Emergency intelligent first-aid repair equipment for fault of main airplane wheel of airplane

Abstract

The invention discloses intelligent emergency repair equipment for faults of a main airplane wheel of an airplane, and belongs to the technical field of aviation ground service support equipment. The equipment mainly comprises an intelligent detection system, a bearing and lifting system, a clamping and lifting control system, an overall control system, a control system, a power supply system and a power system. The intelligent detection system integrates a depth camera and a lightweight YOLOv8 target detection model, and can automatically identify the model and the damage state of an airplane wheel and position a three-dimensional coordinate. The clamping and lifting control system achieves synchronous precise control over the double electric cylinders based on a PID cross coupling algorithm. By means of intelligent sensing, self-adaptive clamping and lifting and closed-loop control, full-automatic recognition, positioning, butt joint and lifting of the airplane wheel with the fault of an unknown model are achieved, the problems that existing equipment is poor in universality, depends on manpower and is low in efficiency are effectively solved, the airplane wheel can serve as a temporary airplane wheel to be transferred along with an airplane after first-aid repair, and the service life of the airplane wheel is prolonged. And the automation level and the safety of emergency repair are obviously improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of aviation ground support equipment, and particularly relates to an emergency intelligent repair equipment for main wheel failure of an airplane. BACKGROUND

[0002] With the rapid development of general aviation industry, the airport operation density and flight frequency continue to grow, and higher requirements are put forward for the timeliness, reliability and intelligent level of aviation ground support. As a key load-bearing component in the process of take-off and landing and taxiing of an airplane, the main wheel of the airplane is prone to failure such as tire burst and structural damage, which is a common emergency in aviation operation. Once such a failure occurs, the faulty airplane will be stranded on the runway or taxiway, which not only affects its own safety, but also may cause a chain reaction of large-scale flight delays, runway closure and other chain reactions, causing huge economic losses and safety hazards. Therefore, it is of great importance to develop efficient and reliable emergency repair equipment for main wheel failure of an airplane to realize rapid response and disposal for the normal operation of the airport.

[0003] At present, the airport mainly relies on special lifting equipment or general lifting equipment for emergency repair of main wheel failure of an airplane (especially tire burst). The existing technical solutions generally have the following outstanding problems and limitations: 1. Poor equipment versatility, limited adaptation to aircraft models: The existing repair equipment is usually designed for the main landing gear structure and wheel size of a specific aircraft model, and can usually only adapt to 1-2 aircraft models. However, modern airports, especially general aviation airports, often operate a variety of small aircraft of different models. This "one machine one type" equipment configuration mode results in the need for the airport to store multiple sets of different emergency equipment, which not only has high purchase cost and occupies a large amount of storage space, but also has complex equipment management, and may delay the best repair opportunity due to improper allocation in an emergency.

[0004] 2. Low automation level, heavily dependent on manual experience: The existing equipment almost completely relies on visual judgment and manual control of the operator in the operation process, from equipment positioning, interfacing with the faulty wheel, to clamping and lifting. It is extremely difficult to perform accurate positioning operations at the bottom of the narrow aircraft fuselage, resulting in long average interfacing time and low repair efficiency. This process requires high professional skills and on-site experience of the operator, and experienced technicians are scarce. In low-visibility or adverse weather conditions such as night, rain, snow and fog, the difficulty and risk of manual operation are multiplied, and it is easy to cause collision between the equipment and the aircraft structure due to inaccurate positioning and operation errors, causing secondary damage.

[0005] 3. Limited Functionality and Lack of Intelligent Detection and Decision Support: Existing equipment is essentially a mechanical lifting tool, lacking the ability to automatically identify, locate, and assess the condition of faulty wheels. It cannot automatically identify different wheel models, determine the type of damage (such as tire blowouts, tears, or wheel detachment), or provide optimal positional decision support for clamping and lifting operations. The entire repair process lacks data support and intelligent guidance, essentially operating in a "blind" manner, further exacerbating the reliance on manual labor and the uncertainty of the operational outcome.

[0006] 4. Insufficient integration and mobility: Many existing devices lack independent power and require towing by other vehicles to position them, increasing coordination steps and preparation time. Furthermore, the equipment has limited functionality; moving a malfunctioning aircraft after lifting still relies on other transfer equipment, making it impossible to achieve integrated "lifting-replacement-transfer" operations and prolonging the aircraft's time at the accident site.

[0007] In summary, existing aircraft main wheel repair equipment has significant shortcomings in terms of versatility, automation, intelligence, and overall efficiency, making it difficult to meet the urgent needs of modern aviation operations for "fast, safe, efficient, and universal" emergency support. Therefore, developing a small, intelligent emergency repair device for aircraft main wheel failures that can autonomously identify faulty wheels, adapt to different aircraft models, is easy to operate, and has a high degree of integration is of paramount practical significance and urgency. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings of the prior art and provide an intelligent emergency repair device for aircraft main engine wheel failure.

[0009] The objective of this invention is achieved through the following technical solution: an intelligent emergency repair device for aircraft main wheel failure, comprising an overall control system, wherein the overall control system is connected to a power system, an operating system, a power supply system, an intelligent detection system, and a clamping and lifting control system, wherein the clamping and lifting control system is connected to a load-bearing and lifting system; The overall control system uses a programmable logic controller as the main control unit, and communicates with each subsystem through the controller local area network bus to coordinate and control each subsystem to complete the emergency repair operation. The power system is used to provide the overall movement and mobility of the equipment; The control system is used to send manual intervention commands to the overall control system. It has two control modes: wireless remote control and wired backup, ensuring that emergency repairs can still be completed manually when the intelligent detection system fails. The intelligent detection system is used to collect visual information of the faulty wheel, and to process the visual information based on a deep learning model to identify the characteristics of the faulty wheel, so as to output the wheel model, damage status and three-dimensional positioning coordinates in space. The clamping control system is used for synchronous motion control of the double electric cylinders arranged bilaterally symmetrically along the feeding direction by adopting a cross-coupling control strategy based on a PID controller according to the optimal clamping position instruction output by the overall control system, so as to drive the bearing and lifting system to complete the clamping and lifting action. The bearing and lifting system comprises a vehicle body bearing structure and a lifting structure, and is used for clamping and lifting the single-side fault aircraft wheel and bearing the entire load of the side main landing gear; the lifting structure is provided with an adjustable wheel clamping mechanism to adapt to wheels of different sizes and can actively compensate for the height difference caused by wheel explosion.

[0010] Preferably, the intelligent detection system comprises a depth camera and a deep learning processing module. The depth camera is used for synchronously collecting color images and depth images, and acquiring the coordinates of the pixel points in the image in the world coordinate system according to the internal and external parameters of the camera. The deep learning processing module is internally provided with a lightweight YOLOV8 aircraft fault main wheel target detection model; the lightweight YOLOV8 aircraft fault main wheel target detection model takes the YOLOV8 model as a basic model, adopts a MobileNetV3-large network fused with an ECA-Net attention mechanism as a backbone feature extraction network of the YOLOV8 model, replaces the PANet neck of the YOLOv8 model with a deep separable convolution version of BiFPN, introduces learnable weight parameters in the feature fusion path in the BiFPN, and introduces a multi-scale hollow convolution module and a parallel key point detection branch in the head network layer, which are used for positioning the wheel grounding center and the hub center and providing geometric constraints for optimal clamping position calculation. The MobileNetV3-large network integrated with the ECA-Net attention mechanism is adopted as the backbone feature extraction network of the YOLOv8 model, the PANet neck of the YOLOv8 model is replaced with the deep separable convolution version of BiFPN, and the multi-scale hollow convolution module and the parallel key point detection branch are introduced in the head network layer, which are used for identifying the model, damage type and optimal clamping position of the aircraft main wheel.

[0011] Preferably, the depth camera comprises an RGB camera, an infrared camera and an infrared projector.

[0012] Preferably, the training and deployment of the lightweight YOLOV8 aircraft fault main wheel target detection model comprises the following steps: A data set construction stage, in which aircraft wheel images under multiple typical working conditions are collected, and the data set is divided into a training set, a verification set and a test set according to a preset ratio by adopting a hierarchical sampling method; In the model training stage, the lightweight YOLOV8 aircraft fault host wheel target detection model is pre-trained on a large general dataset, and then the model is trained on a self-constructed aircraft wheel dataset to observe whether the model result meets the expected requirements, and the model is fine-tuned and optimized for the target detection task. In the model optimization and deployment stage, the model that meets the training standard is subjected to model pruning based on prior knowledge and adaptive multi-resolution inference strategy, and the optimized model is deployed to the intelligent detection system to realize real-time high-performance detection.

[0013] Preferably, the aircraft wheel clamping unit is provided with an adaptive contact surface, which can automatically adjust the clamping angle according to the type of aircraft fault host wheel detected by the intelligent detection system. The lifting structure adopts a multi-stage telescopic electric cylinder structure.

[0014] Preferably, when detecting and positioning the aircraft fault host wheel, the following steps are included: Based on the imaging principle and internal and external parameters of the depth camera, the conversion relationship between the camera coordinate system and the world coordinate system is derived; The intelligent repair equipment is driven to reach the vicinity of the fault host wheel and adjust the position of the depth camera, and then color images and depth images of the fault host wheel are obtained; The color images of the aircraft fault host wheel obtained are input into the deployed lightweight YOLOV8 aircraft fault host wheel target detection model to obtain the pixel coordinates of the center positioning point of the aircraft fault host wheel; According to the pixel coordinates of the center positioning point of the aircraft fault host wheel and the depth information of the aircraft fault host wheel obtained by the depth camera, the position of the aircraft fault host wheel in the world coordinate system and the relative distance from the depth camera are calculated; The three-dimensional coordinates and relative distance information of the detected aircraft fault host wheel are fed back to the overall control system for subsequent navigation and docking control.

[0015] Preferably, when repairing the aircraft fault host wheel, the following steps are included: Receiving dispatching instructions from the tower and ground crew; The overall control system controls the intelligent repair equipment to autonomously navigate to the vicinity of the aircraft fault host wheel according to the dispatching instructions; After reaching the vicinity of the fault host wheel, the intelligent detection system adjusts the detection position to obtain the three-dimensional coordinates and relative distance information of the aircraft fault host wheel and feeds them back to the overall control system; The overall control system calculates the optimal clamping position according to the three-dimensional coordinates and relative distance information and outputs it to the clamping control system; The clamping and lifting control system controls the bearing and lifting system to clamp and lift the fault host wheel according to the received optimal clamping position, in the process, the clamping and lifting control system collects the actual lifting position of the bearing and lifting system in real time and compares it with the expected lifting position, and controls the bearing and lifting system to move towards the expected lifting position based on the cross-coupling control strategy of the PID controller; After the clamping and lifting are completed, the intelligent rescue equipment replaces the fault host wheel as a temporary wheel, moves synchronously with the fault aircraft under the dragging of the tractor, and completes the rescue operation.

[0016] The beneficial effects of the present application are: 1) Strong universality, realizing self-adaptive rescue of fault host wheels of various aircraft models: the present application can automatically identify fault host wheels of different aircraft models and accurately locate the optimal clamping point by combining deep camera vision perception with a lightweight deep learning target detection model. At the same time, the bearing and lifting system adopts symmetric clamping arms with adaptive contact surfaces on the inner side, which can automatically adjust the clamping angle and opening and closing range according to the identified aircraft model. This dual protection of "intelligent identification + mechanical adaptation" enables a single device to cover most general-purpose small aircraft models in the airport, completely solving the problem of poor universality of existing devices and the need to configure multiple sets of equipment, greatly reducing the equipment purchase cost and management complexity of the airport.

[0017] 2) High degree of automation and intelligence, greatly reducing the dependence on operator experience and improving rescue efficiency: the present application builds a complete automation process from "perception - decision - execution". The device can autonomously navigate to the fault point according to the dispatch instruction, automatically complete the identification, positioning and damage discrimination of the host wheel through the intelligent detection system, and automatically plan the optimal docking path and clamping parameters by the overall control system, finally realize precise and smooth automatic clamping and lifting through the clamping and lifting control system. This "one-key" operation process changes the traditional process of relying on manual visual inspection and manual operation, which is time-consuming and long, into fast and accurate automatic operation, greatly shortens the average docking and rescue time, reduces the harsh requirements on operator skills, and significantly improves the emergency response efficiency of the airport.

[0018] 3) Outstanding safety and reliability, effectively avoiding secondary damage and ensuring human-machine safety: the clamping and lifting control system uses a cross-coupling control strategy based on PID to synchronously control the electric cylinders arranged bilaterally and symmetrically, ensuring the absolute stability of the lifting process and avoiding the twisting deformation of the aircraft structure caused by asynchronous lifting.

[0019] 4) High degree of functional integration, realizing the closed-loop operation of the repair process, shortening the aircraft on-site residence time: The invention creatively designs the emergency repair equipment as a mobile platform that can temporarily replace the faulty wheel. After successfully clamping and lifting the faulty aircraft wheel, the equipment can directly serve as a "new wheel" for the aircraft and move synchronously with the faulty aircraft to the maintenance area under the traction of the tow truck. This function realizes the integration of "jacking-replacement-transportation", eliminates complex steps such as hoisting the faulty aircraft to the transport flatbed truck, greatly reduces the residence time of the faulty aircraft on the runway or taxiway, and maximizes the reduction of the impact on the normal operation of the airport.

[0020] 5) The improved YOLOv8 model of the invention can efficiently run on embedded devices through pruning and multi-resolution inference strategies, meeting the demand for real-time processing on site. The overall control system adopts a PLC + CAN bus architecture, which is mature and reliable, easy to maintain and expand. Therefore, the invention not only has advanced technical concepts, but also has good engineering realizability and promotion value, effectively promoting the evolution of aviation ground support equipment towards intelligence and unmanned. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 is the principle block diagram of the equipment of the invention; Figure 2 is the principle block diagram of the clamping and lifting control system; Figure 3 is the structural diagram of the equipment of the invention; Figure 4 is the principle diagram of the conversion between the depth camera and the world coordinate system; Figure 5 is the training and deployment flowchart of the lightweight YOLOV8 aircraft fault main wheel target detection model; Figure 6 is the flowchart of aircraft fault main wheel detection and positioning; Figure 7 is the flowchart of aircraft fault main wheel emergency intelligent repair. DETAILED DESCRIPTION

[0022] The technical solutions of the invention will be described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the invention, not all. Based on the embodiments in the invention, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the invention.

[0023] Reference Figures 1-7The application provides a technical scheme: an airplane main wheel failure emergency intelligent repair equipment, which comprises a general control system, the general control system is connected with a power system, an operation system, a power supply system, an intelligent detection system and a clamping and lifting control system, and the clamping and lifting control system is connected with a bearing and lifting system. The general control system takes a programmable logic controller as a master control unit, is connected with each subsystem through a controller area network bus and coordinates the control of each subsystem to complete the repair work. The power system is used for providing the overall movement and walking ability of the equipment. The operation system is used for sending manual intervention instructions to the general control system, has two control modes of wireless remote control and wired backup, and ensures that the repair task can be completed through manual operation when the intelligent detection system fails. The intelligent detection system is used for collecting visual information of the failure wheel, processing and identifying the characteristics of the failure wheel based on a deep learning model, and outputting the wheel model, damage state and three-dimensional positioning coordinates in space. The clamping and lifting control system is used for synchronously controlling double electric cylinders arranged bilaterally symmetrically along the feeding direction through a cross-coupling control strategy based on a PID controller to drive the bearing and lifting system to complete the clamping and lifting action according to the optimal clamping position instruction output by the general control system. The bearing and lifting system comprises a vehicle body bearing structure and a lifting structure, is used for clamping and lifting the single-side failure wheel of the airplane, and bears the overall load of the single-side main landing gear of the airplane.

[0024] In the embodiment, the bearing and lifting system is the "skeleton" of the entire emergency intelligent repair equipment, directly contacts the airplane, comprises a vehicle body bearing structure and a lifting structure, is responsible for clamping and lifting the single-side failure main wheel of the airplane, bears the overall load of the single-side main landing gear of the airplane, uniformly distributes the force, the lifting structure is mainly responsible for clamping and lifting the single-side failure main wheel of the airplane, has simple structure and high reliability, can realize clamping and lifting integration, has an adjustable wheel clamping mechanism (an adaptive contact surface is arranged on the inner side, the clamping angle can be automatically adjusted according to the airplane failure main wheel model detected by the intelligent detection system), a multi-stage telescopic electric cylinder structure is adopted for the lifting structure, the maximum lifting height is not less than 1.5 meters, different sizes of wheels can be adapted, the height difference caused by wheel explosion can be actively compensated, the airplane is prevented from rolling over, and good dynamic performance is ensured in the process of traction and moving. The clamping and lifting control system needs to accurately control the action of the bearing and lifting system according to the optimal clamping position output by the overall control system. In the process of connecting with the single-side fault main wheel of the airplane, the clamping and lifting control system adopts a classic electromechanical servo system to accurately control the motion of the lifting system, which can safely and smoothly lift the single-side fault main wheel of the airplane. The control principle block diagram of the clamping and lifting control system is shown in Figure 2 The clamping and lifting control system is composed of a main controller, a servo driver, an electric cylinder, and a driving power supply, and is bilaterally symmetrical along the feeding direction. The clamping and lifting control system is arranged with double electric cylinders to eliminate the torsion phenomenon caused by load imbalance. When the clamping and lifting control system receives the optimal clamping position output by the overall control system, it starts to control the double electric cylinders to move towards the expected lifting position. The main controller collects the actual lifting position of the double electric cylinders in real time and compares it with the expected lifting position. Based on the cross-coupling control of the PID controller, the two electric cylinders are synchronously coordinated to lift the main wheel towards the expected lifting position. In order to avoid the gradual increase of the position error of the two electric cylinders and to prevent divergence, a PID controller is added in the feedback channel to obtain a better compensation signal.

[0025] The overall control system is responsible for receiving and analyzing the operation instructions issued by the command center and the control system. The overall control system adopts a PLC controller as the main control unit and uses CAN bus technology to realize information interaction and data communication among the power supply system, the power system, the clamping and lifting control system, and the intelligent detection system.

[0026] The control system is used to ensure that the repair task can be completed manually when the intelligent detection system fails. It mainly includes a transmitter and a receiver. The transmitter is portable and operated by the operator, and the receiver is installed on the airplane main wheel fault emergency intelligent repair equipment. The two are used in cooperation. The transmitter is responsible for collecting operation instructions and state information such as buttons and joysticks, generating corresponding control instructions, and sending them to the receiver through a wireless radio frequency module. The wireless communication is the main control mode, and the cable control is the backup mode. The receiver is responsible for receiving and analyzing the control instruction information and transmitting it to the overall control system through the CAN bus interface to control the airplane main wheel fault emergency intelligent repair equipment.

[0027] The power system is composed of a driving wheel at the rear end of the airplane main wheel fault emergency intelligent repair equipment, a permanent magnet synchronous motor, a speed reducer, an encoder, and a universal driven wheel at the front end of the airplane main wheel fault emergency intelligent repair equipment. They are symmetrically distributed on both sides of the vehicle body and are rigidly connected with the outer frame. The motor provides power for the driving wheel, the speed reducer reduces speed and increases torque, and the torque demand of the airplane main wheel fault emergency intelligent repair equipment is guaranteed. The airport rescue management personnel can use remote control to freely adjust the attitude of the airplane main wheel fault emergency intelligent repair equipment, realize forward, reverse and differential steering operations, etc.

[0028] The intelligent emergency repair equipment for aircraft main wheel failure consists of a lifting structure, telescopic mechanism, wheel clamping mechanism, indicator lights, travel indicator lights, reversing radar, wired interface, lighting, anti-slip mats, and vehicle body shell, etc. Its external layout diagram is shown below. Figure 3 As shown. The vehicle's outer shell uses composite armor materials, ensuring both structural strength and lightweight design. The load-bearing and lifting structures feature a double locking device to effectively prevent accidental slippage during operation. The telescopic mechanism incorporates a high-precision displacement sensor, capable of monitoring changes in the mechanism's position in real time and feeding precise position information back to the clamping and lifting control system, achieving millimeter-level positioning accuracy. The wheel clamping mechanism employs adaptive clamping technology, automatically adjusting the clamping force according to different wheel hub sizes. The indicator light system uses LED light sources and has an intelligent fault self-diagnosis function. When abnormal voltage or circuit faults are detected, it automatically switches to a bright red flashing mode to promptly alert the operator. The travel indicator lights and reversing radar form a linked system, ensuring operational safety even in low-visibility environments. The wired interface supports multi-protocol data transmission, enabling seamless integration with various aviation ground support equipment. The lighting assembly uses a wide-angle lens design, covering a range of 120 degrees. The anti-slip mat surface undergoes special treatment, increasing the coefficient of friction to over 0.8. The vehicle body features a double-layer structure, with the inner layer filled with high-performance energy-absorbing material. In the event of a collision, this material effectively absorbs and disperses impact energy, providing additional safety protection for the equipment. Considering that in the event of a malfunction and inability to move the aircraft tire blowout emergency intelligent repair equipment during field use, a crane is needed to lift the equipment onto a trailer and remove it from the accident site. Therefore, high-strength alloy forged lifting rings are installed around the vehicle body, enabling "four-point balanced lifting." The crane's wire ropes are connected through these lifting rings, allowing for precise control of the vehicle's horizontal attitude (tilt angle ≤3°), preventing rollover or chassis deformation due to center of gravity shift.

[0029] In some embodiments, the intelligent detection system includes a depth camera and a deep learning processing module; The depth camera is used to simultaneously acquire color images and depth images, and to obtain the coordinates of pixels in the image in the world coordinate system based on the camera's intrinsic and extrinsic parameters. The deep learning processing module is internally provided with a lightweight YOLOV8 aircraft fault host wheel target detection model, the lightweight YOLOV8 aircraft fault host wheel target detection model takes a YOLOV8 model as a basic model, adopts a MobileNetV3-large network fused with an ECA-Net attention mechanism as a backbone feature extraction network of the YOLOV8 model, replaces a PANet neck of the YOLOv8 model with a deep separable convolution version of a BiFPN, introduces learnable weight parameters in a feature fusion path in the BiFPN, and simultaneously introduces a multi-scale hollow convolution module and a parallel key point detection branch in a head network layer, for positioning a ground center and a hub center of a wheel, and providing geometric constraints for optimal clamping position calculation; The MobileNetV3-large network fused with the ECA-Net attention mechanism is adopted as the backbone feature extraction network of the YOLOv8 model, the PANet neck of the YOLOv8 is replaced with the deep separable convolution version of the BiFPN, and the multi-scale hollow convolution module and the parallel key point detection branch are introduced in the head network layer, for identifying a model, a damage type and an optimal clamping position of the aircraft host wheel.

[0030] In the embodiment, the intelligent detection system mainly consists of a depth camera and a deep learning processing module, for identifying a fault wheel feature and determining a distance between the fault wheel and the equipment. The depth camera mainly consists of an RGB camera, an infrared camera and an infrared projector, can simultaneously shoot a color image and a depth image, and can obtain coordinates of a pixel point in a world coordinate system according to camera internal and external parameters, a specific principle of which is shown in Figure 4 X W , Y W , Z W ), a coordinate of the P point in the camera coordinate system can be obtained by the following formula: In the formula, is a homogeneous transformation matrix from the camera coordinate system to the world coordinate system; a two-dimensional coordinate of the P point on the imaging plane ( X , Y ) can be obtained by the following formula: ; In the formula, f is a focal length of the depth camera, and formula (2) is substituted into formula (1) to obtain: A pixel coordinate system of the P point can be obtained by the following formula: ​ wherein, x d 、 y d is the normalized coordinate of the point after camera distortion, c x 、 c y is the position of the depth camera optical center in the horizontal and vertical directions in the pixel coordinate system, f x 、 f y is the camera intrinsic parameter.

[0031] In combination with formulas (1)-(4), the following can be obtained: According to the depth camera intrinsic and extrinsic parameters, a point in the pixel coordinate system can be converted to the world coordinate system according to formula (5).

[0032] The improved lightweight YOLOV8 aircraft fault main wheel target detection model is used in the deep learning processing module to identify the model, damage type and optimal clamping position of the aircraft main wheel. The lightweight YOLOV8 aircraft fault main wheel target detection model takes the YOLOV8 model as the basic model, and the model mainly consists of an input layer, a backbone network layer, a neck network layer and a head network layer. The MobileNetV3-large network fused with the ECA-Net attention mechanism is used as the backbone feature extraction network of the YOLOV8 model. The ECA-Net realizes channel attention through one-dimensional convolution, has smaller calculation amount and is more suitable for embedded deployment, can enhance the feature representation capability while maintaining lightweight, replaces the PANet neck of the YOLOV8 with the deep separable convolution version of the BiFPN, uses the deep separable convolution to reconstruct the feature fusion path in the BiFPN, introduces the learnable weight parameter, lets the network adaptively learn the importance of different feature layers, introduces the multi-scale hollow convolution module in the head network layer, expands the receptive field without increasing downsampling, better captures the local detail features of the wheel, increases the parallel key point detection branch, accurately locates the wheel ground center and hub center, and provides geometric constraints for optimal clamping position calculation.

[0033] In some embodiments, the depth camera includes an RGB camera, an infrared camera and an infrared projector.

[0034] In some embodiments, the training and deployment of the lightweight YOLOV8 aircraft fault main wheel target detection model includes the following steps: In the data set construction stage, aircraft wheel images under various typical working conditions are collected, and a hierarchical sampling method is used to divide the data set into a training set, a validation set and a test set according to a preset ratio; In the model training stage, the lightweight YOLOV8 aircraft fault main wheel target detection model is pre-trained on a large general data set, and then the model is trained on the self-constructed aircraft wheel data set. Whether the model result meets the expected requirement is observed, and the model is fine-tuned and optimized for the target detection task. In the model optimization and deployment stage, the model pruning based on prior knowledge and the adaptive multi-resolution inference strategy are implemented for the trained model, and the optimized model is deployed to the intelligent detection system to realize real-time high-performance detection.

[0035] In this embodiment, after the lightweight YOLOV8 aircraft fault main wheel target detection model is built, it is trained and deployed, and the process is as shown in Figure 5 The aircraft wheel images under various typical working conditions are collected to construct a data set. In order to ensure the balance and rationality of the data set division, a hierarchical sampling method is used to divide the data set into a training set, a validation set and a test set according to a ratio of 8:1:1, to ensure the sample independence and distribution consistency among the three. The lightweight YOLOV8 aircraft fault main wheel target detection model is pre-trained on a large general data set, and then the model is trained on the self-constructed aircraft wheel data set. Whether the model training result meets the expected requirement is observed. If it does not meet the expected requirement, the model is fine-tuned and optimized for the target detection. If it meets the expected requirement, the model pruning based on prior knowledge and the adaptive multi-resolution inference strategy are implemented, and the model is deployed on the intelligent detection device to realize real-time high-performance detection.

[0036] The lightweight YOLOV8 aircraft fault main wheel target detection model outputs the detection frame of the detected aircraft fault main wheel and the coordinates of the top-left corner vertex and the bottom-right corner vertex of the detection frame, which are respectively u lt , v lt ), ( u rd , v rd ), and the center point coordinates of the detection frame u m , v m ) are taken as the positioning point of the aircraft fault main wheel. The center point coordinates of the detection frame u m , v m ) can be obtained by the following formula: In the lightweight YOLOV8 aircraft fault host wheel target detection model, the world coordinate system of the depth camera optical center point is X o , Y o , Z o ), the world coordinate system of the aircraft fault host wheel positioning point is X m , Y m , Z m ) can be obtained by formula (5), and under the condition that the specific coordinates of two points in space are known, the distance between the two points, that is, the distance between the aircraft fault host wheel and the depth camera, can be obtained by solving the Euclidean distance between the two points L o can be obtained by the following formula: .

[0037] In some embodiments, the wheel clamping unit is provided with an adaptive contact surface, which can automatically adjust the clamping angle according to the model of the aircraft fault host wheel detected by the intelligent detection system. The lifting structure adopts a multi-stage telescopic electric cylinder structure.

[0038] In some embodiments, when detecting and positioning the aircraft fault host wheel, the following steps are included: Based on the imaging principle and internal and external parameters of the depth camera, the conversion relationship between the camera coordinate system and the world coordinate system is derived; Drive the intelligent repair equipment to the vicinity of the fault host wheel and adjust the position of the depth camera, and then obtain the color image and depth image of the fault host wheel; Input the obtained color image of the aircraft fault host wheel into the deployed lightweight YOLOV8 aircraft fault host wheel target detection model to obtain the pixel coordinates of the aircraft fault host wheel center positioning point; According to the pixel coordinates of the aircraft fault host wheel center positioning point and the depth information of the aircraft fault host wheel obtained by the depth camera, the position of the aircraft fault host wheel in the world coordinate system and the relative distance from the depth camera are calculated; The detected three-dimensional coordinates and relative distance information of the aircraft fault host wheel are fed back to the overall control system for subsequent navigation and docking control.

[0039] In this embodiment, as shown in Figure 6 , the first step is to derive the conversion relationship between the camera coordinate system and the world coordinate system based on the imaging principle and internal and external parameters of the depth camera.

[0040] Second, the intelligent detection device is driven to the vicinity of the aircraft fault host wheel, the position of the depth camera carried on the device is adjusted, and then color images and depth images of the aircraft fault host wheel are obtained.

[0041] Third, the color images of the aircraft fault host wheel are input into the lightweight YOLOV8 aircraft fault host wheel target detection model deployed on the intelligent detection device, and the pixel coordinates of the center positioning point of the aircraft fault host wheel are obtained through the target detection model.

[0042] Fourth, the position of the aircraft fault host wheel in the world coordinate system and the relative distance from the depth camera, i.e., the intelligent monitoring device, are calculated according to the pixel coordinates of the center positioning point of the aircraft fault host wheel output by the target detection model and combined with the depth information of the aircraft fault host wheel obtained by the depth camera.

[0043] Fifth, the detected position and distance information of the aircraft fault host wheel are fed back to the overall control system of the intelligent detection device for subsequent processing.

[0044] In some embodiments, when the aircraft fault host wheel is repaired, the following steps are included: Receiving dispatching instructions from the tower and ground crew; The overall control system controls the intelligent repair device to autonomously navigate to the vicinity of the aircraft fault host wheel according to the dispatching instructions; After arriving at the vicinity of the fault host wheel, the intelligent detection system adjusts the detection position to obtain the three-dimensional coordinates and relative distance information of the aircraft fault host wheel and feeds back to the overall control system; The overall control system calculates the optimal clamping position according to the three-dimensional coordinates and relative distance information and outputs it to the clamping control system; The clamping control system controls the bearing and lifting system to clamp and lift the fault host wheel according to the optimal clamping position received, and in this process, the clamping control system collects the actual lifting position of the bearing and lifting system in real time and compares it with the expected lifting position, and controls the bearing and lifting system to move towards the expected lifting position based on the cross-coupling control strategy of the PID controller; After completing the clamping and lifting, the intelligent repair device replaces the fault host wheel as a temporary host wheel and moves synchronously with the fault aircraft under the traction of the tractor, completing the repair work.

[0045] In this embodiment, as shown in Figure 7 The first step is to receive dispatching instructions from the tower and ground crew.

[0046] Second, the intelligent detection device autonomously navigates to the vicinity of the aircraft fault host wheel according to the dispatching instructions.

[0047] The third step, after reaching the vicinity of the fault host wheel, the intelligent detection system carried by the intelligent detection device adjusts the detection position to obtain the position and distance information of the fault host wheel of the airplane and feeds back to the overall control system.

[0048] The fourth step, the overall control system calculates the optimal clamping position according to the position and distance information of the fault host wheel of the airplane fed back by the intelligent detection system and outputs to the clamping control system.

[0049] The fifth step, the clamping control system controls the bearing and lifting system to clamp and lift the fault host wheel according to the optimal clamping position received, during which the clamping control system collects the actual lifting position of the bearing and lifting system in real time and compares it with the expected lifting position, and controls the bearing and lifting system to clamp and lift the host wheel towards the expected lifting position based on the cross-coupling of the PID controller.

[0050] The sixth step, after clamping and lifting the fault host wheel, the intelligent detection device can be used as a host wheel to replace the fault host wheel of the airplane, and the tow truck can separate the fault airplane from the airport to complete the repair work.

[0051] The above is only the preferred embodiment of the present application, it should be understood that the present application is not limited to the form disclosed herein, should not be considered as excluding other embodiments, and can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein, by the above teaching or related art or knowledge. The modification and change made by the person skilled in the art without departing from the spirit and scope of the present application shall be within the protection scope of the appended claims of the present application.

Claims

1. An intelligent emergency repair device for aircraft main engine wheel failure, characterized in that: It includes an overall control system, which is connected to a power system, an operating system, a power supply system, an intelligent detection system, and a clamping and lifting control system, which is connected to a load-bearing and lifting system; The overall control system uses a programmable logic controller as the main control unit, and communicates with each subsystem through the controller local area network bus to coordinate and control each subsystem to complete the emergency repair operation. The power system is used to provide the overall movement and mobility of the equipment; The control system is used to send manual intervention commands to the overall control system. It has two control modes: wireless remote control and wired backup, ensuring that emergency repairs can still be completed manually when the intelligent detection system fails. The intelligent detection system is used to collect visual information of the faulty wheel, and to process the visual information based on a deep learning model to identify the characteristics of the faulty wheel, so as to output the wheel model, damage status and three-dimensional positioning coordinates in space. The clamping and lifting control system is used to synchronously control the dual electric cylinders arranged symmetrically on both sides along the feed direction according to the optimal clamping position command output by the overall control system and a cross-coupling control strategy based on a PID controller, so as to drive the load-bearing and lifting system to complete the clamping and lifting action. The load-bearing and lifting system includes a vehicle body load-bearing structure and a lifting structure, which are used to clamp and lift a faulty wheel on one side of the aircraft to bear the entire load of the main landing gear on that side; the lifting structure has an adjustable wheel clamping unit to accommodate wheels of different sizes and can actively compensate for the height difference caused by wheel explosion.

2. The intelligent emergency repair equipment for aircraft main wheel failure according to claim 1, characterized in that: The intelligent detection system includes a depth camera and a deep learning processing module; The depth camera is used to simultaneously acquire color images and depth images, and to obtain the coordinates of pixels in the image in the world coordinate system based on the camera's intrinsic and extrinsic parameters. The deep learning processing module incorporates a lightweight YOLOV8 aircraft faulty main wheel target detection model. This model uses the YOLOV8 model as its base model and employs a MobileNetV3-large network with an ECA-Net attention mechanism as the backbone feature extraction network of the YOLOV8 model. The PANet neck of the YOLOv8 model is replaced with a depthwise separable convolutional version of BiFPN, and learnable weight parameters are introduced into the feature fusion path in BiFPN. At the same time, a multi-scale dilated convolutional module and a parallel key point detection branch are introduced into the head network layer to locate the wheel grounding center and the hub center, providing geometric constraints for calculating the optimal clamping position. The MobileNetV3-large network with integrated ECA-Net attention mechanism is used as the backbone feature extraction network of the YOLOv8 model. The PANet neck of YOLOv8 is replaced with a depthwise separable convolutional version of BiFPN. Multi-scale dilated convolutional modules and parallel key point detection branches are introduced into the head network layer to identify the model, damage type and optimal clamping position of the aircraft main wheel.

3. The intelligent emergency repair equipment for aircraft main wheel failure according to claim 2, characterized in that: The depth camera includes an RGB camera, an infrared camera, and an infrared projector.

4. The intelligent emergency repair equipment for aircraft main wheel failure according to claim 2, characterized in that: The training and deployment of the lightweight YOLOV8 aircraft faulty main wheel target detection model includes the following steps: During the dataset construction phase, images of aircraft wheels under various typical working conditions were collected, and the dataset was divided into training set, validation set and test set according to a preset ratio using a hierarchical sampling method. During the model training phase, the lightweight YOLOV8 aircraft faulty main wheel target detection model was pre-trained on a large general dataset. Then, the model was trained on a self-built aircraft wheel dataset. The model results were observed to see if they met the expected requirements, and fine-tuning and optimization were performed for the target detection task. During the model optimization and deployment phase, the trained model is subjected to model pruning and adaptive multi-resolution inference strategies based on prior knowledge, and the optimized model is deployed to the intelligent detection system to achieve real-time high-performance detection.

5. The intelligent emergency repair equipment for aircraft main wheel failure according to claim 1, characterized in that: The aforementioned wheel clamping unit is equipped with an adaptive contact surface, which can automatically adjust the clamping angle according to the model of the faulty main main wheel of the aircraft detected by the intelligent detection system. The lifting structure adopts a multi-stage telescopic electric cylinder structure.

6. The intelligent emergency repair equipment for aircraft main engine wheel failure according to any one of claims 1-5, characterized in that: The following steps are included when inspecting and locating a faulty main engine wheel on an aircraft: Based on the imaging principle and intrinsic and extrinsic parameters of depth cameras, the transformation relationship between the camera coordinate system and the world coordinate system is derived. Drive the intelligent repair equipment to the vicinity of the faulty main wheel and adjust the position of the depth camera, then acquire color and depth images of the faulty main wheel; The acquired color image of the faulty main wheel of the aircraft is input into the deployed lightweight YOLOV8 aircraft faulty main wheel target detection model to obtain the pixel coordinates of the center positioning point of the faulty main wheel of the aircraft. The position of the faulty main wheel in the world coordinate system and its relative distance from the depth camera are calculated based on the pixel coordinates of the center positioning point of the faulty main wheel and the depth information of the faulty main wheel obtained by the depth camera. The detected three-dimensional coordinates and relative distance information of the faulty main wheel of the aircraft are fed back to the overall control system for subsequent navigation and docking control.

7. The intelligent emergency repair equipment for aircraft main engine wheel failure according to any one of claims 1-5, characterized in that: When repairing a faulty main landing gear wheel of an aircraft, the following steps are included: Receive dispatch instructions from the control tower and ground support; The overall control system, based on dispatch instructions, controls the intelligent repair equipment to autonomously navigate to the vicinity of the aircraft's faulty main landing gear; Upon reaching the vicinity of the faulty main wheel, the intelligent detection system adjusts its detection position to obtain the three-dimensional coordinates and relative distance information of the faulty main wheel and feeds it back to the overall control system. The overall control system calculates the optimal clamping position based on the three-dimensional coordinates and relative distance information and outputs it to the clamping and lifting control system. The clamping and lifting control system controls the load-bearing and lifting system to clamp and lift the faulty main engine wheel based on the received optimal clamping position. During this process, the clamping and lifting control system collects the actual lifting position of the load-bearing and lifting system in real time and compares it with the desired lifting position. Based on the cross-coupling control strategy of the PID controller, the system controls the load-bearing and lifting system to move toward the desired lifting position. After the clamping and lifting are completed, the intelligent repair equipment serves as a temporary replacement for the faulty main wheel. It moves synchronously with the faulty aircraft under the towing of the tractor to complete the repair operation.

Images