Operation mode switching method of cage type land-air amphibious unmanned aerial vehicle suitable for complex environment

Abstract

The invention discloses an operation mode switching method of a cage type land-air amphibious unmanned aerial vehicle suitable for a complex environment, and relates to the technical field of unmanned aerial vehicles. Through a flexible land-air amphibious mode switching function, an intelligent control system realizes seamless connection of ground walking and air flying, so that the cage type land-air amphibious unmanned aerial vehicle can adapt to the complex environment; the unmanned aerial vehicle which has dual functions of ground running and air flying and can perform stable operation mode switching becomes an important demand. According to the design of the unmanned aerial vehicle, an automatic and manual switching mode, a 3D visual perception function and a firm cage type structure are integrated, so that the unmanned aerial vehicle can adapt to various complex environments, and the task execution flexibility and safety are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and in particular to an operation mode switching method of a cage-type land-air amphibious unmanned aerial vehicle suitable for complex environments. Background Art

[0002] With the continuous advancement of drone technology, drones are increasingly being used in various fields, including environmental monitoring, disaster relief, military reconnaissance, etc. In complex environments, such as rugged terrain, narrow passages, or areas with dense obstacles, drones face greater challenges. Traditional drones usually only have the ability to fly in the air or walk on the ground, but in these complex environments, a single operating mode may not meet mission requirements.

[0003] In order to improve the adaptability and mission execution efficiency of drones in complex environments, researchers have proposed a land-air amphibious drone design. This design combines the dual functions of walking on the ground and flying in the air, allowing the drone to flexibly switch operating modes under various environmental conditions, thereby effectively coping with different terrains and obstacles. The current commonly used implementation method of land-air amphibious drones is a folding structure. Southwest University disclosed a wheeled air-ground dual-purpose drone with a folding flight bracket in its patent document "A dual-purpose air-ground drone" (patent publication number: CN112046221A). The flight bracket and propeller described in the patent can be folded and retracted in ground mode to reduce the overall size and avoid affecting ground driving or entering narrow spaces. When it is necessary to switch to flight mode, the flight bracket is unfolded, the propeller is exposed and started, and the drone is driven into the air. Although the folding flight bracket has the advantages of compact structure and high flexibility, it still has shortcomings when facing some situations. For example, in a complex environment, the folding bracket is easily affected by wind, sand and debris, causing the joints to jam or deform, making the deformation function ineffective. At the same time, in low light or narrow spaces, the structural strength of the folding bracket may not be sufficient to cope with sudden impacts or collisions, thereby damaging key components of the drone. Summary of the invention

[0004] The present invention provides a method for switching the operation mode of a caged amphibious UAV suitable for complex environments, which can seamlessly switch between the ground walking mode and the air flight mode to meet the needs of various complex environments. The UAV is particularly suitable for irregular terrains such as narrow, rugged, collapsed, and post-disaster rescue, complex terrain exploration, and other mission scenarios.

[0005] The technical solution adopted by the present invention is as follows: a method for switching the operation mode of a cage-type amphibious UAV suitable for a complex environment, wherein the steps are:

[0006] 1. Image preprocessing

[0007] 1.1 Grayscale image conversion: The visual system carried by the drone collects image data of the surrounding environment and converts the image into a grayscale image.

[0008] 1.2 Image region division: The image in the drone’s field of view is divided into 9 regions. The partitioning strategy is used for the drone to perceive and process the surrounding environment in different directions.

[0009] 2. 3D modeling based on binocular vision

[0010] 2.1 Binocular vision system: Using the left and right images obtained by the drone’s binocular camera, the depth information of the environment is obtained through parallax calculation. By matching the feature points in the left and right images, the depth value of each pixel is calculated to generate a three-dimensional environment point cloud;

[0011] 2.2ORB-SLAM algorithm: ORB algorithm is used to extract and match key feature points, complete the positioning and map construction of the UAV, and correct the posture of the UAV in real time;

[0012] The specific method is: divide the drone's field of view into nine 3×3 areas, each area is A i,j , where i∈{1,2,3,}, j∈{1,2,3}, and the area in the direction of the drone is A 2,1 ,A 2,2 ,A 2,3 ;

[0013] The width of the area in the direction of the drone's advance is calculated by the binocular vision system. When the drone is at the center of the image, its width is calculated using the following formula:

[0014]

[0015] Among them, W pix is the pixel width of the drone at the center of the field of view, W real is the real width of the drone, B is the baseline distance between the binocular cameras, and d is the parallax of the binocular cameras;

[0016] After dividing the field of view, the A of the image is analyzed at set intervals. 2,1 ,A 2,2 ,A 2,3 Edge extraction is performed on the region to obtain significant structural information in the image:

[0017]

[0018] Among them, I(w,h) is the gray value of the pixel in the image, and are the gradients of the image in the w direction and h direction respectively, and G(w,h) is the result of edge extraction;

[0019] During the ground driving of the drone, the variance of the edge features reflects the ruggedness of the current terrain. When the variance reaches the threshold, the algorithm will switch the drone to flight mode. The specific formula for calculating the variance and switching the mode is as follows:

[0020]

[0021] where σ 2 is the variance of the image, H, W represent the height and width of the image respectively, and μ represents the average grayscale of the image.

[0022] 3. Dynamic adjustment of the environment and autonomous decision-making

[0023] 3.1 Dynamic environment perception: The control system of the drone monitors the surrounding environment in real time through the processor and sensors, obtains the location information of dynamic objects and terrain features, and the processor adjusts the motion state of the drone in real time based on the sensor data and the results of 3D modeling;

[0024] 3.2 Autonomous decision-making and mode switching: Under normal circumstances, the drone will prioritize ground mode to save power. When it detects an impassable obstacle or complex terrain ahead, the system will automatically switch to flight mode to avoid the obstacle and ensure the successful completion of the mission. In flight mode, the drone can accurately navigate in narrow spaces or complex environments using sensors and three-dimensional map data;

[0025] When the detected variance σ 2 Exceeding the set threshold When the current terrain is judged to be too rugged and not suitable for continuing to drive in ground mode, the drone switches to flight mode. The specific switching conditions are:

[0026]

[0027] in Partition A for the drone's forward direction 2,1 ,A 2,2 ,A 2,3 The variance value of , max(·) is the maximum value function;

[0028] When the above switching conditions are met, the control system will automatically switch to flight mode according to the following formula:

[0029]

[0030] When the drone is in ground mode, in order to keep the drone moving smoothly and efficiently, the algorithm adjusts the output of the flight component according to the ground variance. The formula is as follows:

[0031]

[0032] Among them, F(t) is the output of the flight component at the current time t, F0 is the output of the drone in a stable state, and K p ,K i ,K d are the gain coefficients of proportional control, integral control and differential control, which are set to 0.5, 0.01 and 0.02 respectively; is the variance error between the current image and the road surface in a stable state, is the error integral from time 0 to the current time t, which is used to correct the accumulated error. The differential of the error is used to predict the trend of the error.

[0033] 4. Human Intervention

[0034] Manual intervention: In addition to automatic mode switching, users can manually intervene in the drone through the remote control device and manually switch the drone's movement mode.

[0035] The hardware structure of the cage-type amphibious drone includes:

[0036] It consists of cage structure, flight components, control system, GPS system, binocular stereo vision system, sensor system and battery system;

[0037] The cage structure is composed of the side structural support parts, fence parts, and ring structures through mortise and tenon joints. Screws are used to reinforce the weak parts. Cameras are installed on the left and right sides of the cage structure to form a 3D vision system.

[0038] The cage structure is connected to the drone body through bearings. The flight assembly provides power for the movement of the drone. Through coordination with the cage structure and the control system, the drone can move smoothly in an amphibious environment. A GPS system is installed above the control system for real-time positioning and navigation.

[0039] The 3D vision system is connected to the flight component through bearings. The flight component can adjust the pitch angle of the camera as needed to enhance environmental perception. The 3D vision system communicates with the control system through a data line to achieve data transmission and processing;

[0040] A data interface is reserved on the mainboard of the control system as a connection port for the sensor system, allowing the optional installation of a variety of sensors;

[0041] The bottom of the drone is equipped with a battery system, which is installed through an aviation plug and a card slot to provide stable power support for the entire drone system.

[0042] The beneficial effects created by the present invention are as follows: the present invention realizes seamless connection between ground walking and air flying through the flexible land-air amphibious mode switching function and the intelligent control system, which significantly improves the adaptability and task execution efficiency in complex terrain. Combined with the autonomous navigation and obstacle avoidance functions of high-performance processors, GPS positioning systems and multi-sensors, the UAV can operate stably under harsh conditions and has an automatic return function. In addition, the quick replacement battery system extends the operating endurance time and is suitable for long-term mission requirements. The UAV is particularly suitable for various mission scenarios such as post-disaster rescue and complex terrain exploration, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A top view of the overall structure of the drone

[0044] Figure 2 This is the front view of the overall structure of the drone;

[0045] Figure 3 This is a schematic diagram of the cage structure of a drone;

[0046] Figure 4 It is a schematic diagram of the UAV ground mode motion method;

[0047] Figure 5 Flowchart for switching between ground and air modes;

[0048] Figure 6 This is a schematic diagram of the UAV's field of view partition;

[0049] Figure 7 Schematic diagram of edge detection effect;

[0050] Figure 8 Flowchart of the road condition assessment algorithm. DETAILED DESCRIPTION

[0051] Embodiment 1:

[0052] The overall structure of the drone is as follows Figure 1-3As shown. The drone of the present invention adopts a cage structure 1 made of carbon fiber material to ensure structural strength and rigidity, and can effectively resist external impact. It is internally configured with a flight component 2, and the flight component 2 is at least four rotors, providing vertical take-off and landing and high maneuverability. The control system 3 is responsible for real-time data processing and mode switching, and the GPS system 4 realizes position tracking and automatic return functions. The binocular stereo vision systems 5 on both sides are high-definition cameras 5 for remote control and environmental monitoring, and the sensor system 6 supports autonomous navigation and intelligent obstacle avoidance. The battery system 7 is designed to be easy to replace and is equipped with a quick replacement interface to extend the operating time and endurance.

[0053] 1. Cage structure design

[0054] The main structure of the drone is made of high-strength, lightweight carbon fiber material, which is cut according to design requirements. In order to strengthen the weak parts of the frame structure, screws are used to reinforce it and improve the strength of the overall structure. The main components of the drone are installed inside the cage structure. Figure 3 As shown, the cage structure of the UAV is composed of two structural support parts 1-1, eight fence parts 1-2 and two ring structures 1-3.

[0055] The structural support part 1-1 is a side cage structure made of carbon fiber material. It has holes and bearings in the middle for connecting the flight component 2 and the binocular stereo vision system 5. In order to keep the drone's vision clear and form 3D stereo vision, the binocular stereo vision system 5 is installed on the outside of the cage structure 1.

[0056] Since the fence part 1-2 will be impacted by the airflow generated when the flight assembly is in operation, it is designed as a hollow streamlined structure to balance the weight while reducing the interference with the flight airflow.

[0057] The ring structure 1-3 is used to stabilize the cage structure of the drone and protect the interior. When moving on the ground, it can also be used as the wheel of the drone, connected by carbon fiber screws 1-3a, and used to fix the structural support part 1-1 and the fence part 1-2 to play a structural protection role.

[0058] The structural support part 1-1, the fence part 1-2 and the ring structure 1-3 are connected by mortise and tenon technology to form a cage structure, so as to maintain the stability of the cage structure as much as possible while reducing weight. The above parts are all modularly designed, and when parts are damaged, they can be quickly replaced to restore the normal operation of the drone as soon as possible.

[0059] It consists of a cage structure 1, a flight component 2, a control system 3, a GPS system 4, a binocular stereo vision system 5, a sensor system 6 and a battery system 7;

[0060] Binocular Stereo Vision System5: The drone is equipped with binocular cameras, which are installed at both ends of the drone to avoid being blocked by the cage structure when walking on the ground or flying. The cameras on both sides are calibrated so that the binocular vision system can obtain depth data by calculating the parallax information between the cameras and generate a three-dimensional point cloud map of the environment for the drone. The binocular vision system can sense the position and depth of obstacles, and analyze the three-dimensional data through the processor to determine the path planning and obstacle avoidance plan. It is especially suitable for complex underground environments or small spaces.

[0061] A method for switching operating modes of a caged land-air amphibious UAV suitable for complex environments:

[0062] In order to reduce the amount of calculation and reduce the impact of lighting changes on the UAV algorithm, a strategy based on grayscale image processing and region division is adopted. The UAV captures real-time video streams through a binocular camera. Since changes in lighting conditions may affect the stability of color information, the image is first grayed to retain brightness information, thereby reducing the interference caused by lighting changes. The grayscale formula is as follows:

[0063] Gray=0.299×Red+0.587×Green+0.114×Blue (7)

[0064] Among them, Red, Green, and Blue are the red, green, and blue channel values ​​of the original image respectively.

[0065] Field of view division: Figure 6 As shown in the figure, the drone’s field of view is divided into nine 3×3 areas, each area is A i,j , where j∈{1,2,3,},i∈{1,2,3}. The area in the direction of the drone is A 2,1 ,A 2,2 ,A 2,3 Among them, the width of the area in the direction of the drone's advance is obtained by using binocular vision to calculate the width of the drone at the center of the image on the ground. The calculation formula is as follows:

[0066]

[0067] Where W pix is the pixel width of the drone at the center of the field of view, W real is the real width of the drone. B is the baseline distance between the binocular cameras, and d is the parallax of the binocular cameras.

[0068] After dividing the field of view, the image A is analyzed at regular intervals. 2,1 ,A 2,2 ,A 2,3 Perform edge extraction in the region to obtain significant structural information in the image:

[0069]

[0070] Among them, I(w,h) is the gray value of the pixel in the image, and are the gradients of the image in the w direction and h direction respectively, and G(w,h) is the result of edge extraction.

[0071] During the ground driving of the drone, the variance of the edge features reflects the ruggedness of the current terrain. When the variance reaches a certain value, the algorithm will switch the drone to flight mode. The specific formula for calculating the variance and switching the mode is as follows:

[0072]

[0073] where σ 2 is the variance of the image, H, W represent the height and width of the image respectively, and μ represents the average grayscale of the image.

[0074] When the detected variance σ 2 Exceeding the set threshold When the algorithm determines that the current terrain is too rugged and not suitable for continuing to drive in ground mode, the drone switches to flight mode. The specific switching conditions are:

[0075]

[0076] When the above switching conditions are met, the control system will automatically switch to flight mode according to the following formula:

[0077]

[0078] When the drone is in ground mode, in order to keep the drone moving smoothly and efficiently, the algorithm adjusts the output of the flight component according to the ground variance. The formula is as follows:

[0079]

[0080] Among them, F(t) is the output of the flight component at the current time t, F0 is the output of the drone in a stable state, and K p ,K i ,K d are the gain coefficients of proportional control, integral control and differential control, which are set to 0.5, 0.01 and 0.02 respectively; is the variance error between the current image and the road surface in a stable state, is the error integral from time 0 to the current time t, which is used to correct the accumulated error. The differential of the error is used to predict the trend of the error.

Claims

1. A method for switching the operation mode of a cage-type amphibious UAV suitable for complex environments, characterized in that: The steps are:

1. Image preprocessing 1.1 Grayscale image conversion: The visual system carried by the drone collects image data of the surrounding environment and converts the image into a grayscale image; 1.2 Image region division: The image in the drone’s field of view is divided into 9 regions. The partitioning strategy is used for the drone to perceive and process the surrounding environment in different directions; 2. 3D modeling based on binocular vision 2.1 Binocular vision system: Using the left and right images obtained by the drone’s binocular camera, the depth information of the environment is obtained through parallax calculation. By matching the feature points in the left and right images, the depth value of each pixel is calculated to generate a three-dimensional environment point cloud; 2.2ORB-SLAM algorithm: ORB algorithm is used to extract and match key feature points, complete the positioning and map construction of the UAV, and correct the posture of the UAV in real time; 3. Dynamic adjustment of the environment and autonomous decision-making 3.1 Dynamic environment perception: The control system of the drone monitors the surrounding environment in real time through the processor and sensors, obtains the location information of dynamic objects and terrain features, and the processor adjusts the motion state of the drone in real time based on the sensor data and the results of 3D modeling; 3.2 Autonomous decision-making and mode switching: Under normal circumstances, the drone will prioritize ground mode to save power. When it detects an impassable obstacle or complex terrain ahead, the system will automatically switch to flight mode to avoid the obstacle and ensure the successful completion of the mission. In flight mode, the drone can accurately navigate in narrow spaces or complex environments using sensors and three-dimensional map data; 4. Human Intervention Manual intervention: In addition to automatic mode switching, users can manually intervene in the drone through the remote control device and manually switch the drone's movement mode.

2. The method for switching the operation mode of a cage-type amphibious drone suitable for complex environments according to claim 1, characterized in that: In the above 1.2, the specific method is: The drone’s field of view is then divided into nine 3×3 areas, each of which is A i,j , where i∈{1,2,3,},j∈{1,2,3}, and the area in the direction of the drone is A 2,1 ,A 2,2 ,A 2,3 ; The width of the area in the direction of the drone's advance is calculated by the binocular vision system. When the drone is at the center of the image, its width is calculated using the following formula: Among them, W pix is the pixel width of the drone at the center of the field of view, W real is the real width of the drone, B is the baseline distance between the binocular cameras, and d is the parallax of the binocular cameras; After dividing the field of view, the A of the image is analyzed at set intervals. 2,1 ,A 2,2 ,A 2,3 Perform edge extraction in the region to obtain significant structural information in the image: Among them, I(w,h) is the gray value of the pixel in the image, and are the gradients of the image in the w direction and h direction respectively, and G(w,h) is the result of edge extraction; During the ground driving of the drone, the variance of the edge features reflects the ruggedness of the current terrain. When the variance reaches the threshold, the algorithm will switch the drone to flight mode. The specific formula for calculating the variance and switching the mode is as follows: where σ 2 is the variance of the image, H, W represent the height and width of the image respectively, and μ represents the average grayscale of the image.

3. The method for switching the operation mode of a cage-type amphibious drone suitable for complex environments according to claim 1, characterized in that: In the above 3.2, the specific method is: When the detected variance σ 2 Exceeding the set threshold When the current terrain is judged to be too rugged and not suitable for continuing to drive in ground mode, the drone switches to flight mode. The specific switching conditions are: in Partition A for the drone's forward direction 2,1 ,A 2,2 ,A 2,3 The variance value of , max(·) is the maximum value function; When the above switching conditions are met, the control system will automatically switch to flight mode according to the following formula: When the drone is in ground mode, in order to keep the drone moving smoothly and efficiently, the algorithm adjusts the output of the flight component according to the ground variance. The formula is as follows: Among them, F(t) is the output of the flight component at the current time t, F0 is the output of the drone in a stable state, and K p ,K i ,K d are the gain coefficients of proportional control, integral control and differential control, which are set to 0.5, 0.01 and 0.02 respectively; is the variance error between the current image and the road surface in a stable state, is the error integral from time 0 to the current time t, which is used to correct the accumulated error. The differential of the error is used to predict the trend of the error.

4. The method for switching the operation mode of a cage-type amphibious drone suitable for complex environments according to claim 1, characterized in that: The hardware structure of the cage-type amphibious drone includes: It consists of a cage structure (1), a flight assembly (2), a control system (3), a GPS system (4), a binocular stereo vision system (5), a sensor system (6) and a battery system (7); The cage structure (1) is composed of a side structural support part (1-1), a fence part (1-2), and a ring structure (1-3) through mortise and tenon joints, and is reinforced with screws (1-3a) at weak parts. Cameras are installed on the left and right sides of the cage structure (1) to form a 3D vision system (5); The cage structure (1) is connected to the drone body through a bearing, and the flight component (2) provides power for the movement of the drone. Through coordination with the cage structure (1) and the control system (3), the drone can move smoothly in an amphibious environment. A GPS system (4) is installed above the control system (3) for real-time positioning and navigation. The 3D vision system (5) is connected to the flight assembly (2) via a bearing, and the flight assembly (2) can adjust the pitch angle of the camera as needed to enhance the environmental perception capability. The 3D vision system (5) communicates with the control system (3) via a data line to achieve data transmission and processing; A data interface is reserved above the mainboard of the control system (3) as a connection port for the sensor system (6), allowing the optional installation of a variety of sensors; A battery system (7) is disposed at the bottom of the drone and is installed through an aviation plug and a card slot to provide stable power support for the entire drone system.

Images