Aircraft maintenance intelligent following robot based on UWB and vision and design method

Abstract

The invention belongs to the field of intelligent following robots, and discloses an aircraft maintenance intelligent following robot based on UWB and vision and a design method thereof, a dual-mode following robot is designed and realized based on UWB, a depth camera and a target identification technology, so that the robot can automatically follow the aircraft maintenance under the condition of poor sight or poor communication environment. And objects such as wheel chokes, tools, equipment and aviation materials can be automatically and accurately carried along with maintenance personnel, and automation, intelligence and safety of aircraft maintenance operation are remarkably improved.

Description

Technical Field

[0001] The invention relates to the technical field of intelligent following robots, and in particular to an aircraft maintenance intelligent following robot based on UWB and vision and a design method thereof. Background Art

[0002] Aircraft maintenance is a key link in ensuring flight safety, and its complexity and professionalism are extremely high. With the increasing complexity of aircraft structures and the continuous increase in maintenance tasks, traditional manual maintenance methods can no longer meet the needs of the modern aviation industry. In recent years, with the rapid development of technologies such as robotics, artificial intelligence, and wireless communications, intelligent follower robots have shown great application potential in the fields of industry, medicine, and services. Applying these advanced technologies to the field of aircraft maintenance is expected to solve many of the current challenges.

[0003] According to the actual needs of civil aviation aircraft maintenance personnel, an industrial robot used for target positioning and following targets will have a wide range of application scenarios. At present, there are four main types of four-wheel robot chassis for autonomous operation: Ackerman steering chassis, four-wheel independent drive steering chassis, omni-directional wheel chassis and differential four-wheel chassis. Given the special nature of the outdoor ground at the airport, the four-wheel independent drive and steering chassis and the omni-directional wheel chassis may incur high maintenance costs due to wear problems during operation. At the same time, the Ackerman chassis is difficult to achieve flexible steering in the narrow space at the bottom of the aircraft due to its large turning radius. In contrast, the differential four-wheel chassis, with its ability to turn in place, is more suitable for operation in a flat and high-friction airport environment, and can more effectively meet the needs of airport ground handling work.

[0004] The target following algorithm exists in the following system as a core and important factor. The commonly used lightweight following algorithms mainly include those based on PDOA technology and vision technology. Among them, PDOA (Phase Difference of Arrival) is a positioning technology based on the phase difference of radio waves. It determines the position of the signal source by measuring the phase difference of the signal between different receiving points. PDOA technology is widely used in precise positioning and tracking in wireless communication systems such as ultra-wideband (UWB), Wi-Fi, and Bluetooth. Among them, Margiani et al. designed a UWB-PDOA module based on the DW3200 chip. The module can realize the positioning of the target with only a single chip. The device provides considerable support for the subsequent UWB-PDOA positioning technology. Huang Xin proposed an indoor positioning strategy based on a few base stations using UWB-PDOA technology, while Hong Xinlei used the arrival phase difference of multiple antennas to obtain an indoor positioning algorithm through coordinate transformation and calculation in three-dimensional space. Li Kaijun from Heilongjiang University used Kalman filtering to process the signals collected by the UWB-PDOA system to improve the positioning accuracy. The low-cost indoor positioning AGV designed and developed by Shi is also based on UWB-PDOA technology. Its main features are low power consumption and low cost. The AGV relies on multiple UWB base stations in the room to receive AGV location information, and uses the least squares method to process the data of multiple UWB base stations, thereby realizing high-precision positioning function in indoor environments.

[0005] However, in outdoor scenarios, due to the open environment and the low power limitations of the UWB signal generator, the UWB base station may experience signal loss when it is too far away from the target to be followed. This defect makes it challenging for robots that rely solely on UWB positioning technology to operate outdoors.

[0006] At the same time, the visual positioning technology based on target detection and laser positioning has performed well in target positioning applications. Chen Zhongcai used the Yolov5 neural network and Darknet framework to realize the auxiliary positioning of drones. Song Bowen used the Yolov5s algorithm to identify the iron wire used to bundle pulp bales. Combined with the robot kinematic analysis, the coordinate conversion between the camera and the robot was realized through hand-eye calibration, so as to obtain the three-dimensional information of the bundled wire using the depth camera. Azurmendi used the target detection algorithm combined with the ranging module to realize the positioning of the detection target. Francis combined the deep TOF laser camera with the tracking feature detection algorithm in the early stage, and realized the autonomous control of AGV to a certain extent. However, in low-visibility outdoor scenes such as early morning, night, foggy days, and rainy days, the performance of the visual module will be seriously affected. This defect will reduce the reliability of robots based on visual detection and positioning technology alone. Summary of the invention

[0007] In order to solve the above technical problems, the present invention provides an aircraft maintenance intelligent following robot based on UWB and vision and a design method. The present invention designs and implements a dual-modal following robot based on UWB, depth camera and target recognition technology, so that the robot can autonomously and accurately follow maintenance personnel to carry wheel chocks, tools, equipment, aviation materials and other items under poor visibility or communication environment, thereby significantly improving the automation, intelligence and safety of aircraft maintenance operations.

[0008] The present invention discloses an aircraft maintenance intelligent following robot based on UWB and vision, wherein the following robot comprises: a chassis and a box;

[0009] The chassis is used to move the aircraft maintenance follower robot behind the follower;

[0010] The box is used to transport aircraft wheel chocks, tools, equipment, and aviation materials.

[0011] Preferably, the chassis includes: four independent motor drive wheels, electric brakes, a lithium battery power supply, an operation indicator light and an anti-collision sensor device.

[0012] Preferably, the box is equipped with a UWB base station and a depth camera base for determining the position of the tracking target, and is also equipped with a base for installing laser and ultrasonic TOF sensors for detecting the robot's surroundings and realizing environmental perception function.

[0013] The present invention also discloses a design method of an aircraft maintenance intelligent follower robot based on UWB and vision, and the method is used to implement any one of the following robots, including:

[0014] Using UWB-PDOA algorithm, calculate the relative coordinates of the target and the base station;

[0015] Using the depth camera-target detection algorithm, calculate the coordinates of the target to be followed relative to the camera plane;

[0016] Using the dual-mode fusion algorithm, the relative coordinates of the target and the base station are fused with the coordinates of the follow-up target relative to the camera plane, and the coordinates of the follow-up target relative to the XY plane of the car are calculated;

[0017] According to the XY plane coordinates of the tracking target relative to the car, the linear speed and angular speed required for the robot to reach the target position are calculated to achieve intelligent assisted transportation.

[0018] Preferably, the design method also includes: selecting an obstacle avoidance system that combines ultrasonic and laser TOF sensors to ensure that the follower robot automatically stops when encountering an obstacle to ensure safety.

[0019] Preferably, using the UWB-PDOA algorithm, calculating the relative coordinates of the target and the base station includes:

[0020] d1=t a1 *v,

[0021] d2=t a2 *v,

[0022] d1′=d1*cosα,

[0023] d2′=d2*cosβ,

[0024]

[0025] Among them, t a1 ,t a2 is the time it takes for the Bluetooth tag signal to reach the antenna, v is the signal speed, d1 and d2 are the distances between the tag and the two base station antennas, α and β are the signal incident angles, d1′ and d2′ are the projection distances between the tag and the two base station antennas in the plane rectangular coordinate system, (X t1 ,Y t1 ) is the coordinate of antenna Antenna1, (X t2 ,Y t2 ) is the coordinate of antenna Antenna2, (X t ,Y t ) is the projection coordinate of the label in the XY plane.

[0026] Preferably, using a depth camera-target detection algorithm, calculating the coordinates of the tracking target relative to the camera plane includes:

[0027]

[0028] Where D is the distance between the target and the camera, w is the angle between the target center and the depth camera relative to the vertical Y axis of the photo, L is the actual distance from the camera to the center of the target, (X a ,Y a ) is the coordinate of the target coordinate relative to the camera plane, and (X′, Z′) is the mapping coordinate of the target coordinate relative to the center of the RGB photo.

[0029] Preferably, the relative coordinates of the target and the base station are fused with the coordinates of the follow-up target relative to the camera plane using a dual-mode fusion algorithm, and the calculation of the follow-up target relative to the XY plane coordinates of the vehicle includes:

[0030] X c =k i *X a +k j *X t ,

[0031] Y c =k i *Y a +k j *Y t ,

[0032] k i +k j =1,

[0033] Among them, (X c ,Y c ) is the XY plane coordinate of the tracking target relative to the car calculated by the dual-mode algorithm, k i and k j is the weight parameter.

[0034] Preferably, according to the XY plane coordinates of the tracking target relative to the vehicle, calculating the linear velocity and angular velocity required for the robot to reach the target position includes:

[0035]

[0036] Among them, T a is the time required for the robot to reach the target position from the current position, V x is the linear velocity required for the robot to reach the target position, V r is the angular velocity required for the robot to reach the target position.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] Based on UWB, depth camera and target recognition technology, the present invention designs and implements a dual-mode following robot, so that the robot can autonomously and accurately follow maintenance personnel to carry wheel chocks, tools, equipment, aviation materials and other items under poor visibility or communication environment, significantly improving the automation, intelligence and safety of aircraft maintenance operations.

[0039] The present invention introduces two algorithms, UWB-PDOA and depth camera-visual positioning, and proposes an algorithm that utilizes signal continuity and image confidence to complement each other, integrating the UWB-PDOA algorithm and the depth camera-visual positioning algorithm to make the positioning algorithm more adaptable to the environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0041] Figure 1 A schematic diagram of a box according to an embodiment of the present invention;

[0042] Figure 2 A physical picture of a UWB base station according to an embodiment of the present invention;

[0043] Figure 3 Schematic diagram of the PDOA algorithm according to an embodiment of the present invention;

[0044] Figure 4 A schematic diagram of a target detection algorithm according to an embodiment of the present invention;

[0045] Figure 5 A schematic diagram of calculating the actual coordinates of a follow-up target according to an embodiment of the present invention;

[0046] Figure 6 is a weight parameter function diagram of an embodiment of the present invention;

[0047] Figure 7 Schematic diagram of the detection accuracy of the same target in three following modes in different environments according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0049] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present disclosure should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the embodiments of the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0050] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0051] Embodiment 1

[0052] As can be seen from the background technology, with the continuous growth of the global civil aviation industry, aircraft route maintenance and overhaul work are facing unprecedented challenges. The traditional manual maintenance method is not only inefficient, but also easily affected by human factors and has potential safety hazards.

[0053] The embodiment of the present invention provides an aircraft maintenance intelligent follower robot based on UWB and vision. As an intelligent follower robot that needs to assist in handling work in the outdoor environment of the airport, the physical structure of the intelligent follower robot needs to have a box for placing and transporting goods. The robot chassis needs to have a certain load-bearing capacity, can operate stably on a high-friction and flat airport road surface, and support flexible steering at the bottom of the aircraft wing. Functionally, the dual-mode intelligent follower robot needs to show high target detection accuracy in outdoor environments with low signal strength or poor line of sight conditions, and can ensure the function of accelerating to approach the target when the distance to the target is far, and decelerating to wait for the target when the distance to the target is close. At the same time, as an intelligent robot working in the airport, the robot needs to have strict environmental perception capabilities to avoid obstacles, thereby ensuring the safety of personnel and equipment.

[0054] In terms of equipment, the image processing system based on the vision module needs to use an image processing chip with sufficient AI computing power, and the main control needs to have sufficient communication interfaces to perceive the surrounding environment information of the vehicle and support the ROS operating system.

[0055] According to the above analysis, the follower robot structure needs to have carrying and load-bearing capabilities. Therefore, the follower robot structure of the present invention consists of two parts: a chassis and a box. The chassis is used to move the aircraft maintenance follower robot behind the follower, and the box is used to transport items such as aircraft wheel chocks, tools, equipment, and aviation materials.

[0056] In this embodiment, the chassis designed by the present invention includes four independent motor drive wheels, electric brakes, lithium battery power supply, operation indicator lights and anti-collision sensor devices. Due to the large friction coefficient of the airport road surface, the following task often needs to be implemented under the wings of the aircraft, so a differential four-wheel system that can turn on the spot is used as the drive, and an emergency electric brake is equipped to prevent collisions. The basic parameters of the chassis are shown in Table 1:

[0057] Table 1 Chassis parameters of follower robot

[0058]

[0059]

[0060] In this embodiment, the box designed by the present invention is as follows Figure 1As shown, its structural size is 1000*750*200. It is equipped with a UWB base station and a depth camera base for determining the position of the target to be followed. In addition, a base is left for the installation of sensors. Laser and ultrasonic TOF sensors are also installed to detect the robot's surroundings and realize environmental perception function.

[0061] The technical solution of the present invention, based on UWB, depth camera and target recognition technology, designs and implements a dual-modal following robot, so that the robot can autonomously and accurately follow maintenance personnel to carry wheel chocks, tools, equipment, aviation materials and other items under poor visibility or communication environment, thereby significantly improving the automation, intelligence and safety of aircraft maintenance operations.

[0062] It should be noted that some embodiments of the present disclosure are described above. Other embodiments are within the scope of the attached claims. In some cases, it should be understood that the size of the serial number of each step in the above embodiment does not mean the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention. The actions or steps recorded in the claims can be performed in an order different from that in the above embodiment and still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0063] Embodiment 2

[0064] Based on the same inventive concept, corresponding to any of the above embodiments, the present invention also provides a design method for an aircraft maintenance intelligent following robot based on UWB and vision, the method is used to implement any of the following robots, according to demand analysis, the following robot needs to show high target detection accuracy in outdoor environments with low signal strength or poor line of sight conditions, so the present invention combines the advantages of UWB technology and depth camera-target detection technology, and complements the disadvantages of the two technologies, and designs an intelligent robot following algorithm that integrates UWB-PDOA and depth camera-target detection dual modes. Specifically:

[0065] Using UWB-PDOA algorithm, calculate the relative coordinates of the target and the base station;

[0066] Using the depth camera-target detection algorithm, calculate the coordinates of the target to be followed relative to the camera plane;

[0067] Using the dual-mode fusion algorithm, the relative coordinates of the target and the base station are fused with the coordinates of the follow-up target relative to the camera plane, and the coordinates of the follow-up target relative to the XY plane of the car are calculated;

[0068] According to the XY plane coordinates of the tracking target relative to the car, the linear speed and angular speed required for the robot to reach the target position are calculated to achieve intelligent assisted transportation.

[0069] In this embodiment, UWB technology uses impulse pulse signals and has extremely high signal time resolution, which can provide accurate target positioning services to the following system even in low visibility environments. Therefore, the present invention uses a dual-antenna UWB base station and tag (with low power consumption, strong anti-interference ability and high positioning accuracy) with low power consumption and strong anti-interference ability. Figure 2 ) is used as one of the positioning modules, and the PDOA algorithm is used to calculate the relative coordinates of the target and the base station.

[0070] like Figure 3 As shown in the figure, the PDOA algorithm transmits data by sending and receiving extremely narrow pulses. It relies on measuring the phase difference between the signals arriving at different receivers, calculates the time when the signals arrive at the two base station antennas, and calculates the distances d1 and d2 between the tag and the two base station antennas by the time when the signals arrive at the base station antennas and the signal transmission rate. Then, according to the geometric relationship between d1, d2 and the signal incident angles α and β, the projection distances d1′ and d2′ of the tag relative to the base station in the XY plane are calculated.

[0071] The coordinates of antenna Antenna1 are (X t1 ,Y t1 ), the coordinates of antenna Antenna2 are (X t2 ,Y t2 ), Bluetooth tag signal reaches the antenna time t a1 ,t a2 , time multiplied by the signal speed v, we get the distances d1 and d2 between the tag and the two base station antennas. Then, combined with the signal incident angles α and β, we calculate the projection distances d′1 and d′2 between the tag and the two base station antennas in the plane rectangular coordinate system. Finally, we list the equations and solve the projection coordinates (X t ,Y t ).

[0072] d1=t a1 *v Formula (1)

[0073] d2=t a2 *v Formula (2)

[0074] d1′=d1*cosα Formula (3)

[0075] d2′=d2*cosβ Formula (4)

[0076]

[0077] It is worth noting that the coordinates (Xt ,Y t ) will be passed into the dual-modal fusion algorithm as part of the target coordinates.

[0078] In this embodiment, the depth camera-target detection algorithm relies on a convolutional neural network similar to the human visual system for target detection, while the depth camera uses laser TOF technology to detect the actual distance of all targets in the field of view from the camera. After aligning the two images, high-precision target positioning can be achieved in a good visual environment.

[0079] The pictures taken by the RGB camera are processed as follows Figure 4 The pre-trained YOLOV5 neural network shown outputs a target detection result image with object labels and the mapped coordinates of the object relative to the center of the RGB photo.

[0080] The depth camera obtains point cloud data through laser TOF, where each pixel of the depth image represents the actual distance between the target shown by the current pixel and the camera. After aligning the depth image with the image output by the target detection task, the distance between the target and the depth camera can be obtained. In order to simplify the calculation process and reduce computing power, we use the geometric center of the target as the reference point.

[0081] like Figure 5 As shown, the coordinates of the following target relative to the camera plane are (X a ,Y a ,Z a ), the distance between the target and the camera is D.

[0082]

[0083] It is known that in the captured RGB picture, the mapping coordinates of the follow target relative to the center of the RGB photo are (X′ a ,Z′ a ), the horizontal FOV angle of the camera is θx, and the horizontal pixel of the camera is X′ max , the vertical pixel is Z′ max Therefore, the angle w between the center of the tracking target and the depth camera relative to the vertical Y axis of the photo and the actual distance L from the camera to the center of the tracking target satisfy:

[0084] L=D*cosw Formula (6)

[0085]

[0086] It can be solved as:

[0087]

[0088] It is worth noting that the coordinates (Xa ,Y a ) will also be passed into the dual-modal fusion algorithm as part of the target coordinates.

[0089] In this embodiment, in order to achieve accurate target position determination in various scenarios, we merge the coordinates obtained by the UWB-PDOA algorithm and the depth camera-target detection algorithm to obtain the XY plane coordinates (X c ,Y c ), which satisfies:

[0090] X c =k i *X a +k j *X t Formula (10)

[0091] Y c =k i *Y a + k j *Y t Formula (11)

[0092] k i +k j =1 Formula (12)

[0093] where k i and k j is the weight parameter, which is the average confidence parameter v of the depth camera image i and UWB signal continuity parameter v u The calculation formula satisfies:

[0094]

[0095] Where m is the number of picture frames m = 1, 2, 3..., when the average confidence value is greater than the confidence threshold v θ When i is the average confidence value of the n frames of images calculated, which reflects the reliability of the target detection system. vu is the number of signal packets actually received by the sensor (packets / second) a and the maximum number of supported received signal packets u max Division calculation, which reflects the coherence of the signal system.

[0096] Weight parameter k i , k j By Figure 6It is calculated by the sigmoid-like function, which has the characteristics of strong correction when one of the UWB-PDOA and depth camera-target detection dual systems is interfered with, so that the system can minimize the impact of the environment on positioning accuracy and improve anti-interference ability.

[0097] The system then calculates the coordinates (X c ,Y c ), calculate the linear velocity V required for the robot to reach the target position x , angular velocity V r ,in:

[0098]

[0099] Where T a It is the time required for the robot to reach the target position from the current position. Taking time as a quantitative value can ensure that the robot accelerates at long distances and decelerates at close distances, thereby improving the overall equipment safety performance. K is the turning angle coefficient used to optimize the robot's angular velocity. A suitable coefficient can make the robot more agile and smoother during the turning process.

[0100] As a four-wheel differential robot, the robot linear speed V x , angular velocity V r , wheelbase L, wheel radius R and left and right wheel speeds V1 and V2 satisfy:

[0101]

[0102] In addition, to ensure safety, emergency braking condition judgment is added to the control algorithm to ensure that the following system can max It will automatically stop when the aircraft is in flight, ensuring the safety of personnel, aircraft and robots.

[0103]

[0104] In this embodiment, in an airport environment, especially in areas with dense personnel and equipment, the handling tool needs to operate efficiently and safely in a limited space. Therefore, the obstacle avoidance module is crucial. The present invention uses an obstacle avoidance system that combines ultrasonic and laser TOF sensors to ensure that the follower robot will automatically stop when encountering an obstacle to ensure safety.

[0105] Embodiment 3

[0106] In order to verify the superiority of UWB and visual dual-modal system in multi-environment following applications, the present invention uses a physical prototype to conduct comparative tests of single machine vision module positioning and following targets, single UWB module positioning and following targets, and UWB-PDOA and depth camera-target detection dual-modal positioning and following targets in two environments: daytime and nighttime. The percentage errors of the test results are shown in Table 2 and Figure 7 shown.

[0107] Table 2 Detection accuracy of the same target in three following modes in different environments

[0108] Time(Mode) / Distance 1～10(M) 10～20(M) 20～35(M) Daytime (machine vision) ±1.2％ ±4.3％ ±13.6％ Daytime (UWB positioning) ±0.9％ ±7.1％ ±23.2％ Daytime (dual mode) ±1.1％ ±5.3％ ±14.3％ Night (machine vision) ±6.6％ ±25.3％ NA Night (UWB positioning) ±0.9％ ±7.1％ ±23.2％ Night (dual mode) ±2.5％ ±7.6％ ±23.2％

[0109] According to the test results, it can be seen that in a stable light environment, the positioning error of the machine vision module changes smoothly at different distances and is relatively stable. However, the positioning error of the UWB module increases significantly at long distances.

[0110] In an environment with large changes in light conditions, the error of the machine vision module increases significantly, indicating that machine vision can hardly provide effective positioning data in a night environment. The UWB module is not affected by light and still maintains good accuracy at short and medium distances.

[0111] The dual-modal system shows similar accuracy to machine vision in stable light environments, and its performance at medium and long distances is significantly better than that of a single machine vision module in environments with large light changes, and its accuracy is similar to that of the UWB module. Taking all factors into consideration, the dual-modal system can significantly improve the reliability of tracking and positioning in night environments by combining the two technologies. This shows that the dual-modal system can combine the advantages of UWB and vision in a variety of environments, thereby providing stable and low errors in various distance ranges.

[0112] The test results show that regardless of the lighting conditions, the dual-modal system combining UWB and vision shows superior performance compared to the single-modal system. Especially in environments with large changes in lighting conditions (such as at night), the dual-modal system can effectively make up for the shortcomings of machine vision and achieve higher tracking accuracy and stability.

[0113] The present invention has developed a dual-mode civil aircraft maintenance intelligent following robot according to the lighting and signal conditions in different airport environments, which can assist aircraft maintenance personnel in carrying items and improve work efficiency. The specific work done by the present invention includes the following parts.

[0114] (1) Review domestic and foreign literature on target positioning, understand the current research status of UWB-PDOA and vision / LiDAR assisted positioning, analyze the pros and cons of various positioning methods through actual application scenarios, and propose a fusion positioning method that combines UWB-PDOA and vision / LiDAR positioning dual modes.

[0115] (2) The present invention analyzes the actual needs of airport ground staff and designs the robot's body and chassis separately. The chassis is equipped with an N8250E-HD industrial computer for controlling the motor drive, a REALSENSE D345 depth camera for visual positioning, and a UWB base station for PDOA positioning.

[0116] (3) The present invention introduces two algorithms, UWB-PDOA and depth camera-visual positioning, and proposes an algorithm that uses signal continuity and image confidence to complement each other, integrating the UWB-PDOA algorithm and the depth camera-visual positioning algorithm, so that the positioning algorithm has stronger environmental adaptability.

[0117] (4) The present invention conducted comparative tests on target tracking using a single machine vision module, target tracking using a single UWB module, and target tracking using dual-modality positioning using UWB-PDOA and depth camera-target detection in both daytime and nighttime environments. By comparing the positioning accuracy of the three at different distances, it was concluded that the dual-modality tracking and positioning system has higher accuracy and better environmental adaptability.

[0118] The embodiments of the present invention are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention should be included in the scope of protection of this disclosure.

Claims

1. Aircraft maintenance intelligent following robot based on UWB and vision, characterized by: The following robot comprises: a chassis and a box body; The chassis is used to move the aircraft maintenance follower robot behind the follower; The box is used to transport aircraft wheel chocks, tools, equipment, and aviation materials.

2. The following robot according to claim 1, characterized in that: The chassis includes: four independent motor drive wheels, electric brakes, a lithium battery power supply, an operating indicator light and an anti-collision sensor device.

3. The following robot according to claim 1, characterized in that: The box is equipped with a UWB base station and a depth camera base for determining the position of the target to be followed. It is also equipped with a base for installing laser and ultrasonic TOF sensors for detecting the robot's surroundings and realizing environmental perception function.

4. A design method for an aircraft maintenance intelligent following robot based on UWB and vision, the method being used to implement the following robot according to any one of claims 1 to 3, characterized in that: include: Using UWB-PDOA algorithm, calculate the relative coordinates of the target and the base station; Using the depth camera-target detection algorithm, calculate the coordinates of the target to be followed relative to the camera plane; Using the dual-mode fusion algorithm, the relative coordinates of the target and the base station are fused with the coordinates of the follow-up target relative to the camera plane, and the coordinates of the follow-up target relative to the XY plane of the car are calculated; According to the XY plane coordinates of the tracking target relative to the car, the linear speed and angular speed required for the robot to reach the target position are calculated to achieve intelligent assisted transportation.

5. The design method according to claim 4, characterized in that: The design method also includes: selecting an obstacle avoidance system that combines ultrasonic and laser TOF sensors to ensure that the following robot will automatically stop when encountering an obstacle to ensure safety.

6. The design method according to claim 4, characterized in that: Using the UWB-PDOA algorithm, the relative coordinates of the target and the base station are calculated including: d1=t a1 *v, d2=t a2 *v, d1′=d1*cosα, d2′=d2*cosβ, Among them, t a1 ,t a2 is the time it takes for the Bluetooth tag signal to reach the antenna, v is the signal speed, d1 and d2 are the distances between the tag and the two base station antennas, α and β are the signal incident angles, d′1 and d′2 are the projection distances between the tag and the two base station antennas in the plane rectangular coordinate system, (X t1 , Y t1 ) is the coordinate of antenna Antenna1, (X t2 ,Y t2 ) is the coordinate of antenna Antenna2, (X t , Y t ) is the projection coordinate of the label in the XY plane.

7. The design method according to claim 6, characterized in that: Using the depth camera-target detection algorithm, the coordinates of the target to be followed relative to the camera plane are calculated, including: Where D is the distance between the target and the camera, w is the angle between the target center and the depth camera relative to the vertical Y axis of the photo, L is the actual distance from the camera to the center of the target, (X a , Y a ) is the coordinate of the target coordinate relative to the camera plane, and (X′, Z′) is the mapping coordinate of the target coordinate relative to the center of the RGB photo.

8. The design method according to claim 7, characterized in that: Using the dual-mode fusion algorithm, the relative coordinates of the target and the base station are fused with the coordinates of the follow-up target relative to the camera plane. The calculation of the follow-up target relative to the XY plane coordinates of the car includes: X c =k i *X a +k j *X t , AND c =k i *AND a +k j *AND t , k i +k j =1, Among them, (X c , Y c ) is the XY plane coordinate of the tracking target relative to the car calculated by the dual-mode algorithm, k i and k j is the weight parameter.

9. The design method according to claim 8, characterized in that: According to the XY plane coordinates of the tracking target relative to the robot, the linear velocity and angular velocity required for the robot to reach the target position are calculated: Among them, T a is the time required for the robot to reach the target position from the current position, V x is the linear velocity required for the robot to reach the target position, V r is the angular velocity required for the robot to reach the target position.

Images