Method and system for sensing and controlling wing shape of morphing aircraft

Abstract

The invention discloses a wing shape sensing and control method and system for a morphing aircraft, and belongs to the technical field of aircraft flight detection and control. When the error between the actual angle of the wing and the required angle exceeds a preset error threshold value, predicting the motion state of the feature points through Kalman filtering, updating the feature points to obtain the variable quantity of the sweepback angle and the span of the wing, performing timestamp and synchronization processing on the variable quantity and the state quantity, and performing multi-source information fusion to obtain the wing. State estimation values of the sweepback and the span of the wing are obtained; and taking the state estimation value, the current flight attitude information of the aircraft and the error between the actual angle of the wing and the required angle as the input of a controller, and carrying out self-adaptive adjustment on the sweepback angle and the elongation until the error between the actual angle of the wing and the required angle meets an error threshold value. According to the method, the precision and robustness of dynamic deformation monitoring and control of the morphing wing can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of aircraft flight detection and control, and more particularly to a variable aircraft wing shape sensing and control method and system. BACKGROUND

[0002] With the increasing demand for multi-mission, multi-flight envelope aircraft, the traditional fixed configuration wing has been difficult to maintain optimal performance in different flight states.

[0003] Variable wing (variable sweep angle, variable span) technology can optimize lift, drag and maneuverability of the aircraft in different working conditions by actively adjusting the wing shape. For example, at high speed cruise, a large sweep angle and a small span are used to reduce drag; at take-off and low speed flight stage, a small sweep angle and a large span are used to improve lift.

[0004] Currently, the common methods for wing shape deformation monitoring include contact sensors such as strain gauges and fiber Bragg gratings. This method can only measure a single point, and cannot cover and detect the overall shape deviation or local structural abnormalities of the wing. Moreover, such contact sensors can affect the structure of the wing itself. The existing contact sensors cannot real-time sense the actual wing deformation state. When there is a gap in mechanical transmission, external force interference or environmental influence, the control command and the actual wing shape may not be consistent. The existing wing deformation control strategy is mostly based on pre-setting or model predictive control, and lacks dynamic adjustment of wing shape according to flight attitude (climbing, cruising, diving, etc.).

[0005] In summary, in the existing variable wing technology, when the monitoring process of wing deformation (including sweep angle and span change) is disturbed by external force or environment, the variable wing cannot be adjusted in attitude accurately only by sensor detection. SUMMARY

[0006] To solve the above problems, the present application provides a variable aircraft wing shape sensing and control method and system. When the error between the actual angle of the wing and the required angle exceeds the preset error threshold, the feature points are updated in real time through state prediction, and then the change amount of the sweep angle and the span of the wing is determined. After time stamp alignment and synchronization processing of the sweep angle and the span state amount of the wing obtained by the sensor, and multi-source information fusion, the random noise and systematic deviation caused by a single sensor can be effectively suppressed, the stability and reliability of state estimation can be improved, and stable estimation of the span and sweep angle parameters can be realized.

[0007] To solve the above technical problems, the present application discloses a variable aircraft wing shape sensing and control method, comprising the following steps: The positions of preset feature points on the inner and outer wing sections of the variant wing are obtained to determine the actual wing angle; the sweep angle and span of the wing are obtained through sensors mounted on the aircraft's rotation axis. When the error between the actual wing angle and the required angle exceeds the preset error threshold, the feature point becomes visually ineffective. The motion state of the feature point is predicted by Kalman filtering to obtain the updated feature point. The updated feature point is then subjected to three-dimensional reconstruction and deformation calculation in sequence, and the changes in the wing sweep angle and span are estimated. After aligning and synchronizing the state variables and changes of the wing sweep angle and span with timestamps, and then performing multi-source information fusion, the state estimates of the wing sweep angle and span are obtained. The system acquires the current flight attitude information of the aircraft; using the state estimate, the current flight attitude information of the aircraft, and the error between the actual wing angle and the required angle as inputs to the controller, it adaptively adjusts the sweep angle and span until the error between the actual wing angle and the required angle meets the error threshold.

[0008] Preferably, the step of obtaining the positions of preset feature points on the surfaces of the inner and outer wing sections of the variant wing to determine the actual wing angle specifically includes: Extract the center coordinates of feature points at the zero-point position of the wing, acquire the first frame of the zero-point position image, and select all feature points on the wing surface through manual interaction; Using any feature point corner as a reference, a search area of ​​a certain radius is set around it. The similarity is calculated point by point through a sliding template to generate a matching score map. The position with the highest similarity in the matching score map is selected to determine the sub-pixel coordinates of the feature point corner. Based on the sub-pixel coordinates of the feature point corner points, a coordinate measuring machine is used to obtain the three-dimensional coordinates of the feature point in the wing coordinate system. After performing coordinate transformation on the three-dimensional coordinates, straight line fitting is performed on the feature points of the inner and outer wing sections respectively to obtain the corresponding straight line equations representing the directions of the inner and outer wing sections. By determining the included angle between the two straight lines, the actual angle of the wing is obtained.

[0009] Preferably, when the error between the actual wing angle and the required angle exceeds a preset error threshold, the feature point experiences visual failure. Kalman filtering is then used to predict the motion state of the feature point to obtain an updated feature point. Specifically, this includes: Obtain the image of the feature point locations in each frame; By establishing a mapping relationship between the actual wing angle and the feature point position, feature point detection and matching are performed on the image of each frame of feature point position to obtain the sub-pixel level image coordinates of all feature point corners. Based on the subpixel-level image coordinates of all feature point corners, the confidence level of the feature points is determined to determine whether the feature points have experienced visual failure. When the confidence level of a feature point exceeds the confidence threshold, the feature point becomes visually ineffective. A motion state prediction method based on Kalman filtering is used to update the position and track the trajectory of the feature point that has become visually ineffective, and the updated feature point is obtained.

[0010] Preferably, the step of sequentially performing 3D reconstruction and deformation calculation on the updated feature points to obtain real-time changes in the wing's sweep angle and span specifically includes: Obtain the subpixel-level image coordinates of all feature point corners after the update, and obtain the world coordinates of all feature point corners through 3D reconstruction; Based on the updated world coordinates of all feature points and corner points, the feature points of the inner wing section and the outer wing section are evenly distributed in two rows. Each row of the inner wing section includes more than two feature points; each row of the outer wing section includes at least one feature point. Taking a certain feature point of a certain row of the selected inner wing section as a reference point, all feature points are projected onto the tangent plane of the wing surface where the reference point is located. The tangent plane is the z-axis plane. In the tangent plane, with the reference point as a straight line, a vector A is obtained by fitting a straight line to all feature points of the inner wing segment on the straight line; a vector B is obtained by fitting a least-squares straight line to all feature points of the other row of the inner wing segment; vectors A and B are averaged and normalized to obtain the wing's direction vector C along the wingspan direction; by calculating the angular change of the direction vector C, the change in the wing's sweep angle is obtained. The distance change of each feature point of the outer wing segment relative to all feature points of the inner wing segment on the tangent plane of the reference point is determined and averaged to obtain the amount of expansion and contraction of the outer wing segment relative to the inner wing segment, i.e., the amount of span change.

[0011] Preferably, the step of using the state estimate, the aircraft's current flight attitude information, and the error between the actual and required wing angles as inputs to the controller to adaptively adjust the sweep angle and span until the error between the actual and required wing angles meets an error threshold specifically includes: The error between the actual and required wing angles, the state estimate, and the aircraft's flight attitude information are used as inputs to the fuzzy PID controller. When the error between the actual and required wing angles exceeds a preset error threshold, the PID parameter k is dynamically adjusted. p k i k d : When the aircraft's current flight attitude is in a rapid climb or dive phase, the fuzzy PID controller increases k... d and reduce k i The output control command is a large sweep angle and a small span; When the aircraft is currently in the cruise phase, the fuzzy PID controller increases k... i The output control command is a small sweep angle and a large span.

[0012] Preferably, the step of aligning and synchronizing the state variables and changes of the wing sweep angle and span with timestamps and then fusing multi-source information to obtain state estimates of the wing sweep angle and span specifically includes: Align the wing sweep angle and span state variables and changes to a unified timestamp, and use linear interpolation or spline interpolation to process the asynchronous data to obtain the state variables and changes after timestamp alignment and synchronization. The state variables and changes after timestamp alignment and synchronization are fused with multi-source information, and the state prediction of the wing sweep angle and span is performed by the Kalman filter algorithm to obtain the state estimates of the wing sweep angle and span.

[0013] Preferably, the current flight attitude information of the aircraft is obtained by encoding through sensors mounted on the aircraft's rotation axis, including pitch angle, roll angle, angular velocity, and acceleration.

[0014] Preferably, it also includes a vari-plane wing shape sensing and control system, comprising: The attitude measurement module is used to acquire the current flight attitude information of the aircraft; The visual online monitoring module is used to acquire the positions of preset feature points on the inner and outer wing sections of the variant wing to determine the actual wing angle; it acquires the wing's sweep angle and span state variables through sensors mounted on the aircraft's rotation axis; when the error between the actual wing angle and the required angle exceeds a preset error threshold, the feature points experience visual failure, and the motion state of the feature points is predicted using Kalman filtering to obtain updated feature points; the updated feature points are then subjected to 3D reconstruction and deformation calculations to estimate the changes in the wing's sweep angle and span. The multi-source information fusion module is used to align and synchronize the state variables and changes of the wing sweep angle and span after time stamping, and then perform multi-source information fusion to obtain the state estimates of the wing sweep angle and span. The control execution module is used to adaptively adjust the sweep angle and span using the state estimate, the current flight attitude information of the aircraft, and the error between the actual wing angle and the required angle as inputs to the controller, until the error between the actual wing angle and the required angle meets the error threshold.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The proposed method for variator wing shape perception and control acquires the positions of preset feature points on the inner and outer wing sections of the variator wing to determine the actual wing angle. When the error between the actual wing angle and the required angle exceeds a preset error threshold, the feature points experience visual failure. Kalman filtering is used to predict the motion state of the feature points, and position updates and trajectory tracking are performed on the visually failed feature points to maintain continuous observation. By timestamping and synchronizing the wing's sweep angle and span state variables and then fusing multi-source information, random noise and systematic biases from single sensors can be effectively suppressed, improving the stability and reliability of state estimation. A fuzzy PID controller automatically adjusts the wing morphology to achieve stable estimation of span and sweep angle parameters, improving the aerodynamic efficiency and mission adaptability of the aircraft. The method proposed in this invention uses a non-contact visual measurement method to compensate for contact measurement methods such as strain gauges and accelerometers. It can perform full-field deformation observation without affecting the wing structure. The structure is simple and easy to maintain, so as to realize real-time monitoring of the wing deformation process and improve the stability of vision in complex environments such as changes in lighting, cloud cover, and smoke interference. Attached Figure Description

[0016] Figure 1 This is a flowchart of the wing shape sensing and control method for variant aircraft proposed in this invention; Figure 2 A schematic diagram of a variant wing provided for an embodiment of the present invention; Figure 3 The embodiments of the present invention provide a basis for... Figure 2 The diagram shows the sweep angle of the wing during rotational deformation and the rear span during telescoping deformation. Figure 4 A schematic diagram of a vari-shaped aircraft wing shape sensing and control system provided in an embodiment of the present invention; Figure 5 Detailed modules comprising the visual online monitoring module provided in the embodiments of the present invention; In the diagram: 11. Inner wing section; 12. Outer wing section; 13. Linear motion unit; 14. Rotational motion unit; 21-28. Feature points. Detailed Implementation

[0017] The following will refer to the appendices in the embodiments of the present invention. Figures 1-5 The technical solutions in the embodiments of the present invention will be clearly and completely described. It should be understood that the terminology used in the present invention is only for describing particular implementation methods and is not intended to limit the present invention.

[0018] Example like Figure 1 As shown, this invention proposes a method for sensing and controlling the shape of a variant aircraft wing, comprising the following steps: S1: Obtain the positions of preset feature points on the inner and outer wing sections of the variant wing to determine the actual wing angle; obtain the sweep angle and span of the wing through sensors mounted on the aircraft's rotation axis. S2: When the error between the actual wing angle and the required angle exceeds the preset error threshold, the feature point becomes visually ineffective. The motion state of the feature point is predicted by Kalman filtering to obtain the updated feature point. The updated feature point is then subjected to three-dimensional reconstruction and deformation calculation in sequence. The changes in the wing sweep angle and span are estimated. S3: After timestamping and synchronizing the state variables and changes of the wing sweep angle and span, and then fusing multi-source information, the state estimates of the wing sweep angle and span are obtained. S4: Obtain the current flight attitude information of the aircraft; use the state estimate, the current flight attitude information of the aircraft, and the error between the actual wing angle and the required angle as inputs to the controller, and adaptively adjust the sweep angle and span until the error between the actual wing angle and the required angle meets the error threshold.

[0019] In step S1, the positions of preset feature points on the surfaces of the inner and outer wing sections of the variant wing are obtained to determine the actual wing angle, specifically including: The inner wing section can rotate around the hinge at the root of the fuselage to change the sweep angle, and the outer wing section can extend and retract along the wingspan to change the span. Several high-contrast feature points are evenly distributed on the surfaces of the inner and outer wing sections. Extract the center coordinates of feature points at the zero-point position of the wing, acquire the first frame of the zero-point position image, and select all feature points on the wing surface through manual interaction; Using any feature point corner as a reference, a search area of ​​a certain radius is set around it. The similarity is calculated point by point through a sliding template to generate a matching score map. The position with the highest similarity in the matching score map is selected to determine the sub-pixel coordinates of the feature point corner. Based on the sub-pixel coordinates of the feature point corner points, a coordinate measuring machine is used to obtain the three-dimensional coordinates of the feature point in the wing coordinate system. After performing coordinate transformation on the three-dimensional coordinates, straight line fitting is performed on the feature points of the inner and outer wing sections respectively to obtain the corresponding straight line equations representing the directions of the inner and outer wing sections. By determining the included angle between the two straight lines, the actual angle of the wing is obtained.

[0020] In step S2, when the error between the actual wing angle and the required angle exceeds a preset error threshold, the feature point experiences visual failure. The motion state of the feature point is then predicted using Kalman filtering to obtain updated feature points, specifically including: Obtain the image of the feature point locations in each frame; By establishing a mapping relationship between the actual wing angle and the feature point position, feature point detection and matching are performed on the image of each frame of feature point position to obtain the sub-pixel level image coordinates of all feature point corners. Based on the subpixel-level image coordinates of all feature point corners, the confidence level of the feature points is determined to determine whether the feature points have experienced visual failure. When the confidence level of a feature point exceeds the confidence threshold, the feature point becomes visually ineffective. A motion state prediction method based on Kalman filtering is used to update the position and track the trajectory of the feature point that has become visually ineffective, and the updated feature point is obtained.

[0021] The updated feature points are then subjected to 3D reconstruction and deformation calculations. The changes in the wing's sweep angle and span are estimated, specifically including: Obtain the subpixel-level image coordinates of all feature point corners after the update, and obtain the world coordinates of all feature point corners through 3D reconstruction.

[0022] Based on the updated world coordinates of all feature point corners, such as Figure 2 The diagram shown is a schematic of a variant wing provided in an embodiment of the present invention. The feature points of the inner wing section and the outer wing section are evenly distributed in two rows. Each row of the inner wing section includes two or more feature points; each row of the outer wing section includes at least one feature point. Taking a certain feature point of a certain row of the inner wing section as a reference point, all feature points are projected onto the tangent plane of the wing surface where the reference point is located. The tangent plane is the z-axis plane. In the tangent plane, using feature point 21 as the reference point and the reference point as the straight line, vector A is obtained by fitting a straight line to all feature points 21, 22, and 23 on the inner wing segment along the straight line; vector B is obtained by fitting a least-squares straight line to all feature points 24, 25, and 26 on the other row of the inner wing segment; vectors A and B are averaged and normalized to obtain the wing's direction vector C along the wingspan; by calculating the angular change of the direction vector C, the change in the wing's sweep angle is obtained. The distance changes of feature points 27 and 28 of the outer wing segment relative to all feature points 21-26 of the inner wing segment in the tangential plane where feature point 21 is located (as a reference point) are determined and averaged to obtain the expansion and contraction changes of the outer wing segment relative to the inner wing segment, i.e., the change in span.

[0023] like Figure 3 As shown, this is for Figure 2 A schematic diagram of the sweep angle for rotational deformation and the elongation for stretching deformation.

[0024] In step S3, the wing sweep angle and span state variables and their changes are aligned to a unified timestamp, ensuring that the wing sweep angle and span state variables and the estimated changes in wing sweep angle and span have a unified time reference. Linear interpolation or spline interpolation is used to process the asynchronous data, and the timestamp-aligned and synchronized state variables and their changes are obtained by compensating for the time difference. The timestamp-aligned and synchronized state variables and their changes are then fused with the wing sweep angle and span state variables acquired by the sensors using a Kalman filter algorithm to obtain high-precision state estimates of the wing sweep angle and span.

[0025] In step S4, the state estimate, the aircraft's current flight attitude information, and the error between the actual and required wing angles are used as inputs to the controller. The sweep angle and span are adaptively adjusted until the error between the actual and required wing angles meets the error threshold. Specifically, this includes: The error between the actual and required wing angles, the state estimate, and the aircraft's flight attitude information are used as inputs to the fuzzy PID controller. When the error between the actual and required wing angles exceeds a preset error threshold, the PID parameter k is dynamically adjusted. p k i k d In particular, under highly dynamic operating conditions such as rapid ascent and descent, the fuzzy PID controller increases k... d and reduce k i To suppress rapid disturbances and overshoot; under steady-state cruise conditions, by increasing k i To improve control precision.

[0026] When the aircraft is in a rapid climb or dive phase, the control command output by the fuzzy PID controller is a large sweep angle and a small span.

[0027] When the aircraft is in the cruise phase, the control command output by the fuzzy PID controller is a small sweep angle and a large span.

[0028] like Figure 4 As shown, the present invention also proposes a wing shape sensing and control system for a morphing aircraft, comprising: The attitude measurement module is used to acquire the current flight attitude information of the aircraft using an inertial measurement unit (IMU). The IMU is an attitude sensor mounted on the rotation axis of the aircraft. It obtains the current flight attitude information of the aircraft through encoding. Its flight attitude information includes pitch angle, roll angle, angular velocity and acceleration.

[0029] The visual online monitoring module acquires the positions of preset feature points on the inner and outer wing sections of the variant wing to determine the actual wing angle. Sensors mounted on the aircraft's rotation axis acquire the wing's sweep angle and span. When the error between the actual and required wing angles exceeds a preset error threshold, the feature points experience visual failure. Kalman filtering is used to predict the motion state of the feature points, resulting in updated feature points. The updated feature points are then subjected to 3D reconstruction and deformation calculations to estimate the changes in the wing's sweep angle and span. Multi-threaded parallel processing enables high-speed processing from image acquisition to feature point matching and deformation calculation, achieving real-time data transmission of wing deformation data.

[0030] Specifically, such as Figure 5 As shown, the visual online monitoring module includes an image acquisition module, a feature point detection and matching module, a 3D reconstruction and deformation calculation module, a displacement visualization module, and a serial communication module.

[0031] The image acquisition module is mounted on the fuselage or tail of the aircraft, with a field of view covering the entire wing, and is used to acquire feature point images of the wing surface; the feature point detection and matching module is used to extract the center coordinates of the feature points and perform matching; the 3D reconstruction and deformation calculation module is used to perform 3D reconstruction on the extracted feature points, and can calculate the changes in the sweep angle and span of the wing before and after deformation.

[0032] The feature point detection and matching module includes two steps: first frame selection and template matching. Under normal visual conditions, when the error between the actual wing angle and the required angle corresponding to the detected feature point meets the preset error threshold, the feature point detection at the wing zero point position is first performed. That is, the image acquisition module first acquires an image of the wing zero point position, and then selects all feature points on the wing surface through manual interaction to obtain the integer pixel position of each feature point corner. Then, within a specified search window, the feature point corner position is refined from integer pixels to sub-pixel level using image grayscale gradient and iterative least squares optimization. A small area of ​​image region is extracted as the template image centered on this area for subsequent matching.

[0033] In subsequent images acquired by the image acquisition module, a search region of a certain radius is defined around the initial feature point corner as a reference, based on a template matching algorithm. For each pixel within the search region, an image region centered on that point and with the same size as the template image is selected as a candidate target, and its similarity is calculated with the template image to obtain the similarity value for each pixel, ultimately generating a matching score map. Then, the position with the highest similarity is selected from the score map to initially determine the integer pixel coordinates of the feature point corner. Further, a fitting window is constructed within the locally high similarity region of the matching score map, and a quadratic surface fitting method is used to model the similarity distribution, thereby accurately estimating the maximum matching position and achieving sub-pixel-level matching accuracy. Finally, the coordinates of the initial reference feature point corner are updated to the coordinates of the pixel with the highest score in the current frame for use in the matching of the next frame.

[0034] Each frame of the image can obtain the sub-pixel level image coordinates of all feature points and corners through the feature point detection and matching module. Then, the world coordinates of all feature points and corners are obtained through the 3D reconstruction and deformation calculation module. Finally, deformation calculation is performed using the obtained world coordinates of the feature points and corners.

[0035] Specifically, the validity of a feature point is determined by assessing its confidence level. When the confidence level of a feature point exceeds the confidence threshold, the feature point is lost. The validity of a feature point is determined by assessing its confidence level. When a feature point is lost due to external disturbances, a motion state prediction method based on Kalman filtering is used to update the position and extrapolate the trajectory of the feature point to obtain the updated feature point.

[0036] The updated feature points are detected and tracked by template matching. After obtaining the sub-pixel level image coordinates of all updated feature point corners, 3D reconstruction and deformation calculation are performed on them to obtain the changes in the wing sweep angle and span in real time.

[0037] The displacement visualization module is used to convert the feature point displacements and wing morphology parameters output by the visual online monitoring module into graphical data and display them in real time as dynamic curves, so as to realize intuitive visualization monitoring of the wing deformation process.

[0038] The serial communication module is used to enable data interaction between the visual online monitoring module and the control execution module via serial communication.

[0039] The multi-sensor fusion module timestamps and synchronizes the changes in wing sweep angle and span when vision fails and when vision is normal, and then fuses multi-source information to obtain state estimates of wing sweep angle and span. This fusion design can effectively suppress random noise and systematic bias caused by a single sensor, and improve the stability and reliability of state estimation.

[0040] By using an encoder to record the number of rotations of the stepper motor and servo motor, the theoretical rotation angle of the inner wing section and the expansion / contraction change of the outer wing section relative to the inner wing section can be calculated separately.

[0041] The stepper motor can be replaced by any device capable of linear output, such as a linear actuator, and the encoder can be any sensor that records the output of the linear output device.

[0042] The control execution module includes control commands and a variator wing. It uses state estimates, the aircraft's current flight attitude information, and the error between the actual and required wing angles as inputs to a fuzzy PID controller. By identifying errors and triggering corrections, it adaptively adjusts the sweep angle and span of the aircraft at different flight stages, generating corresponding control commands (when the aircraft is detected to be climbing, a control command with a large sweep angle and small span is generated). The variator wing includes a linear motion unit and a rotary motion unit. The linear motion unit drives a joystick to achieve variable sweep according to the control commands from the control execution module, and uses a stepper motor to drive a lead screw to rotate the variator wing around its root to achieve variable sweep motion. The rotary motion unit achieves variable span according to the control commands from the control execution module, using a continuously rotating servo to drive a steel cable to extend and retract the outer wing section along the wingspan direction to achieve span changes.

[0043] The control execution module generates control commands based on the flight attitude information measured by the attitude measurement module. It employs a fuzzy PID controller, using the flight attitude information acquired by the IMU as an auxiliary input to dynamically adjust the PID parameter k. p k i k d This enables attitude adaptive fuzzy PID control until the error between the actual wing angle and the required angle meets the error threshold.

[0044] When the error between the actual wing angle and the required wing angle caused by external disturbances exceeds the preset error threshold, a visual failure smoothing switching mechanism based on the confidence level of real-time acquisition of changes in wing sweep angle and span is used to update the position and extrapolate the trajectory of the feature point, resulting in the updated feature point.

[0045] The visual failure smooth switching mechanism includes a visual confidence determination unit, a prediction update unit, and a switching execution unit. The visual confidence determination unit determines the validity of visual data based on the number of feature points, matching residuals, and confidence thresholds. When visual feature points are lost due to external interference (lighting changes or partial occlusion), the prediction update unit uses a motion state prediction method based on Kalman filtering to update the position and extrapolate the trajectory of the visual feature points to obtain updated feature points to maintain continuous observation. The updated feature points are then sequentially reconstructed in three dimensions and their deformation is calculated to obtain changes in the wing's sweep angle and span in real time.

[0046] The visual failure smooth switching mechanism is responsible for improving the system's stability in complex environments such as changes in lighting, cloud cover, and smoke interference.

[0047] The visual online measurement module designed in this invention uses a non-contact visual measurement method to compensate for contact measurement methods such as strain gauges and accelerometers. It can perform full-field deformation observation without affecting the airfoil structure. It has a simple structure and is easy to maintain.

[0048] The variant aircraft wing shape perception and control system designed in this invention can realize real-time monitoring of the wing deformation process and visualization of feature point displacement.

[0049] Based on the Kalman filter algorithm, this invention aligns and synchronizes the sweep angle and span state variables of the wing acquired by the sensor and the estimated changes in the sweep angle and span of the wing, and then performs multi-source information fusion to establish a control-oriented wing aerodynamic shape parameter calculation model, which can achieve stable estimation of span and sweep angle parameters.

[0050] The visual failure smooth switching mechanism integrated in this invention can improve the stability of vision in complex environments such as changes in lighting, cloud cover, and smoke interference through real-time response and closed-loop control of feature point positions.

[0051] This invention automatically adjusts the wing shape using a fuzzy PID controller based on the aircraft's flight status information until the error between the actual wing angle and the required angle meets the error threshold. This enables stable estimation of the span and sweep angle parameters, improving the aerodynamic efficiency and mission adaptability of the aircraft.

[0052] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0053] Furthermore, unless otherwise stated, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. All references to this specification are incorporated by way of citation to disclose and describe methods relating to those references. In the event of any conflict with any incorporated reference, the content of this specification shall prevail.

Claims

1. A method for sensing and controlling the wing shape of a morphing aircraft, characterized in that, Includes the following steps: The positions of preset feature points on the inner and outer wing sections of the variant wing are obtained to determine the actual wing angle; the sweep angle and span of the wing are obtained through sensors mounted on the aircraft's rotation axis. When the error between the actual wing angle and the required angle exceeds the preset error threshold, the feature point becomes visually ineffective. The motion state of the feature point is predicted by Kalman filtering to obtain the updated feature point. The updated feature point is then subjected to three-dimensional reconstruction and deformation calculation in sequence, and the changes in the wing sweep angle and span are estimated. After aligning and synchronizing the state variables and changes of the wing sweep angle and span with timestamps and then fusing multi-source information, the state estimates of the wing sweep angle and span are obtained. The system acquires the current flight attitude information of the aircraft; using the state estimate, the current flight attitude information of the aircraft, and the error between the actual wing angle and the required angle as inputs to the controller, it adaptively adjusts the sweep angle and span until the error between the actual wing angle and the required angle meets the error threshold.

2. The method for sensing and controlling the wing shape of a morphing aircraft according to claim 1, characterized in that, The process of obtaining the positions of preset feature points on the inner and outer wing sections of the variant wing to determine the actual wing angle specifically includes: Extract the center coordinates of feature points at the zero-point position of the wing, acquire the first frame of the zero-point position image, and select all feature points on the wing surface through manual interaction; Using any feature point corner as a reference, a search area of ​​a certain radius is set around it. The similarity is calculated point by point through a sliding template to generate a matching score map. The position with the highest similarity in the matching score map is selected to determine the sub-pixel coordinates of the feature point corner. Based on the sub-pixel coordinates of the feature point corner points, a coordinate measuring machine is used to obtain the three-dimensional coordinates of the feature point in the wing coordinate system. After performing coordinate transformation on the three-dimensional coordinates, straight line fitting is performed on the feature points of the inner and outer wing sections respectively to obtain the corresponding straight line equations representing the directions of the inner and outer wing sections. By determining the included angle between the two straight lines, the actual angle of the wing is obtained.

3. The method for sensing and controlling the wing shape of a morphing aircraft according to claim 2, characterized in that, When the error between the actual wing angle and the required angle exceeds a preset error threshold, the feature point experiences visual failure. Kalman filtering is then used to predict the motion state of the feature point, resulting in an updated feature point. Specifically, this includes: Obtain the image of the feature point locations in each frame; By establishing a mapping relationship between the actual wing angle and the feature point position, feature point detection and matching are performed on the image of each frame of feature point position to obtain the sub-pixel level image coordinates of all feature point corners. Based on the subpixel-level image coordinates of all feature point corners, the confidence level of the feature points is determined to determine whether the feature points have experienced visual failure. When the confidence level of a feature point exceeds the confidence threshold, the feature point becomes visually ineffective. A motion state prediction method based on Kalman filtering is used to update the position and track the trajectory of the feature point that has become visually ineffective, and the updated feature point is obtained.

4. The method for sensing and controlling the wing shape of a morphing aircraft according to claim 3, characterized in that, The process of sequentially performing 3D reconstruction and deformation calculations on the updated feature points to obtain real-time changes in the wing's sweep angle and span specifically includes: Obtain the subpixel-level image coordinates of all feature point corners after the update, and obtain the world coordinates of all feature point corners through 3D reconstruction; Based on the updated world coordinates of all feature points and corner points, the feature points of the inner wing section and the outer wing section are evenly distributed in two rows. Each row of the inner wing section includes more than two feature points; each row of the outer wing section includes at least one feature point. Taking a certain feature point of a certain row of the selected inner wing section as a reference point, all feature points are projected onto the tangent plane of the wing surface where the reference point is located. The tangent plane is the z-axis plane. In the tangent plane, with the reference point as a straight line, a vector A is obtained by fitting a straight line to all feature points of the inner wing segment on the straight line; a vector B is obtained by fitting a least-squares straight line to all feature points of the other row of the inner wing segment; vectors A and B are averaged and normalized to obtain the wing's direction vector C along the wingspan direction; by calculating the angular change of the direction vector C, the change in the wing's sweep angle is obtained. The distance change of each feature point of the outer wing segment relative to all feature points of the inner wing segment on the tangent plane of the reference point is determined and averaged to obtain the amount of expansion and contraction of the outer wing segment relative to the inner wing segment, i.e., the amount of span change.

5. The method for sensing and controlling the wing shape of a morphing aircraft according to claim 1, characterized in that, The controller uses the state estimate, the aircraft's current flight attitude information, and the error between the actual and required wing angles as inputs to adaptively adjust the sweep angle and span until the error between the actual and required wing angles meets an error threshold. Specifically, this includes: The error between the actual and required wing angles, the state estimate, and the aircraft's flight attitude information are used as inputs to the fuzzy PID controller. When the error between the actual and required wing angles exceeds a preset error threshold, the PID parameter k is dynamically adjusted. p k i k d : When the aircraft's current flight attitude is in a rapid climb or dive phase, the fuzzy PID controller increases k... d and reduce k i The output control command is a large sweep angle and a small span; When the aircraft is currently in the cruise phase, the fuzzy PID controller increases k... i The output control command is a small sweep angle and a large span.

6. The method for sensing and controlling the wing shape of a morphing aircraft according to claim 1, characterized in that, The process of aligning and synchronizing the state variables and changes of the wing sweep angle and span using timestamps, followed by multi-source information fusion to obtain state estimates of the wing sweep angle and span, specifically includes: Align the wing sweep angle and span state variables and changes to a unified timestamp, and use linear interpolation or spline interpolation to process the asynchronous data to obtain the state variables and changes after timestamp alignment and synchronization. The state variables and changes after timestamp alignment and synchronization are fused with multi-source information, and the state prediction of the wing sweep angle and span is performed by the Kalman filter algorithm to obtain the state estimates of the wing sweep angle and span.

7. The method for sensing and controlling the wing shape of a morphing aircraft according to claim 1, characterized in that, The current flight attitude information of the aircraft is obtained by encoding through sensors mounted on the aircraft's rotation axis, including pitch angle, roll angle, angular velocity, and acceleration.

8. A wing shape sensing and control system for a morphing aircraft, characterized in that, include: The attitude measurement module is used to acquire the current flight attitude information of the aircraft; The visual online monitoring module is used to acquire the positions of preset feature points on the inner and outer wing sections of the variant wing to determine the actual wing angle; it acquires the wing's sweep angle and span state variables through sensors mounted on the aircraft's rotation axis; when the error between the actual wing angle and the required angle exceeds a preset error threshold, the feature points experience visual failure, and the motion state of the feature points is predicted using Kalman filtering to obtain updated feature points; the updated feature points are then subjected to 3D reconstruction and deformation calculations to estimate the changes in the wing's sweep angle and span. The multi-source information fusion module is used to align and synchronize the state variables and changes of the wing sweep angle and span after time stamping, and then perform multi-source information fusion to obtain the state estimates of the wing sweep angle and span. The control execution module is used to adaptively adjust the sweep angle and span using the state estimate, the current flight attitude information of the aircraft, and the error between the actual wing angle and the required angle as inputs to the controller, until the error between the actual wing angle and the required angle meets the error threshold.

Images