Full-automatic inspection and printing integrated control system based on Raspberry Pi and APM secondary development

Abstract

The invention discloses a full-automatic reconnaissance and attack integrated control system based on Raspberry Pi and APM secondary development, which provides regional position information according to a ground station, automatically carries out route planning, automatically identifies and searches a launching target, carries out target positioning, intelligently decides a system terminal, and automatically completes a precise launching task. The position and speed parameters of the unmanned aerial vehicle, the real-time pose of the launched target and the surrounding environment information can be displayed on the ground station in real time during the period. The system can automatically complete a reconnaissance and attack integrated task, and has the characteristics of modular and detachable design, low cost and complete functions, so that accurate launching of the fixed-wing unmanned aerial vehicle and rapid deployment of the unmanned aerial vehicle system for target recognition are realized.

Description

Technical Field

[0001] The present invention relates to the field of fixed-wing UAV identification and delivery technology, and in particular to a fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM. Background Art

[0002] With the continuous development of general aviation transportation business and the gradual increase in battlefield demand, fixed-wing UAV integrated automatic control systems have been given great attention. Many countries are competing to develop fixed-wing UAV systems, among which drones used for military reconnaissance and strike and civilian material transportation account for a large proportion. The diversification of application scenarios and the universality of different types of UAVs have put forward new requirements for the development of such systems for fixed-wing UAVs.

[0003] Existing technologies still face obstacles in rapidly targeting and delivering drones, and the drone market is lacking. Existing solutions fall into three main categories: large reconnaissance and strike drones for regional strikes, rotary-wing bomb drones for targeted delivery, and cruise missile strikes. Large reconnaissance and strike drones have large radar reflectors, making them vulnerable to being shot down, difficult to maintain, and lacking airport runway resources, making them unsuitable for civilian use. Rotary-wing bomb drones have slow flight speeds, are hampered by high altitudes, and have limited system integrity, resulting in slow communication with support. Cruise missiles are expensive to develop and produce, resulting in low efficiency for single missions. Furthermore, due to cost and other limitations, a single reconnaissance and strike drone can only perform a single function, failing to meet the needs of diverse sectors. Cross-platform adaptation is difficult, hindering the expansion of diverse application scenarios. Therefore, developing a multimodal drone identification and delivery system that enables system control and intelligent decision-making for fixed-wing drones, enabling identification and delivery of target areas or objects, has become a critical technical challenge. Summary of the Invention

[0004] In view of the above shortcomings, the present invention proposes a fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM.

[0005] The Raspberry Pi is a microcomputer, and the APM is a type of open-source flight control firmware. To achieve the above objectives, the present invention provides the following solution: a fully automatic reconnaissance and strike control system based on the Raspberry Pi and APM, suitable for fixed-wing unmanned aerial vehicles (UAVs). The system comprises four components: an identification and positioning system, an autonomous route planning system, a precision delivery system, and a joint communications system. The system is characterized by a comprehensive hardware infrastructure that can be deployed on a fixed-wing aircraft with an appropriate payload, transforming the aircraft into a reconnaissance and strike UAV capable of performing reconnaissance and identification, target tracking, and precision strikes.

[0006] The identification and positioning system includes a visual recognition module and a target positioning module. The ROS system framework is deployed on the host computer, a Raspberry Pi. After image preprocessing using OpenCV, the recognition algorithm analyzes the image information sent back by the camera, outputs the relative position of the target and the drone, and summarizes it into the positioning function package established in the ROS framework. The ROS framework then uses the longitude and latitude conversion algorithm stored in the node to obtain the target's precise longitude and latitude in the Earth coordinate system.

[0007] Furthermore, the visual recognition module employs OpenCV image preprocessing, an open-source computer vision learning software library. Based on data enhancement for motion image deblurring, it uses image preprocessing and image enhancement techniques to process the segmented image. The recognition algorithm detects the target prediction box in the image, obtains the minimum rectangular box containing the prediction box, and obtains the coordinates of the four corners of the minimum rectangular box. The rectangle containing the prediction box is then segmented to accurately segment the target image. This reduces noise and interference, enhances image contrast and edge features, and thus improves image recognition accuracy, ultimately identifying the intended target.

[0008] Furthermore, the recognition algorithm utilizes an improved YOLOv5 algorithm, embeds a spatial attention mechanism module in the backbone network, and adds a small target detection head to improve recognition speed and image recognition accuracy. For aerial drone images, which contain a lot of complex and disordered background information and few target features, the SimAM module is introduced to eliminate interference and focus on extracting the key features of small targets, improving detection speed and rapidly locating the target position in the image. Each grid cell can predict multiple targets of different categories and output their center coordinates, width, height, and confidence level.

[0009] Furthermore, the target positioning module uses the center coordinates, width, height and confidence of the output target as parameters, and performs calculations on the real-time longitude and latitude information collected by the lower flight control system through MAVROS. The node uses Python language to call the geographic coordinate system conversion library (pyproj) to output the precise longitude and latitude of the target.

[0010] Preferably, for specific targets in photos taken during the fixed-wing UAV's constant altitude cruise, the camera's internal and external parameter data used in the conversion are the camera's factory execution standards to ensure the accuracy of the conversion.

[0011] The autonomous route planning system is divided into APM secondary development route planning firmware and ROS route intelligent decision-making package. The software is based on APM firmware and ROS system framework, and the hardware is based on open source flight control and RTK navigation modules.

[0012] Furthermore, the flight controller uses APM's redeveloped route planning firmware to read the airspeed sensor and IMU's attitude information, obtain the longitude and latitude for differential positioning navigation, and control the coordinated operation of various modules to achieve drone attitude control and perform fusion filtering on the sensor data. The sensor data is shared with the Raspberry Pi host computer via a USB data cable. The flight controller also receives waypoint files from the host computer and outputs actuator control signals using the attitude control algorithm.

[0013] Furthermore, the aforementioned sensors include a gyroscope, accelerometer, magnetometer, barometer, GPS, and airspeed meter. The accelerometer, magnetometer, gyroscope, and barometer are integrated into the flight control system. Due to the limited accuracy of the accelerometer, magnetometer, gyroscope, and barometer, data from the RTK navigation module is fused with these data to compensate for inertial navigation errors and obtain precise position information for the fixed-wing drone. Preferably, differential positioning navigation based on network RTK technology is used to achieve decimeter-level positioning of the drone, and an airspeed meter is used to replace the barometer to obtain precise airspeed.

[0014] When the ground station sends information about the target area for the drone's launch, the flight controller receives the instructions, performs path planning, and sends the location information to the Raspberry Pi host computer via the USB serial port. The Raspberry Pi host computer runs the ROS route intelligent decision package to parse and select a route, and then transmits it to the flight controller. The flight controller then sends uORB-type information to the servo motors and servo actuators via the SBus protocol to control the drone's heading.

[0015] Furthermore, the above-mentioned ROS route intelligent decision package is deployed in the Raspberry Pi host computer. After obtaining the specific latitude and longitude position of the ground object, the Raspberry Pi host computer controller independently decides the path passing over the object. The path template comes from the route file package manually planned before takeoff. According to the L1 navigation and guidance algorithm, the expected path and the longitude and latitude of the impact point are calculated, and the algorithm intelligent decision is made to obtain the best route. The new waypoint file is transmitted to the underlying flight control for executing the new path.

[0016] The precision delivery system is deployed in the lower-level flight control system, using the NUTTX operating system and a redevelopment of the open-source flight firmware APM. The algorithm redeveloped the commands_logic command in the APM open-source code to adapt it for airdrop scenarios. Upon receiving the airdrop command, its primary function is to calculate the distance the aircraft needs to fly to accurately deliver the object to the target, wait for a period of time to compensate for mechanical errors, and then trigger the airdrop. Upon successful target delivery, a message is transmitted back to the ground station for display.

[0017] Furthermore, the entire launch process was simulated using MATLAB software, recreating the entire process using physical formulas. The basic horizontal projection motion was recreated using a programming language, and the drag coefficient of the simulated projectile at different volumes and masses was calculated, taking into account the drag in normal air. The launch process was simulated multiple times, and then fitted and corrected.

[0018] The joint communication system consists of a flight control communication link, a reconnaissance communication link and a flight control and airborne computer joint communication link.

[0019] Furthermore, the flight control link system is an important component of the UAV system. Its main task is to establish an air-ground two-way data transmission channel for completing the long-distance remote control, reconnaissance and strike mission information transmission of the UAV by the ground control station.

[0020] Furthermore, the reconnaissance communication link transmits target image information acquired by the upper-level control system to the ground-side PC via a redeveloped cloud-based OSS, ensuring accurate synchronization of reconnaissance data with the ground station and providing effective data for the deployment process. An industrial-grade 4G module provides network services for the upper-level control system, enabling the redeveloped Alibaba Cloud OSS transmission to achieve cloud-based transmission without distance restrictions and frequency band conflicts.

[0021] Furthermore, the aforementioned joint communication link between the flight controller and the onboard computer enables real-time, two-way communication via the ROS system framework. MAVROS is used to convert ROS topics into MAVLINK-formatted messages and send them to the flight controller. Within the flight controller, the MAVLINK module converts these MAVLINK-formatted messages into uORB messages, enabling messaging between the various modules of the drone.

[0022] Preferably, a subscriber is added to the flight mode topic in ROS, and a callback function is added to the subscriber to realize automatic scheduling of tasks. When the flight mode is switched to task mode, the callback function calls the task list pull service subscription topic and uploads the new task file.

[0023] Compared with the existing technology, the beneficial effects of the present invention are:

[0024] It provides a fully automatic reconnaissance and strike integrated UAV system, which combines patrol reconnaissance and image recognition tasks in real time and synchronizes them. It can strike ground targets and deliver supplies at high altitude and high speed, realizing fully automatic reconnaissance and strike integration.

[0025] Normal use in high-intensity and harsh environments: The system adopts a highly stable hardware system and a software technology solution with high fault tolerance, and can be used normally in high-intensity and harsh environments. It adopts an independently designed system structure and constructed mathematical model to eliminate scene interference factors, improve the drone's recognition and operation capabilities and work efficiency under high-speed flight conditions, and ensure flight safety and stability while achieving the established mission objectives.

[0026] The device itself is universal and suitable for installation on most fixed-wing drones. The device itself is light and simple, the installation process is simple and easy to operate, and it can be deployed quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a schematic diagram of the subsystem composition of a fully automatic reconnaissance and strike control system based on the secondary development of Raspberry Pi and APM in the present invention;

[0028] Figure 2 This is a hardware structure diagram of a fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM of the present invention;

[0029] Figure 3 This is a schematic diagram of the workflow of the subsystem identification and positioning system of the fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM of the present invention;

[0030] Figure 4 This is a schematic diagram of the workflow of the precision delivery system subsystem of the fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM of the present invention;

[0031] Figure 5 This is a subsystem joint communication system diagram of a fully automatic reconnaissance and strike integrated control system based on the secondary development of Raspberry Pi and APM of the present invention;

[0032] Figure 6 This is a supplementary diagram illustrating the subsystem joint communication system of the fully automatic reconnaissance and strike integrated control system based on the secondary development of Raspberry Pi and APM of the present invention;

[0033] Figure 7 This is a schematic diagram of the connection and working logic of the components of a fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM of the present invention;

[0034] Figure 2 In: 1-Open source hardware autopilot (supports APM firmware); 2-Raspberry Pi; 3-Industrial-grade 4G module; 4-UAV GPS (supports RTK technology); 5-UAV data transmission module; 6-Airspeed meter; 7-Servo; 8-Camera; DETAILED DESCRIPTION

[0035] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0036] See also Figure 2 , the present invention provides a technical solution: The following is the connection principle of the hardware part:

[0037] A fully automatic reconnaissance and strike integrated control system based on the secondary development of Raspberry Pi and APM is suitable for fixed-wing UAVs. The hardware components include: 1- open source hardware autopilot (supporting APM firmware); 2- Raspberry Pi; 3- industrial-grade 4G module; 4- GPS for UAV (supporting RTK technology); 5- UAV data transmission module; 6- airspeed meter; 7- servo; 8- camera; It is characterized in that: all the hardware are assembled in a box that can accommodate Raspberry Pi inside, the open source hardware autopilot-1 is fixed on the top of the box to ensure the normal use of firmware such as the level, and the UAV GPS-4 is connected to the open source hardware through an interface. The open-source hardware autopilot-1 is fixed on top of it to facilitate the normal use of high-precision RTK. The drone data transmission module-5 and airspeed meter-6 are both connected to the open-source hardware autopilot-1 through specific interfaces and arranged in an open area on the box to avoid obstruction and stacking to ensure normal use. The Raspberry Pi-2 is placed inside the box, and the industrial-grade 4G module-3 is connected to the external port of the Raspberry Pi-2 to provide network services for the entire system. The camera-8 with its own gimbal is fixed at the bottom of the box to acquire images from below in real time and transmit the data back to the Raspberry Pi-2 for processing via the serial port protocol. Furthermore, the Raspberry Pi-2 is connected to the open-source hardware autopilot-1 via a USB Type-C cable for data transmission.

[0038] The following is a detailed description of the working method of the fully automatic detection and strike control system based on the secondary development of Raspberry Pi and APM.

[0039] The present invention is a fully automatic reconnaissance and strike integrated control system based on the secondary development of Raspberry Pi and APM. The specific process is that after the ground station sends the target delivery area, the host computer Raspberry Pi receives the instruction, connects the target area to the template route, automatically searches for the target after reaching the target area, identifies the delivery target and returns the target location, and the Raspberry Pi host computer converts the coordinate information in the instruction into longitude and latitude, automatically determines the route path that can pass over the object again, and can automatically find the optimal route, and send the route path to the lower computer flight control through the serial port USB. The lower computer parses the received instruction and sends the instruction to the drone power equipment through the uORB message to realize automatic driving. At the same time, the precise delivery system and the joint communication system run in real time, cooperating with the Raspberry Pi host computer and the lower computer flight control to perform real-time positioning and delivery calculations on the target. After the identification and positioning system and the autonomous route planning system output the target information, the target object can be correctly identified, and the delivery system based on the secondary development of APM can complete precise delivery.

[0040] like Figure 3 The recognition and positioning system adopts the improved yolov5 algorithm, embeds a spatial attention mechanism module in the backbone network (Backbone) and adds a small target detection head to improve the recognition speed and image recognition accuracy. Subsequently, it combines OpenCV image segmentation, uses data enhancement methods based on motion image deblurring, uses image preprocessing and image enhancement technology to process the segmented image, reduce noise and interference, enhance image contrast and edge features, and thus improve image recognition accuracy. The main framework of the entire target autonomous recognition and reconnaissance system is referenced Figure 1 .

[0041] In this example, the target prediction frame in the image is detected by the recognition algorithm, the minimum rectangular frame containing the prediction frame in the image transmitted back by the onboard computer is obtained, and the four vertex coordinates of the minimum rectangular frame are obtained. After the coordinates are determined, the rectangle containing the prediction frame is segmented to maximize the accuracy of the target image segmentation. When YOLOv5 detects the segmented image, the image orientation obtained will be inconsistent due to the arbitrary placement of the ground target, which will interfere with the subsequent recognition accuracy. Therefore, this embodiment pre-processes the segmented image and connects the center of gravity of the different color blocks on the target to create a unit vector. In order to ensure that the final rotation direction of all images is consistent, the unit vector normalization is performed. That is, the segmented image is rotated with the default pixel point in the upper left corner of the image as the rotation center, and the image is rotated so that the unit vector connecting the center of gravity coordinates is rotated to be parallel to the (0, -1) vector. The error angle range is controlled to (-15°, 15°) to ensure the applicability of the rotated image.

[0042] The Raspberry Pi then converts the image to HSV color space. Based on the defined color ranges, a bitwise AND operation is performed on the white and red (or blue) areas of the original image. After conversion to grayscale and binarization, the largest contours of each color are found and extracted. Based on this, the center of gravity of the color blocks is calculated using the center of gravity formula. Based on the coordinate formula for the center of gravity of the color blocks, the coordinates of the centers of gravity of the red (or blue) and white blocks are calculated, and the unit vector from the center of gravity of the white block to the center of gravity of the red (or blue) block is calculated. A ROS framework is built on the Raspberry Pi host computer, and a positioning function package is established within ROS. For the coordinate vectors of specific color blocks in photos taken by the fixed-wing drone during altitude control, the upper-level control system obtains real-time latitude and longitude information and maps them to the photos. This is then implemented using the geographic coordinate system conversion library (pyproj), using the output target's center coordinates, width, height, and confidence level as parameters. To ensure accuracy, the camera's internal and external parameters used in the conversion are factory-standard.

[0043] like Figure 4 The precise delivery can achieve targeted delivery, which is secondary developed based on the commands_logic command in the APM open source code to be suitable for airdrop scenarios.

[0044] Furthermore, define the variable servo_delay to obtain the aircraft's attitude and state. Taking into account the drag of the target in normal air, this example uses 350ml of mineral water as the drop object and calculates the air drag coefficient based on the volume and mass of the bottled water. And considering that the reconnaissance and strike UAV needs to strike ground targets at high altitude and high speed, the speed of the simulated missile is approximated to the speed of the aircraft at the moment of strike, and the wind direction and wind speed are approximately calculated by calculating the difference between the ground speed and the airspeed. The drag coefficient and wind speed are brought into the calculation to calculate the acceleration of the simulated missile in the horizontal and vertical directions. The flight control will calculate the time s and distance for the simulated missile to reach the target in the current environment in real time. When the calculated distance is between 5-30cm in the longitude and latitude of the center of the target, the flight control determines to strike. Considering the delay problem of flight control calculation, a final delay time is reserved in the secondary development firmware, and a timer is used to execute commands in milliseconds to reduce errors. When s is less than 0 and s is greater than 8ms, the distance value is reset and returned. When s is between 0 and 8ms, the final delay time is calculated, and the distance value is reset and returned after waiting, triggering the airdrop action and sending a successful delivery message.

[0045] like Figure 5 、 Figure 6The joint communication system is primarily comprised of a sky-side drone, a ground-side PC, and the cloud. The system consists of a flight control communication link, a reconnaissance communication link, and a joint flight control and onboard computer communication link. A bidirectional air-to-ground data transmission channel is established for remote control of drones from the ground control station, as well as for transmission of reconnaissance and strike mission information. Reconnaissance enables monitoring of drone status, while strike missions allow remote control of drones and mission equipment. Target information acquired by onboard mission sensors is transmitted to the ground-side PC via a redeveloped cloud-based OSS.

[0046] Furthermore, the reconnaissance communication link uses a simple and lightweight industrial-grade 4G module to build a secondary developed cloud-based OSS transmission, achieving cloud-based transmission without distance restrictions and frequency band conflict interference. When the fixed-wing drone arrives above the target area, the ROS package calls the node, enables the transmission Python script, and uploads the target data of the recognition function package. The target data is stored as an object in the storage space (Bucket) of Alibaba Cloud OSS in the form of target latitude and longitude + corresponding target image. The ground-side PC specifies a local folder through the Python script and synchronizes the target data from the cloud to the local storage in real time.

[0047] Furthermore, the communication link between the flight control and the onboard computer is implemented using the ROS framework. Specifically, a ROS node runs on the airborne end, and ROS adds a subscriber to the flight mode topic. This subscriber has a callback function that, when the flight mode switches to mission mode, calls the task list pull service to subscribe to the topic and upload the new mission file. The onboard computer connects to the flight control's serial port via a USB-to-serial module. This connection is primarily used by the onboard computer to obtain flight control information and send control commands, using the MAVLINK transmission protocol. The onboard computer uses MAVROS to convert topics into MAVLINK-formatted messages and sends them to the flight control. The MAVLINK module in the flight control converts these MAVLINK-formatted messages into uORB messages for use in messaging between the drone's modules. Conversely, uORB messages are sent back to the ROS topic, enabling real-time two-way communication.

[0048] It should be noted that, in this application, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0049] The foregoing is merely a list of specific embodiments of the present application, intended to enable those skilled in the art to understand or implement the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the broadest scope consistent with the principles and novel features of the present application.

Claims

1. A fully automatic detection and strike control system based on the secondary development of Raspberry Pi and APM, characterized by: The fully automatic reconnaissance and strike control system, based on the secondary development of Raspberry Pi and APM, includes: an identification and positioning system, which uses image information captured by the camera to identify and locate targets through the ROS system framework and OpenCV image processing technology; an autonomous route planning system, which is used to automatically plan the UAV's flight route based on the target location information; a precision delivery system, which executes precise material delivery or strike missions according to the target positioning information; and a joint communication system, which is used to transmit the UAV's position and speed parameters, delivery target posture and surrounding environment information to the ground station in real time.

2. The fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM according to claim 1 is characterized in that: The recognition and positioning system is divided into a visual recognition module and a target positioning module. It adopts the improved YOLOv5 algorithm, embeds a spatial attention mechanism module in the backbone network, and adds a small target detection head to improve the recognition speed and image recognition accuracy. Subsequently, it combines OpenCV image segmentation and data enhancement methods based on motion image deblurring. It uses image preprocessing and image enhancement technology to process the segmented image, reduce noise and interference, enhance image contrast and edge features, and improve target recognition accuracy.

3. The identification and positioning system according to claim 2, characterized in that: The target localization module is improved by using the spatial attention mechanism and the SimAM module to enhance the recognition capability of small targets and eliminate background interference.

4. The fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM according to claim 1 is characterized in that: The autonomous route planning system is divided into an APM secondary development route planning firmware and a ROS route intelligent decision package. The APM secondary development route planning firmware reads the posture information of the airspeed meter (6) and the IMU, obtains the longitude and latitude of the differential positioning navigation (4), and writes the autonomous flight control firmware to control the joint operation of each module to realize the autonomous flight control of the UAV.

5. The autonomous route planning system according to claim 4, characterized in that: The ROS route intelligent decision package is deployed in the Raspberry Pi host computer (2). After obtaining the specific longitude and latitude of the ground object, it autonomously decides the path passing over the object. The path template comes from the route file package manually planned before takeoff. The expected path and the longitude and latitude of the impact point are calculated according to the L1 navigation and guidance algorithm. The algorithm intelligent decision is performed to obtain the best route, and the new waypoint file is transmitted to the lower computer flight control (1) for executing the new path.

6. The fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM according to claim 1 is characterized in that: The precise delivery system is a secondary development based on the commands_logic command in the APM open source code to suit airdrop scenarios. After receiving the airdrop command, it calculates the distance the aircraft needs to fly and triggers the airdrop action after waiting for a period of time.

7. The fully automatic detection and strike integrated control system based on the secondary development of Raspberry Pi and APM according to claim 1 is characterized in that: The joint communication system is mainly composed of flight control communication link, reconnaissance communication link and flight control and airborne computer joint communication link. The task of the flight control communication link is to establish an air-to-ground two-way data transmission channel to complete the ground control station's long-distance remote control, reconnaissance and strike mission information transmission for the UAV.

8. The joint communication system according to claim 7, characterized in that: The reconnaissance communication link is to transmit the target image information obtained by the upper control system to the ground-side PC through the secondary development of the cloud OSS, use the industrial-grade 4G module (3) to provide network services for the upper control system, and use the OSS cloud to achieve transmission without distance restrictions and frequency band conflict interference.

9. The joint communication system according to claim 7, characterized in that: The flight control and the Raspberry Pi host computer jointly communicate via a joint ROS framework to achieve two-way real-time communication. The MAVROS software package is used to convert ROS topics into MAVLINK format messages and send them to the flight control. The flight control converts the messages into uORB messages through the MAVLINK module, enabling communication between the various modules of the drone.

Images