minimum allowable take-off temperature, the standard allowable take-off wind speed, the maximum allowable take-off wind speed and the minimum allowable take-off wind speed of the drone model, setting the temperature corresponding to the three-dimensional model of the initial area to the minimum allowable take-off temperature, setting the wind speed corresponding to the three-dimensional model of the initial area to the minimum allowable take-off wind speed, setting the load capacity of the drone model to zero, and performing a second test on the drone model; Step S3, if the remaining power of the drone model when returning is less than the power threshold or it cannot return, the test is stopped and the drone model training is determined to have failed; Step S4, if the remaining power of the drone model when returning is greater than or equal to the power threshold, the temperature and wind speed corresponding to the three-dimensional model of the initial area are set unchanged, and the load of the drone model is increased upward with a fixed load until the ideal maximum load is reached. When the test shows that the remaining power of the drone model when returning is less than the power threshold or it cannot return, the corresponding load is recorded as the maximum load of the drone model; After obtaining the maximum load capacity of the drone at the current temperature and wind speed, the maximum load capacity of the drone model at fixed temperature and different wind speeds is tested with wind speed as the variable and temperature as the quantity. Similarly, taking temperature as the variable and wind speed as the quantity, the maximum load capacity of the drone model at different temperatures and fixed wind speeds was tested; In this way, the maximum load capacity of the UAV model under different wind speeds and temperatures is obtained; Step S5, taking the ideal maximum load capacity of the drone model and the maximum load capacity of the drone model at different wind speeds and different temperatures as simulated flight data of the drone model; The simulation training module sends the simulated flight data of the UAV model to the real training module.

[0025] In this embodiment, the obstacle sensing module is used to identify and avoid obstacles in the flight path during the flight of the drone; like Figure 5As shown, in this embodiment, the drone is equipped with a perception sensor, which includes but is not limited to a laser radar and an infrared sensor, etc. The working process of the obstacle perception module is as follows: Construct obstacle perception area for drones; Specifically, the geometric center of the drone is used as the center of the circle, the maximum sensing length of the sensing sensor in the drone is used as the radius, and the constructed spherical area is used as the obstacle sensing area of ​​the drone; Set the drone's load to zero and send a flight command to the drone. When the drone flies along the flight path, the obstacle sensing area issues an alarm, the drone stops flying, and an obstacle model is drawn through the obstacle sensing area. If the obstacle model is an obstacle that the drone cannot pass through, the drone will return immediately and generate a path update instruction; If the obstacle model is an obstacle that the drone can pass through, the drone records the obstacle coordinates corresponding to the obstacle model; The obstacle perception module sends the path update instruction or the obstacle coordinates to the path planning module, and the path planning module updates the flight path of the UAV according to the path update instruction and the obstacle coordinates.

[0026] Furthermore, the data acquisition module is also used to collect the real-time wind speed and real-time temperature of the initial area and send them to the real training module; The real training module is used to adjust the load of the UAV; Specifically, simulated flight data of the UAV model is obtained, and a load capacity is configured for the UAV based on the real-time wind speed and real-time temperature of the initial area; Furthermore, a fixed percentage of the load in the simulated flight data of the UAV model is used as the initial load of the UAV, and the remaining power of the UAV after returning along the flight path with the initial load is obtained; If the remaining power of the drone is greater than zero, the simulated flight data of the drone model is determined to be accurate, and training continues; If the remaining power of the drone is zero or it cannot return, it is determined that the simulated flight data of the drone model is inaccurate, and an optimized simulation training instruction is generated; The fixed load is increased upward, and the actual load corresponding to the remaining power of the drone when returning is zero is recorded, and compared with the maximum load corresponding to the current wind speed and temperature. If the actual load is greater than or equal to the maximum load, the simulated flight data of the drone model is determined to be accurate, and the simulated flight data of the drone model is saved to the database; if the actual load is less than the maximum load, the simulated flight data of the drone model is determined to be inaccurate, and an optimized simulation training instruction is generated; The real training module sends the simulation training instruction to the simulation training module.

[0027] In this embodiment, the data acquisition module is also used to collect the first detection target and the second detection target in the area to be delivered, and send them to the target recognition module; The target recognition module is used for the drone to automatically recognize the target in the area to be dropped; In this embodiment, the drone is equipped with a high-definition camera, and the working process of the target recognition module is as follows: The area to be placed is constructed with the center of the initial area as the dot and the second length as the radius; When the UAV reaches the edge of the area to be dropped along the flight path, target recognition is initiated; like Figure 6 As shown, when the UAV detects the first detection target through target recognition, the UAV flies to the top of the center of the first detection target and generates a material delivery instruction; The first detection target is a special pattern or a special shaped object on the ground in the area to be dropped. The special pattern or special object is a pattern or object that has been previously input into the image recognition model of the drone for training. The drone can perform image recognition locally. When the UAV fails to detect the first detection target through target recognition, it starts to detect the second detection target. If the second detection target is detected, a target delivery area is constructed with the second detection target as the center and the third length as the radius. At the same time, the UAV obtains the delivery area edge of the target delivery area and flies to just above the edge of the delivery area to generate a material delivery instruction. If the second detection target is not detected, the drone will return immediately; Specifically, the second detection target is a person wearing special colors or an object with special colors in the area to be delivered. The drone captures the special colors in the area to be delivered and compares them with the local preset colors. If the local preset color is the same as the special color, the person with the special color or the object with the special color is determined as the second detection target; If the local preset color is different from the special color, no action is performed; Wherein, the first length is greater than the second length, the second length is greater than the third length, and the third length is greater than zero; The target recognition module sends the material delivery instruction to the material delivery module.

[0028] The material delivery module is used to deliver materials to the center of the first detection target or the edge of the target delivery area.

[0029] In this application, if corresponding calculation formulas appear, the above calculation formulas are all dimensionless and take their numerical calculations. The weight coefficients, proportional coefficients and other coefficients in the formulas are set to a result value obtained by quantifying each parameter. The size of the weight coefficient and the proportional coefficient can be determined as long as it does not affect the proportional relationship between the parameter and the result value.

[0030] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A military training system based on artificial intelligence and big data, characterized in that: include: A database module, used to store the building names of all buildings in the initial area, the road names of all roads and the longitude and latitude of the area center; A data acquisition module, used to acquire three-dimensional geographic data of the initial area; A training modeling module, used for converting the three-dimensional geographic data of the initial area into a three-dimensional model; The path planning module is used to plan the flight path of the UAV from the initial coordinates to the initial area and the flight path of the UAV model; The database module is also used to store the allowed flight data of the drone model; A simulation training module is used to perform simulation training on the UAV model and obtain simulated flight data of the UAV model; The obstacle perception module is used to identify and avoid obstacles in the flight path during the flight of the drone, generate path update instructions or obtain obstacle coordinates; The path planning module is used to update the flight path of the drone based on the path update instructions and obstacle coordinates; The data acquisition module is also used to collect the real-time wind speed and real-time temperature of the initial area; The real training module is used to adjust the load of the UAV and generate simulation training instructions; The data collection module is also used to collect the first detection target and the second detection target in the area to be released; The target recognition module is used for the drone to automatically identify the target in the area to be delivered and obtain the material delivery instruction; The material delivery module is used to deliver materials to the center of the first detection target or the edge of the target delivery area.

2. The military training system based on artificial intelligence and big data according to claim 1 is characterized in that: The initial area is a circular area formed by taking the area center of the initial area as the circle point and the first length as the radius; The three-dimensional geographic data of the initial area includes road width, building height, building area and building external structure in the initial area.

3. A military training system based on artificial intelligence and big data according to claim 2, characterized in that: The working process of the training modeling module is as follows: The latitude and longitude of the center of the initial area are taken as the origin O of the three-dimensional rectangular coordinate system, the due east direction is taken as the positive direction of the X axis, the due north direction is taken as the positive direction of the Y axis, and the direction perpendicular to the ground is taken as the positive direction of the Z axis, to establish a three-dimensional rectangular coordinate system OXYZ; Preprocessing the three-dimensional geographic data of the initial area, and then converting the preprocessed three-dimensional geographic data into three-dimensional grid data; The preprocessing steps are as follows: denoising the three-dimensional geographic data of the initial area and filling the blank values ​​in the denoised three-dimensional geographic data; gridding the preprocessed three-dimensional geographic data to obtain the three-dimensional grid data of the initial area; The three-dimensional grid data of the initial area are constructed one by one in the three-dimensional rectangular coordinate system, and then the building names of all buildings in the initial area are matched with the buildings in the three-dimensional rectangular coordinate system, and the road names of all roads are matched with the roads in the three-dimensional rectangular coordinate system to construct a three-dimensional model of the initial area.

4. The military training system based on artificial intelligence and big data according to claim 3 is characterized in that: The planning process of the path planning module is as follows: Constructing a UAV model in the three-dimensional model of the initial area, obtaining the initial coordinates of the UAV model in the three-dimensional rectangular coordinate system, and setting the target coordinates of the UAV model; Among them, the initial coordinates of the drone model belong to the initial area; Construct the flight path of the drone model; Set the load of the drone model to zero, and then send flight instructions to the drone model, and the drone model flies according to the flight path; Among them, the flight instruction is to command the drone model to fly from the initial coordinates along the flight path to the target coordinates, and automatically return along the flight path after staying at the target coordinates for a preset time.

5. The military training system based on artificial intelligence and big data according to claim 1 is characterized in that: The allowable flight data includes the maximum allowable take-off temperature, minimum allowable take-off temperature, standard allowable take-off wind speed, maximum allowable take-off wind speed and minimum allowable take-off wind speed of the UAV model. The minimum allowable take-off temperature is used as the left endpoint and the maximum allowable take-off temperature is used as the right endpoint to obtain the standard allowable take-off temperature range of the UAV model. Similarly, the standard allowable take-off wind speed range can be obtained.

6. The military training system based on artificial intelligence and big data according to claim 5 is characterized in that: The training process of the simulation training module is as follows: Step S1, setting the load capacity of the drone model to zero, performing a first test on the drone model in the standard allowable takeoff temperature range and when the wind speed is zero, obtaining the remaining power of the drone model when returning home, if the remaining power of the drone model when returning home is less than the power threshold or it cannot return home, then stop the test and determine that the drone model training has failed; if the remaining power of the drone model when returning home is greater than or equal to the power threshold, then increase the fixed load capacity upward until the remaining power of the drone model when returning home is zero, and record the corresponding load capacity as the ideal maximum load capacity; wherein, the middle value of the standard allowable takeoff temperature range is taken for the first test; Step S2, obtaining the maximum allowable take-off temperature, the minimum allowable take-off temperature, the standard allowable take-off wind speed, the maximum allowable take-off wind speed and the minimum allowable take-off wind speed of the drone model, setting the temperature corresponding to the three-dimensional model of the initial area to the minimum allowable take-off temperature, setting the wind speed corresponding to the three-dimensional model of the initial area to the minimum allowable take-off wind speed, setting the load capacity of the drone model to zero, and performing a second test on the drone model; Step S3: If the remaining power of the drone model when returning is less than the power threshold or it cannot return, the test is stopped and the drone model training is determined to have failed.

7. The military training system based on artificial intelligence and big data according to claim 6 is characterized in that: The training process of the simulation training module also includes: Step S4, if the remaining power of the drone model when returning is greater than or equal to the power threshold, the temperature and wind speed corresponding to the three-dimensional model of the initial area are set unchanged, and the load of the drone model is increased upward with a fixed load until the ideal maximum load is reached. When the test shows that the remaining power of the drone model when returning is less than the power threshold or it cannot return, the corresponding load is recorded as the maximum load of the drone model; After obtaining the maximum load capacity of the drone at the current temperature and wind speed, the maximum load capacity of the drone model at fixed temperature and different wind speeds is tested with wind speed as the variable and temperature as the quantity. Similarly, taking temperature as the variable and wind speed as the quantity, the maximum load capacity of the drone model at different temperatures and fixed wind speeds was tested; In this way, the maximum load capacity of the UAV model under different wind speeds and temperatures is obtained; Step S5, taking the ideal maximum load capacity of the drone model and the maximum load capacity of the drone model at different wind speeds and different temperatures as simulated flight data of the drone model.

8. The military training system based on artificial intelligence and big data according to claim 1 is characterized in that: The working process of the obstacle perception module is as follows: Construct the obstacle perception area of ​​the drone, specifically: The geometric center of the drone is taken as the center of the circle, the maximum sensing length of the sensing sensor in the drone is taken as the radius, and the constructed spherical area is taken as the obstacle sensing area of ​​the drone. Set the drone's load to zero and send a flight command to the drone. When the drone flies along the flight path, the obstacle sensing area issues an alarm, the drone stops flying, and an obstacle model is drawn through the obstacle sensing area. If the obstacle model is an obstacle that the drone cannot pass through, the drone will return immediately and generate a path update instruction; If the obstacle model is an obstacle that the UAV can pass through, the UAV records the obstacle coordinates corresponding to the obstacle model.

9. The military training system based on artificial intelligence and big data according to claim 7 is characterized in that: The working process of the real training module is: Get simulated flight data of the drone model and configure the payload for the drone based on the real-time wind speed and temperature in the initial area; A fixed percentage of the load in the simulated flight data of the UAV model is used as the initial load of the UAV, and the remaining power of the UAV after returning along the flight path with the initial load is obtained; If the remaining power of the drone is greater than zero, the simulated flight data of the drone model is determined to be accurate, and training continues; If the remaining power of the drone is zero or it cannot return, it is determined that the simulated flight data of the drone model is inaccurate, and an optimized simulation training instruction is generated; Increase upward with a fixed load, record the actual load corresponding to the time when the remaining power of the UAV is zero when returning, and compare it with the maximum load corresponding to the current wind speed and temperature. If the actual load is greater than or equal to the maximum load, the simulated flight data of the UAV model is determined to be accurate, and the simulated flight data of the UAV model is saved to the database; if the actual load is less than the maximum load, the simulated flight data of the UAV model is determined to be inaccurate, and an optimized simulation training instruction is generated.

10. The military training system based on artificial intelligence and big data according to claim 1, characterized in that: The working process of the target recognition module is as follows: The area to be placed is constructed with the center of the initial area as the dot and the second length as the radius; When the UAV reaches the edge of the area to be dropped along the flight path, target recognition is initiated; When the UAV detects the first detection target through target recognition, the UAV flies to the top of the center of the first detection target and generates a material delivery instruction; The first detection target is a special pattern or a special shaped object on the ground in the area to be dropped. The special pattern or special object is a pattern or object that has been previously input into the image recognition model of the drone for training. The drone can perform image recognition locally. When the UAV fails to detect the first detection target through target recognition, it starts to detect the second detection target. If the second detection target is detected, a target delivery area is constructed with the second detection target as the center and the third length as the radius. At the same time, the UAV obtains the delivery area edge of the target delivery area and flies to just above the edge of the delivery area to generate a material delivery instruction. If the second detection target is not detected, the drone will return immediately; Specifically, the second detection target is a person wearing special colors or an object with special colors in the area to be delivered. The drone captures the special colors in the area to be delivered and compares them with the local preset colors. If the local preset color is the same as the special color, the person with the special color or the object with the special color is determined as the second detection target; If the local preset color is different from the special color, no action is performed; The first length is greater than the second length, the second length is greater than the third length, and the third length is greater than zero.

Images