Unmanned aerial vehicle autonomous landing method and system

By judging the environment in the drone, obtaining satellite or image position information, and using optimization model and Kalman filter prediction model for landing path planning, the problem of inaccurate landing of the drone is solved, and the drone in the park is realized. The stability of safe, precise landing and coordinated operation of unmanned vehicles.

CN118915806APending Publication Date: 2024-11-08NINGBO UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410878583.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-02
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

When landing in the park, drones are susceptible to visual light, obstruction of field of view and unstable satellite signals carried by drones, resulting in inaccurate landing.

Method used

The drone obtains satellite or image position information by determining whether the environment meets whether it meets the satellite or image position information. If it meets the environment, it obtains the corresponding information and lands. If it is not met, it hoveres and waits. Returns to determine whether the environment meets the satellite position information, and uses the optimization model and extended Kalman. The filter prediction model and the landing path planning model perform accurate landing.

Benefits of technology

It has achieved safe and accurate landing of drones in the park, improved the stability of coordinated operation between drones and unmanned vehicles, and enhanced the accuracy and reliability of landing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118915806A_ABST
    Figure CN118915806A_ABST
Patent Text Reader

Abstract

The invention relates to an unmanned aerial vehicle autonomous landing method and system, and the method comprises the steps: firstly judging whether a surrounding environment meets a condition for obtaining the satellite position information of a landing position through an unmanned aerial vehicle, and directly obtaining the satellite position information of the landing position when the surrounding environment meets the condition for obtaining the satellite position information of the landing position; when the condition of obtaining the satellite position information of the landing position is not met, the unmanned aerial vehicle judges whether the surrounding environment meets the condition of obtaining the image position information of the landing position or not, and when the condition of obtaining the image position information of the landing position is met, the image position information of the landing position is directly obtained; and when the condition of obtaining the image position information of the landing position is not met, the unmanned aerial vehicle hovers and waits, and after a preset time, the unmanned aerial vehicle is returned to judge whether the current driving environment of the unmanned aerial vehicle meets the condition of obtaining the satellite position information of the landing position or not.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of autonomous landing of unmanned aerial vehicles, and in particular to an autonomous landing method and system for unmanned aerial vehicles. Background Art

[0002] With the rapid development of drone and autonomous driving technology, the field of logistics and cargo transportation in the park is ushering in a broader development space and innovation possibilities. However, in the park scene where drones are required to land on unmanned vehicles in real time and accurately, they are facing a series of changes and challenges, such as drones are easily affected by visual light, field of vision is blocked, and the satellite signal of the unmanned vehicle is unstable. This paper aims to propose a series of innovative optimization methods and collaborative strategies to address these challenges, so as to ensure that drones can land safely and accurately on unmanned vehicles, and achieve the coordinated operation and stability of the two. Summary of the invention

[0003] Based on this, it is necessary to provide an autonomous landing method and system for drones to address the problem that traditional drones are easily affected by visual light, obstructed field of view, and unstable satellite signals carried by unmanned vehicles.

[0004] The present application provides a method for autonomous landing of a drone, comprising:

[0005] The drone determines whether the current driving environment of the drone meets the conditions for obtaining the satellite position information of the landing location;

[0006] If the current driving environment of the UAV meets the conditions for obtaining the satellite position information of the landing position, the UAV obtains the satellite position information of the landing position and performs the landing action according to the satellite position information of the landing position;

[0007] If the current driving environment of the drone does not meet the conditions for obtaining the satellite position information of the landing position, the drone determines whether the current driving environment of the drone meets the conditions for obtaining the image position information of the landing position;

[0008] If the current driving environment of the drone meets the conditions for obtaining the image position information of the landing position, the drone obtains the image position information of the landing position, and performs a landing action according to the image position information of the landing position;

[0009] If the current driving environment of the drone does not meet the conditions for obtaining the image position information of the landing position, the drone hovers and waits, and returns to the drone to determine whether the current driving environment of the drone meets the conditions for obtaining the satellite position information of the landing position.

[0010] The present application also provides a drone autonomous landing system, comprising:

[0011] Autonomous vehicles;

[0012] A drone, comprising a drone body and a processing device, wherein the processing device is used to execute the drone autonomous landing method as described above.

[0013] The present application relates to an autonomous landing method and system for a drone, wherein the drone first determines whether the surrounding environment meets the conditions for obtaining satellite position information of the landing position, and directly obtains the satellite position information of the landing position when the conditions for obtaining the satellite position information of the landing position are met; when the conditions for obtaining the satellite position information of the landing position are not met, the drone determines whether the surrounding environment meets the conditions for obtaining image position information of the landing position, and directly obtains the image position information of the landing position when the conditions for obtaining the image position information of the landing position are met; when the conditions for obtaining the image position information of the landing position are not met, the drone hovers and waits, and after a preset time, returns to the drone to determine whether the current driving environment of the drone meets the conditions for obtaining the satellite position information of the landing position. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A schematic diagram of a flow chart of a method for autonomous landing of a drone provided in one embodiment of the present application.

[0015] Figure 2 A schematic diagram of the structure of an autonomous landing system for a drone provided in one embodiment of the present application.

[0016] Figure 3 A schematic diagram of the structure of a drone in a drone autonomous landing system provided in one embodiment of the present application.

[0017] Figure 4 A schematic diagram of the structure of an unmanned vehicle in a drone autonomous landing system provided in one embodiment of the present application.

[0018] Reference numerals:

[0019] 11. Unmanned vehicle; 111. Unmanned vehicle body; 112. Landing platform; 113. Landing sign image;

[0020] 114. second positioning member; 12. drone; 121. drone body; 122. processing device;

[0021] 123. Camera; 124. First positioning member; 125. Receiving device. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0023] like Figure 1 As shown, in one embodiment of the present application, the autonomous landing method of the drone includes the following steps S100 to S500:

[0024] S100, the UAV determines whether the current driving environment of the UAV satisfies the conditions for obtaining the satellite position information of the landing position.

[0025] S200: If the current driving environment of the drone satisfies the condition for obtaining the satellite position information of the landing position, the drone obtains the satellite position information of the landing position and performs a landing action according to the satellite position information of the landing position.

[0026] S300, if the current driving environment of the drone does not meet the conditions for obtaining the satellite position information of the landing position, the drone determines whether the current driving environment of the drone meets the conditions for obtaining the image position information of the landing position;

[0027] S400: If the current driving environment of the drone meets the conditions for obtaining the image position information of the landing position, the drone obtains the image position information of the landing position and performs a landing action according to the image position information of the landing position.

[0028] S500: If the current driving environment of the drone does not meet the conditions for obtaining the image position information of the landing position, the drone hovers and waits, and returns to the drone to determine whether the current driving environment of the drone meets the conditions for obtaining the satellite position information of the landing position.

[0029] Specifically, before step S100, the process also includes respectively creating an optimization model, an extended Kalman filter prediction model and a landing path planning model.

[0030] Creating an optimization model involves:

[0031] Get the backbone network, neck network and output end of the YOLOv5-Lite model;

[0032] For the backbone network of the YOLOv5-Lite model, CBRM is used as the initial step of the backbone network of the YOLOv5-Lite model; further, multiple Shuffle_Blocks of ShuffleNetV2 are placed into the backbone network of the YOLOv5-Lite model; further, a SE_Layer is added after each Shuffle_Block of ShuffleNetV2, and a SE_Layer is also added after the last Shuffle_Block of ShuffleNetV2;

[0033] Specifically, the standard Shuffle_Block is still used for feature extraction and channel shuffling between two adjacent Shuffle_Blocks of ShuffleNetV2;

[0034] For the neck network of the YOLOv5-Lite model, a SE_Layer is added after each C3False, and the results of the neck network of the YOLOv5-Lite model are input to the output end of the YOLOv5-Lite model by SE_Layer;

[0035] For the output end of the YOLOv5-Lite model, the result obtained by the neck network of the YOLOv5-Lite model is Conv2d and then output.

[0036] Furthermore, after the ShuffleNetV2 lightweight module Shuffle_Block was added to YOLOv5-Lite, compared with the original YOLOv5-Lite, while maintaining a high accuracy, FLOPS was reduced from 15.9G to 3.7G, the parameter size was reduced from 14.3MB to 3.4MB, the FPS was faster from 101f / s to 110f / s, and the inference speed was reduced from 49.533ms to 16.319ms; after adding the SE_Layer attention mechanism, the accuracy of the trained model was improved by 0.07%, the accuracy was 99.99%, the recall rate remained unchanged, and the FLOPS was reduced by 2.8G. For airborne computers with small computing power, small FLOPS is more suitable, the parameter size is also reduced by 2.5MB, the frame rate is 8f / s faster, and the inference speed is reduced by 10.405ms.

[0037] Creating an extended Kalman filter prediction model involves:

[0038] Get the drone position P at time t u =[x u ,y u ,z u ] T , target position P g =[x g ,y g ,z g ] T , acceleration of the target position and the speed of the target location

[0039] Create the equation of state:

[0040] X(t+1)=F(t)X(t)+B(t)u(t)+w(t)

[0041] Among them, X(t) is the position of the drone and the target at time t; X(t+1) is the position of the drone and the target at time t+1; u(t) is the acceleration of the target position; w(t) is the process noise, which is represented by p(w(t))~N(0,Q) and is a Gaussian distribution with a covariance matrix of Q; F(t) is the state transfer matrix; B(t) is the control input matrix, which is converted from external control input to the change of state vector, and is used to describe how small acceleration affects the state;

[0042] The difference between time t and the next time t+1 is Δt, which is the period of system data transmission;

[0043] The state transfer function of the extended Kalman filter at this time is:

[0044] f(x(t),u(t),w(t)):X(t+1)=f(X(t),u(t))+w(t) The state transfer matrix and control input matrix expressions are:

[0045]

[0046] Based on the received UAV position and target position, the position and angle information are selected as the observation matrix to obtain the system measurement equation:

[0047] Z(t)=h(X(t))+v(t)

[0048] Where Z(t) is the observation value at time t, obtained by the sensor, v(t) is the measurement noise, represented by p(v(t))~N(0,R), where R is the measurement noise covariance matrix;

[0049] Get the attitude of the drone at time t, including the pitch angle θ, roll angle φ and yaw angle

[0050] Further observation parameters of the pitch angle, roll angle and yaw angle of the drone are:

[0051]

[0052] Then the measurement equation is expressed as:

[0053]

[0054] The Jacobian matrix is ​​used to linearize the observation equation to obtain the linearized representation of the observation equation:

[0055]

[0056] Further carefully predict the target location;

[0057]

[0058] in, is the prior prediction of the state at time t+1, is the a priori prediction of the covariance at time t, u(t+1) is the control input at time t+1. P(t+1|t) is the a priori prediction of the covariance at time t, P(t|t) is the a priori prediction of the covariance at time t, F(t+1) is the state transfer matrix, and Q(t+1) is the process noise covariance matrix.

[0059] The Kalman gain is used to weigh prediction and measurement uncertainty, where K(t+1) is the Kalman gain at time t+1, H(t+1) is the Jacobian matrix of the observation matrix, and R(t+1) is the covariance matrix of the observation noise. is the posterior prediction of the state at time t+1, P(t+1|t+1) is the posterior prediction of the covariance at time t, Z(t+1) is the actual observation at time t+1, and I is the identity matrix.

[0060] Creating a landing path planning model includes:

[0061] Get the current status of the drone, including its position and speed;

[0062] Next, the location of the landing position is obtained, which is provided by satellite position information or image position information;

[0063] Inputting the obtained landing position into the extended Kalman filter prediction model, starting the extended Kalman filter prediction model, and obtaining the predicted position of the landing position output by the extended Kalman filter prediction model;

[0064] The predicted landing position is input into the A* path planning algorithm, the A* path planning algorithm is started, and the optimal path from the current position of the UAV to the predicted landing position output by the A* path planning algorithm is obtained.

[0065] In this embodiment, the drone first determines whether the surrounding environment meets the conditions for obtaining the satellite position information of the landing position. When the conditions for obtaining the satellite position information of the landing position are met, the satellite position information of the landing position is directly obtained; when the conditions for obtaining the satellite position information of the landing position are not met, the drone determines whether the surrounding environment meets the conditions for obtaining the image position information of the landing position. When the conditions for obtaining the image position information of the landing position are met, the image position information of the landing position is directly obtained; when the conditions for obtaining the image position information of the landing position are not met, the drone hovers and waits, and after a preset time, returns to the drone to determine whether the current driving environment of the drone meets the conditions for obtaining the satellite position information of the landing position.

[0066] In one embodiment of the present application, the image position information of the landing position includes a landing mark image.

[0067] Specifically, the landing mark image is a map with a set landing mark, and the surface of the map with the set landing mark is flat.

[0068] In one embodiment of the present application, the drone obtains the image position information of the landing position including the following steps S400a to S400c:

[0069] S400a, acquiring a landing mark image.

[0070] S400b, inputting the landing mark image into the optimization model, starting the optimization model, and obtaining the target landing mark image output by the optimization model.

[0071] S400c, parsing the target landing mark image to obtain position information of the landing mark in the target landing mark image.

[0072] Specifically, the landing sign image is obtained by taking pictures from the perspective of a drone;.

[0073] In this embodiment, the landing mark image obtained by shooting from the perspective of the drone is input into the optimization model, and the optimization model is started to obtain the target landing mark image output by the optimization model. The target landing mark image is analyzed to obtain the position information of the landing mark in the target landing mark image.

[0074] In one embodiment of the present application, the satellite position information of the landing position includes satellite coordinate information of the landing position.

[0075] Specifically, the satellite coordinate information of the landing position is three-dimensional coordinate data of the landing position, and the data represents a point, which is any point in a circle with a preset length as the radius and the center point of the landing mark as the center in the landing mark image.

[0076] In one embodiment of the present application, if the current driving environment of the drone meets the condition of obtaining the satellite position information of the landing position, the drone obtains the satellite position information of the landing position, and performs a landing action according to the satellite position information of the landing position, including the following S201 to S206:

[0077] S201, the UAV establishes a three-dimensional geographic coordinate system with the take-off point of the UAV as the origin.

[0078] S202, the UAV receives the satellite position information of the landing position sent by the unmanned vehicle.

[0079] S203, the drone converts the drone's location information and the satellite location information of the landing location into a three-dimensional geographic coordinate system to obtain the drone's location coordinates and the landing location coordinates.

[0080] S204, input the position coordinates of the UAV and the position coordinates of the landing position into the extended Kalman filter prediction model, start the extended Kalman filter prediction model, and obtain the predicted position coordinates of the UAV and the predicted position coordinates of the landing position after a preset time output by the extended Kalman filter prediction model.

[0081] S205, inputting the predicted position coordinates of the UAV and the predicted position coordinates of the landing position into the landing path planning model, starting the landing path planning model, and obtaining the predicted landing path output by the landing path planning model.

[0082] S206, the drone uses the predicted landing path to execute the landing action.

[0083] Specifically, the three-dimensional geographic coordinate system is established based on the take-off point of the UAV, and the satellite position information of the landing position at each moment and the satellite position information of the UAV need to be converted into the three-dimensional geographic coordinate system to establish a unified position reference.

[0084] In this embodiment, the satellite position information of the landing position and the satellite position information of the UAV at a moment are obtained, and the obtained satellite position information of the landing position and the satellite position information of the UAV are input into the extended Kalman filter prediction model, the extended Kalman filter prediction model is started to obtain the predicted position coordinates of the UAV at the next moment and the predicted position coordinates of the landing position output by the extended Kalman filter prediction model, and then the predicted position coordinates of the UAV at the next moment and the predicted position coordinates of the landing position are input into the landing path planning model, the landing path planning model is started, and the predicted landing path output by the landing path planning model is obtained.

[0085] In one embodiment of the present application, if the current driving environment of the drone meets the condition for obtaining the image position information of the landing position, the drone obtains the image position information of the landing position, and performs the landing action according to the image position information of the landing position, including the following steps S401 to S407:

[0086] S401, the UAV establishes a three-dimensional geographic coordinate system with the take-off point of the UAV as the origin.

[0087] S402, the UAV obtains the location information of the landing mark.

[0088] S403, the drone analyzes the location information of the landing mark, and the drone obtains the coordinate information of the landing mark.

[0089] S404, the drone converts the coordinate information of the drone and the coordinate information of the landing mark into a three-dimensional geographic coordinate system to obtain the position coordinates of the drone and the position coordinates of the landing position.

[0090] S405, input the position coordinates of the UAV and the position coordinates of the landing position into the extended Kalman filter prediction model, start the extended Kalman filter prediction model, and obtain the predicted position coordinates of the UAV and the predicted position coordinates of the landing position after a preset time output by the extended Kalman filter prediction model.

[0091] S406, input the predicted position coordinates of the UAV and the predicted position coordinates of the landing position into the landing path planning model, start the landing path planning model, and obtain the predicted landing path output by the landing path planning model.

[0092] S407, the drone uses the predicted landing path to execute the landing action.

[0093] Specifically, the three-dimensional geographic coordinate system is established based on the take-off point of the drone, and the location information of the landing position at each moment and the location information of the drone need to be converted into the three-dimensional geographic coordinate system to establish a unified location reference.

[0094] In this embodiment, the coordinate information of the landing mark based on the three-dimensional geographic coordinate system at the current moment and the coordinate information of the UAV are input into the extended Kalman filter prediction model, the extended Kalman filter prediction model is started to obtain the predicted position coordinates of the UAV at the next moment and the predicted position coordinates of the landing position output by the extended Kalman filter prediction model, and then the predicted position coordinates of the UAV at the next moment and the predicted position coordinates of the landing position are input into the landing path planning model, the landing path planning model is started, and the predicted landing path output by the landing path planning model is obtained.

[0095] The present application also provides an autonomous landing system for unmanned aerial vehicles. The autonomous landing system for unmanned aerial vehicles applies the autonomous landing method for unmanned aerial vehicles mentioned in the aforementioned embodiment. It should be noted that, for the sake of simplicity, all the devices, equipment, devices and other structures mentioned in this application are numbered in the embodiment section of the autonomous landing system for unmanned aerial vehicles, and are not numbered in the various embodiments of the aforementioned autonomous landing method for unmanned aerial vehicles. It should be noted that the devices, equipment, devices and other structures mentioned in the various embodiments of the aforementioned autonomous landing method for unmanned aerial vehicles, and the devices, equipment, devices and other structures mentioned in the embodiments of the autonomous landing system for unmanned aerial vehicles, if the names are the same, belong to the same devices, equipment, devices, thereby indicating that the autonomous landing system for unmanned aerial vehicles can apply each embodiment of the aforementioned autonomous landing method for unmanned aerial vehicles.

[0096] like Figure 2As shown, in one embodiment of the present application, a drone autonomous landing system is also provided, wherein the drone 12 autonomous landing system comprises an unmanned vehicle 11 and a drone 12

[0097] The drone 12 includes a drone 12 body and a processing device. The processing device is used to execute the autonomous landing method of the drone 12 as described above.

[0098] In this embodiment, the state of the unmanned vehicle 11 can be driving or parked. The processing device on the main body of the drone 12 analyzes and calculates the position information of the unmanned vehicle 11 and the position information of the drone 12 to obtain the flight path of the drone 12 landing on the unmanned vehicle 11 to help the drone 12 land.

[0099] like Figure 3 As shown, in one embodiment of the present application, the drone 12 further includes a camera 121 , a first positioning member 122 and a receiving device 123 .

[0100] The camera 121 is used to obtain an image with a landing mark.

[0101] The first positioning element 122 is used to obtain the position information of the main body of the drone 12 .

[0102] The receiving device 123 is used to receive the location information of the landing position sent by the unmanned vehicle 11.

[0103] Specifically, the camera 121 is installed on the main body of the drone 12, and the camera 121 includes at least two lens modules, which are used to identify landing signs. The setting of multiple lens modules can improve the coverage of landing sign recognition and further improve the recognition efficiency of landing signs. Multiple lens modules can capture the landing signs into images and then transmit them to the processing device; the first positioning member 122 is the inertial measurement unit (IMU) of the drone 12 itself, which can obtain the three-axis attitude angle (or angular velocity) and acceleration of the drone 12 itself in real time, and then obtain the position coordinates of the drone 12 based on the take-off point of the drone 12 as the origin.

[0104] like Figure 4 As shown, in one embodiment of the present application, the unmanned vehicle 11 includes an unmanned vehicle 11 body, a landing platform 112 , a landing mark image 113 and a second positioning member 114 .

[0105] The landing platform 112 is disposed on the main body of the unmanned vehicle 11 .

[0106] The landing mark image 113 is attached to the top surface of the landing platform 112 .

[0107] The second positioning element 114 is used to obtain position information of the landing mark image 113 .

[0108] Specifically, the second positioning component 114 includes a GPS locator and an inertial measurement unit (IMU) of the unmanned vehicle 11 itself. The inertial measurement unit can obtain the three-axis attitude angle (or angular velocity) and acceleration of the unmanned vehicle 11 itself in real time, and then the position information of the unmanned vehicle 11 is obtained by fusing the position information of the GPS locator and the position information based on the inertial measurement unit, and the second positioning component 114 is set in the middle part of the landing mark in the landing mark image 113.

[0109] In this embodiment, the unmanned vehicle 11 is also provided with a processor, and the processor has an Apollo communication system and a ROS system built in; Apollo is used to obtain the position information of the second positioning member 114 in real time; Apollo uses the Cyber ​​RT communication mechanism to send the obtained position information of the second positioning member 114 to the ROS system (slave) via UDP transmission, and then the ROS system (slave) sends the obtained position information of the second positioning member 114 to the ROS system (host) via TCP transmission. The ROS system (host) is set on the processing device, and then the processing device obtains the position information of the second positioning member 114, which further enables the unmanned vehicle to transmit accurate positioning information to the drone, as well as the real-time positioning and collaborative transmission operations of the multi-robot system.

[0110] The technical features of the above-described embodiments may be arbitrarily combined, and the execution order of the method steps is not limited. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0111] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for autonomous landing of a drone, characterized in that: The autonomous landing method of the UAV comprises: The drone determines whether the current driving environment of the drone meets the conditions for obtaining the satellite position information of the landing location; If the current driving environment of the UAV meets the conditions for obtaining the satellite position information of the landing position, the UAV obtains the satellite position information of the landing position and performs the landing action according to the satellite position information of the landing position; If the current driving environment of the drone does not meet the conditions for obtaining the satellite position information of the landing position, the drone determines whether the current driving environment of the drone meets the conditions for obtaining the image position information of the landing position; If the current driving environment of the drone meets the conditions for obtaining the image position information of the landing position, the drone obtains the image position information of the landing position, and performs a landing action according to the image position information of the landing position; If the current driving environment of the drone does not meet the conditions for obtaining the image position information of the landing position, the drone will hover and wait, and return to the drone to determine whether the current driving environment of the drone meets the conditions for obtaining the satellite position information of the landing position.

2. The autonomous landing method of a drone according to claim 1, characterized in that: The image position information of the landing position includes a landing mark image.

3. The autonomous landing method of a drone according to claim 2, characterized in that: The image location information of the landing location obtained by the drone includes: Acquire landing mark image; Inputting the landing mark image into the optimization model, starting the optimization model, and obtaining the target landing mark image output by the optimization model; The target landing mark image is analyzed to obtain the position information of the landing mark in the target landing mark image.

4. The autonomous landing method of a drone according to claim 3, characterized in that: The satellite position information of the landing position includes satellite coordinate information of the landing position.

5. The autonomous landing method of a drone according to claim 4, characterized in that: If the current driving environment of the UAV satisfies the condition for obtaining the satellite position information of the landing position, the UAV obtains the satellite position information of the landing position, and performs a landing action according to the satellite position information of the landing position, including: The drone establishes a three-dimensional geographic coordinate system with the drone’s takeoff point as the origin; The drone receives the satellite position information of the landing location sent by the unmanned vehicle; The UAV converts the UAV position information and the satellite position information of the landing position into a three-dimensional geographic coordinate system to obtain the UAV position coordinates and the landing position coordinates; Inputting the position coordinates of the UAV and the position coordinates of the landing position into the extended Kalman filter prediction model, starting the extended Kalman filter prediction model, and obtaining the predicted position coordinates of the UAV and the predicted position coordinates of the landing position after a preset time output by the extended Kalman filter prediction model; Inputting the predicted position coordinates of the UAV and the predicted position coordinates of the landing position into the landing path planning model, starting the landing path planning model, and obtaining the predicted landing path output by the landing path planning model; The drone uses the predicted landing path to execute the landing action.

6. The autonomous landing method of a drone according to claim 5, characterized in that: If the current driving environment of the drone meets the condition for obtaining the image position information of the landing position, the drone obtains the image position information of the landing position, and performs a landing action according to the image position information of the landing position, including the steps of: The drone establishes a three-dimensional geographic coordinate system with the drone’s takeoff point as the origin; The drone obtains the location information of the landing mark; The drone analyzes the location information of the landing mark, and the drone obtains the coordinate information of the landing mark; The UAV converts the coordinate information of the UAV and the coordinate information of the landing mark into a three-dimensional geographic coordinate system to obtain the position coordinates of the UAV and the position coordinates of the landing position; Inputting the position coordinates of the UAV and the position coordinates of the landing position into the extended Kalman filter prediction model, starting the extended Kalman filter prediction model, and obtaining the predicted position coordinates of the UAV and the predicted position coordinates of the landing position after a preset time output by the extended Kalman filter prediction model; Input the predicted position coordinates of the drone and the predicted position coordinates of the landing position into the landing path planning model, start the landing path planning model, and obtain the predicted landing path output by the landing path planning model; The drone uses the predicted landing path to execute the landing action.

7. An autonomous landing system for a drone, characterized in that: The UAV autonomous landing system comprises: Autonomous vehicles; A drone, comprising a drone body and a processing device, wherein the processing device is used to execute the drone autonomous landing method as described in any one of claims 1 to 6 above.

8. The autonomous landing system for unmanned aerial vehicles according to claim 7, characterized in that: The drone also includes: A camera for acquiring images with landing markings; The first positioning member is used to obtain the position information of the drone body; The receiving device is used to receive the landing position information sent by the unmanned vehicle.

9. The autonomous landing system for unmanned aerial vehicles according to claim 8, characterized in that: The unmanned vehicle comprises: Unmanned vehicle body; A landing platform is arranged on the unmanned vehicle body; A landing sign image, attached to the top surface of the landing platform; The second positioning element is used to obtain the position information of the landing mark image.