Positioning landing method for manned aircraft

Abstract

The invention relates to the technical field of manned aircrafts, in particular to a positioning landing method for a manned aircraft. The method comprises the following steps: collecting image data of a landing area through a binocular camera, and extracting visual feature information of an obstacle; obtaining distance, speed and azimuth angle information of the obstacle through a millimeter wave radar; performing space-time alignment and feature level fusion on the visual data and the radar data to generate a dynamic obstacle map; and planning an obstacle avoidance path based on the dynamic obstacle map, adjusting the attitude and landing point of the aircraft, and ensuring safe landing. The method has the effect of improving ground obstacle recognition, can deal with severe weather and prevent safety accidents, improves the obstacle recognition accuracy by more than or equal to 30%, and supports complex weather (rain / fog) and night environments; the real-time correction response time of the landing trajectory is less than or equal to 50ms, and the collision risk is reduced; the system is compatible with an existing aircraft architecture, and hardware transformation cost is low.

Description

Technical Field

[0001] The present application relates to the technical field of manned aircraft, and in particular to a positioning and landing method for manned aircraft. Background Art

[0002] In recent years, with the development of the global economy and advancements in science and technology, the low-altitude economy has rapidly emerged as a new, integrated economic sector. This refers to an economic model that utilizes civilian manned and unmanned aircraft, operating in low-altitude airspace generally below 1,000 meters in vertical altitude, extending up to 3,000 meters as needed, and driving the integrated development of related sectors through diverse low-altitude flight activities, including passenger and cargo transport, and other operations. It boasts a wide reach, a long industrial chain, and strong growth and driving force. It has been widely applied in scenic tourism, urban security, medical care, emergency rescue, agricultural and forestry plant protection, and power inspection, and is becoming a significant driver of economic growth.

[0003] Currently, many countries around the world are actively promoting the development of the low-altitude industry, striving to seize this new development opportunity. China also attaches great importance to the development of the low-altitude economy and has already established a leading edge in some areas. For example, China is the world's largest producer of consumer drones, with DJI becoming a leading company in the consumer drone industry. A number of companies are accelerating the development of electric vertical take-off and landing aircraft (eVTOL) and actively promoting their pilot applications in scenic areas and other scenarios. Express delivery companies and terminal delivery companies, such as SF Express and Meituan, have launched multiple drone delivery routes in cities such as Shenzhen and Shanghai.

[0004] In the invention patent with authorization announcement number CN108750129B, the applicant Guangzhou Ehang Intelligent Technology Co., Ltd. adopted a pure visual solution when solving the problem of positioning and landing of manned aircraft. However, in actual use, when encountering some bad weather, the camera itself cannot clearly capture the obstacles under the manned aircraft, causing the manned aircraft to be unstable when landing, and even possible reversal of the manned aircraft, which in turn poses a safety hazard to the passengers in the manned aircraft. Summary of the Invention

[0005] In order to solve the technical problem of unsafe landing of existing manned aircraft, this application provides a solution based on pure vision plus millimeter wave radar.

[0006] The positioning landing method for manned spacecraft includes the following four steps: 1. Use binocular cameras to collect image data of the landing area and extract visual feature information of obstacles; 2. Obtain the distance, speed and azimuth angle information of obstacles through millimeter wave radar; 3. Perform spatio-temporal alignment and feature-level fusion on visual data and radar data to generate a dynamic obstacle map; 4. Plan an obstacle avoidance path based on the dynamic obstacle map, adjust the attitude and landing point of the aircraft to ensure a safe landing.

[0007] Furthermore, for spatio-temporal alignment, the gyroscope and accelerometer data of the inertial measurement unit are used to synchronize the sensor coordinate systems, with an error range ≤ 2 cm.

[0008] Furthermore, for feature-level fusion, an adaptive Kalman filtering algorithm is adopted, and the weight allocation is based on sensor confidence parameters, including the illumination intensity of the camera and the signal-to-noise ratio of the millimeter-wave radar.

[0009] Furthermore, the millimeter-wave radar operates in the frequency band of 76 - 81 GHz, with a detection accuracy ≤ 3 cm, a maximum detection distance of 100 m, and supports rain and fog penetration ability.

[0010] Furthermore, the binocular camera includes an infrared fill light module, supports obstacle recognition in night and low-light environments, and has a resolution ≥ 1080P.

[0011] Furthermore, the dynamic obstacle map adopts a probability grid model, fuses the millimeter-wave radar point cloud data and the visual SLAM results, and has an update frequency ≥ 20 Hz.

[0012] Furthermore, the controller calculates the optimal landing trajectory in real time based on the fused data, and uses the PID algorithm to dynamically adjust the rotor thrust and yaw angle, with a response time ≤ 50 ms.

[0013] Specifically, the binocular camera: captures the RGB image and depth information of the landing area, and identifies the obstacle contour and texture features; Specifically, the millimeter-wave radar: measures the distance, relative speed, and azimuth angle of the obstacle through FMCW frequency-modulated continuous wave, and operates stably through rain and fog.

[0014] Specifically, for the data fusion processing module, the spatio-temporal alignment unit: synchronizes the sensor coordinate systems based on IMU data (such as the data of the gyroscope and accelerometer); Specifically, the feature-level fusion algorithm: Matches the millimeter-wave radar point cloud data with the visual SLAM (simultaneous localization and mapping) results to generate a 3D obstacle probability grid map; Optimizes the obstacle positioning accuracy through dynamic weighted fusion by Kalman filtering (millimeter-level error ≤ 5 cm).

[0015] Specifically, the control decision-making module Path planning unit: calculates the optimal landing trajectory based on the fused map and avoids the safe area of obstacles; Attitude adjustment unit: Dynamically adjust the rotor thrust and yaw angle based on a PID controller to achieve a soft landing.

[0016] Beneficial effects: 1. The obstacle recognition accuracy is improved by ≥30%, supporting complex weather (rain / fog) and night environments; 2. The response time for real-time correction of the landing trajectory is ≤50 ms, reducing the collision risk; 3. The system is compatible with the existing aircraft architecture, and the hardware transformation cost is low. Description of the drawings

[0017] Figure 1 It is a flow block diagram of the present invention. Detailed implementation manners

[0018] A positioning and landing method for a manned aircraft includes the following four major steps: 1. Collect image data of the landing area through a binocular camera, and extract the visual feature information of obstacles; 2. Obtain the distance, speed, and azimuth angle information of obstacles through a millimeter-wave radar; 3. Perform spatio-temporal alignment and feature-level fusion on the visual data and radar data to generate a dynamic obstacle map; 4. Plan an obstacle avoidance path based on the dynamic obstacle map, adjust the aircraft attitude and landing point to ensure a safe landing.

[0019] Furthermore, for spatio-temporal alignment, the gyroscope and accelerometer data of the inertial measurement unit are used to synchronize the sensor coordinate systems, and the error range is ≤2 cm.

[0020] Furthermore, for feature-level fusion, an adaptive Kalman filtering algorithm is used, and the weight allocation is based on sensor confidence parameters, including the illumination intensity of the camera and the signal-to-noise ratio of the millimeter-wave radar.

[0021] Furthermore, the millimeter-wave radar operates in the frequency band of 76 - 81 GHz, the detection accuracy is ≤3 cm, the maximum detection distance is 100 m, and it supports rain and fog penetration ability.

[0022] Furthermore, the binocular camera includes an infrared supplementary lighting module, supports obstacle recognition in night and low-light environments, and the resolution is ≥1080P.

[0023] Furthermore, the dynamic obstacle map adopts a probability grid model, fuses the millimeter-wave radar point cloud data and the visual SLAM result, and the update frequency is ≥20 Hz.

[0024] Furthermore, the controller calculates the optimal landing trajectory in real time according to the fused data, and dynamically adjusts the rotor thrust and yaw angle using the PID algorithm, and the response time is ≤50 ms.

[0025] Specifically, a binocular camera: captures RGB images and depth information of the landing area, and identifies the contours and texture features of obstacles; Specifically, a millimeter-wave radar: measures the distance, relative speed, and azimuth angle of obstacles through FMCW frequency-modulated continuous wave, and works stably through rain and fog.

[0026] Specifically, a data fusion processing module, a spatio-temporal alignment unit: synchronizes the sensor coordinate systems based on IMU data (such as gyroscopes and accelerometers); Specifically, a feature-level fusion algorithm: matches the millimeter-wave radar point cloud data with the results of visual SLAM (simultaneous localization and mapping) to generate a 3D obstacle probability grid map; Through Kalman filter dynamic weighted fusion, optimize the obstacle positioning accuracy (millimeter-level error ≤ 5 cm).

[0027] Specifically, a control decision-making module: A path planning unit: calculates the optimal landing trajectory based on the fusion map and avoids the safe areas of obstacles; An attitude adjustment unit: dynamically adjusts the rotor thrust and yaw angle based on a PID controller to achieve a soft landing.

[0028] Technical key points description: 1. Sensor collaboration advantage: The all-weather detection ability of the millimeter-wave radar (penetrating rain and fog) complements the high-precision visual recognition of the camera, improving the reliability of obstacle detection.

[0029] 2. Dynamic path planning: Generate a probability map by fusing data, and optimize the landing trajectory in combination with the aircraft dynamics model to avoid static and dynamic obstacles.

[0030] 3. Redundancy and fault tolerance mechanism: Enhance the system robustness through sensor backup and ultrasonic near-ground detection to ensure safety under extreme conditions.

[0031] 4. Introduce an ultrasonic sensor as redundancy to make up for the risk of single sensor failure, referring to the fault tolerance design of medical detection equipment.

[0032] The above system also includes other components well-known to those skilled in the art such as communication buses and communication interfaces, and their settings and functions are known in the art, so they will not be elaborated here.

[0033] The above are all preferred embodiments of this application. The protection scope of this application is not limited accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of this application should be covered within the protection scope of this application.

Claims

1. A positioning and landing method for a manned aircraft, characterized in that Including: Collect image data of the landing area through a binocular camera and extract the visual feature information of obstacles; Obtain the distance, speed and azimuth angle information of obstacles through a millimeter-wave radar; Perform spatio-temporal alignment and feature-level fusion on the visual data and radar data to generate a dynamic obstacle map; Plan an obstacle avoidance path based on the dynamic obstacle map, adjust the attitude of the aircraft and the landing point to ensure a safe landing.

2. The positioning and landing method for a manned aircraft according to claim 1, characterized in that: The spatio-temporal alignment uses the gyroscope and accelerometer data of the inertial measurement unit to synchronize the sensor coordinate systems, and the error range is ≤2 cm.

3. The positioning and landing method for a manned aircraft according to claim 1, characterized in that: The feature-level fusion uses an adaptive Kalman filtering algorithm, and the weight allocation is based on the sensor confidence parameters, including the illumination intensity of the camera and the signal-to-noise ratio of the millimeter-wave radar.

4. The positioning and landing method for a manned aircraft according to claim 1, characterized in that: The millimeter-wave radar operates in the frequency band of 76 - 81 GHz, the detection accuracy is ≤3 cm, the maximum detection distance is 100 m, and it supports rain and fog penetration ability.

5. The positioning and landing method for a manned aircraft according to claim 1, characterized in that: The binocular camera includes an infrared supplementary light module, supports obstacle recognition in night and low-light environments, and the resolution is ≥1080P.

6. The positioning and landing method for a manned aircraft according to claim 1, characterized in that: The dynamic obstacle map uses a probability grid model, fuses the millimeter-wave radar point cloud data and the visual SLAM result, and the update frequency is ≥20 Hz.

7. The positioning and landing method for a manned aircraft according to claim 1, characterized in that: The controller calculates the optimal landing trajectory in real time according to the fused data, and uses the PID algorithm to dynamically adjust the rotor thrust and yaw angle, and the response time is ≤50 ms.

8. The positioning and landing method for a manned aircraft according to claim 1, characterized in that: It includes redundant design. When any sensor fails, it automatically switches to the data of the backup sensor and activates the ultrasonic near-ground detection module as a supplement.

9. The positioning and landing method for a manned aircraft according to claim 1, characterized in that: During the landing process, the driver is fed back with obstacle warning information through the multi-language interaction module, which supports voice and visual interface display.

10. The positioning and landing method for a manned aircraft according to any one of claims 1 to 9, characterized in that: For complex terrains, an altimeter and a lidar are integrated to assist in height calibration, and the vertical height error is ≤0.5 m.

Images