Method, system and device for accurate positioning of three-dimensional adjustable connection of flight simulator and storage medium

Abstract

The application relates to the field of flight simulation control and precision measurement, and provides a precise positioning method, system and device of a three-dimensional adjustable connecting piece of a flight simulator and a storage medium, and solves the problems of flight simulation control and precision measurement.The method comprises the following steps: acquiring real-time stroke length, rotation angle change data and connecting piece vibration images of a motion control rod, outputting displacement prediction parameters through a platform kinematics model, and generating three-dimensional point cloud data; three-dimensional visual coordinates are generated through sub-pixel recognition and distortion compensation of the images through a marker point recognition algorithm; space pose offset is obtained by fusing the point cloud and visual data and processing through a Kalman filter; finally, displacement compensation instructions are generated, the connecting piece pose is adjusted through a closed-loop controller, and precise positioning of the connecting piece relative to a flight target reference coordinate system is realized.The application improves the precision and stability of the connecting piece pose control in a dynamic vibration environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of flight simulation control and precision measurement, and particularly relates to a precise positioning method, system and device for a three-dimensional adjustable connecting piece of a flight simulator and a storage medium. BACKGROUND

[0002] In the field of flight simulation training, the core component of the high-precision motion platform, the three-dimensional adjustable connecting piece, needs to maintain an accurate pose relationship with the aircraft model in a continuous vibration environment. Since the six-degree-of-freedom platform will produce complex motion when simulating air flow disturbance, the connecting piece is prone to millimeter-level deviation, which puts high requirements on the positioning system: sub-millimeter positioning accuracy needs to be achieved within a millisecond response time, while overcoming dynamic environmental factors such as optical obstruction and vibration interference.

[0003] At present, one existing scheme for this requirement adopts a real-time monitoring system based on a laser tracker. The system arranges high-reflectivity target balls on the surface of the connecting piece, acquires the spatial coordinates of the target balls in real time through multiple laser trackers, predicts the motion trajectory of the connecting piece in combination with the kinematic model of the platform, and finally drives the electric actuator for position correction through a PID controller. This scheme relies on high-precision laser ranging principle and closed-loop control architecture, and can achieve a certain degree of dynamic compensation.

[0004] However, this existing scheme has obvious limitations in actual application: the laser tracking system is sensitive to line-of-sight obstruction, and the target ball surface is prone to reflection signal attenuation in a vibration environment; there is a modeling error between the kinematic model of the platform and the measured data, resulting in a deviation between the predicted trajectory and the actual displacement; the single sensing mode has limited anti-interference capability in a complex vibration environment, ultimately affecting the accuracy and stability of the pose compensation. SUMMARY

[0005] The present application provides a precise positioning method, system, device and storage medium for a three-dimensional adjustable connecting piece of a flight simulator, to solve the problem of flight simulation control and precision measurement in the prior art.

[0006] To solve the above technical problems, in a first aspect, the present application provides a precise positioning method for a three-dimensional adjustable connecting piece of a flight simulator, comprising:

[0007] acquiring real-time stroke length, instantaneous angle change data of a motion control lever in a flight target, and vibration interference image data of a three-dimensional adjustable connecting piece;

[0008] inputting the real-time stroke length and the instantaneous angle change data into a pre-trained kinematic model of the platform, outputting displacement prediction parameters of the flight target, and generating three-dimensional coordinate point cloud data of the three-dimensional adjustable connecting piece based on the displacement prediction parameters;

[0009] The vibration interference image data is sub-pixel recognized and optical distortion compensated by using a marker point recognition algorithm to generate three-dimensional visual coordinate data;

[0010] Based on the three-dimensional coordinate point cloud data and the three-dimensional visual coordinate data, a preset Kalman filter is used for fusion processing to output a spatial pose offset of the three-dimensional adjustable connecting piece;

[0011] According to the spatial pose offset, a displacement compensation instruction is generated, and the displacement compensation instruction is input to a closed-loop position controller of the flight target, and the spatial pose of the three-dimensional adjustable connecting piece is adjusted through the closed-loop position controller to realize accurate positioning of the three-dimensional adjustable connecting piece relative to the reference coordinate system of the flight target.

[0012] Optionally, based on the three-dimensional coordinate point cloud data and the three-dimensional visual coordinate data, a preset Kalman filter is used for fusion processing to output a spatial pose offset of the three-dimensional adjustable connecting piece, comprising:

[0013] The three-dimensional coordinate point cloud data is converted into a first observation vector, and the three-dimensional visual coordinate data is converted into a second observation vector;

[0014] The first observation vector and the second observation vector are synchronously input to the Kalman filter, wherein the Kalman filter takes the displacement prediction parameter as a state transition constraint;

[0015] Through a time update unit of the Kalman filter, a state estimation value at the next time is predicted based on the state transition constraint;

[0016] Through a measurement update unit of the Kalman filter, based on the state estimation value, the first observation vector and the second observation vector, a spatial pose offset of the three-dimensional adjustable connecting piece is generated.

[0017] Optionally, based on the state estimation value, the first observation vector and the second observation vector, the spatial pose offset of the three-dimensional adjustable connecting piece is generated, comprising:

[0018] Based on the first observation vector and the second observation vector, a Kalman gain matrix is calculated by an adaptive weight distribution network in the measurement update unit;

[0019] Based on the Kalman gain matrix, the state estimation value is weighted and corrected to obtain a corrected state estimation value;

[0020] The position deviation component and the angle deviation component are parsed from the corrected state estimation value;

[0021] motion smoothing filtering is respectively performed on the position deviation component and the angle deviation component;

[0022] The filtered position deviation component and the filtered angle deviation component are combined to generate a spatial pose offset of the three-dimensional adjustable connecting piece.

[0023] Optionally, based on the first observation vector and the second observation vector, a Kalman gain matrix is calculated by an adaptive weight allocation network in a measurement update unit, comprising:

[0024] The first observation vector and the second observation vector are input into an adaptive weight allocation network, and an environmental disturbance coefficient is generated by an environment perception module of the adaptive weight allocation network;

[0025] The feature importance scores of the first observation vector and the second observation vector are respectively calculated by an attention mechanism module of the adaptive weight allocation network;

[0026] The real-time confidence weights of the first observation vector and the second observation vector are respectively generated in combination with the environmental disturbance coefficient and the feature importance scores;

[0027] Based on the real-time confidence weights, a Kalman gain matrix is generated by a Kalman gain calculation formula.

[0028] Optionally, the sub-pixel recognition and optical distortion compensation of the vibration interference image data are performed by using a marker point recognition algorithm to generate three-dimensional visual coordinate data, comprising:

[0029] The camera internal parameters, lens distortion coefficients, spatial position relationship and attitude parameters of a binocular vision system in a flight target are obtained;

[0030] The pixel area of a visual marker point is extracted from the vibration interference image data, the gray distribution barycenter of the pixel area is calculated, and the gray distribution barycenter is taken as a sub-pixel level coordinate position;

[0031] Based on the camera internal parameters and the lens distortion coefficients, an optical distortion compensation model is established;

[0032] The sub-pixel level coordinate position is input into the optical distortion compensation model, and the sub-pixel level coordinate position is nonlinearly transformed by polynomial coefficients stored in the optical distortion compensation model to output a corrected coordinate after eliminating lens distortion;

[0033] According to the spatial position relationship and the attitude parameters, a stereo matching algorithm is used to calculate the horizontal position difference value of the corrected coordinate in left and right images of the binocular vision system, and disparity information is generated based on the horizontal position difference value;

[0034] Based on the parallax information, three-dimensional visual coordinate data of the visual marker points in the camera coordinate system is generated by the principle of triangulation.

[0035] Optionally, the generating the three-dimensional coordinate point cloud data of the three-dimensional adjustable connecting piece based on the displacement prediction parameter comprises:

[0036] The displacement prediction parameter is converted into a space scanning trigger instruction of a plurality of time difference ranging base stations, and the space scanning trigger instruction is used to control the plurality of time difference ranging base stations to emit ultra-wideband signals according to a preset time sequence;

[0037] The echo signal of the ultra-wideband signal reflected by the surface of the three-dimensional adjustable connecting piece is received, and the signal propagation time difference of each time difference ranging base station is measured based on the echo signal;

[0038] The signal propagation path length is calculated according to the signal propagation time difference;

[0039] Based on the signal propagation path length, a three-dimensional coordinate set of the surface reflection points of the three-dimensional adjustable connecting piece is generated in combination with the spatial geometric distribution relationship of each time difference ranging base station;

[0040] The three-dimensional coordinate set is subjected to spatial clustering processing based on the spatial position change amount in the displacement prediction parameter, to form three-dimensional coordinate point cloud data.

[0041] Optionally, the generating displacement compensation instructions according to the spatial pose offset and inputting the displacement compensation instructions into a closed-loop position controller of the flight target, and adjusting the spatial pose of the three-dimensional adjustable connecting piece through the closed-loop position controller, comprises:

[0042] The spatial pose offset is decomposed into linear displacement components in three orthogonal directions and three rotation angle components;

[0043] The displacement compensation amount required for each motion axis is calculated according to the linear displacement component and the rotation angle component in combination with the motion characteristics of the electrically controlled displacement mechanism;

[0044] The displacement compensation amount is converted into displacement compensation instructions in the form of pulse control signals;

[0045] The displacement compensation instructions are input into a digital signal processing unit of the closed-loop position controller, the displacement compensation instructions are adjusted in real time through the digital signal processing unit, and the spatial pose of the three-dimensional adjustable connecting piece is adjusted by a servo motor of the electrically controlled displacement mechanism based on the adjusted displacement compensation instructions.

[0046] In a second aspect, the application provides a precise positioning system for a three-dimensional adjustable connecting piece of a flight simulator, comprising:

[0047] an acquisition module, configured to acquire real-time stroke length, instantaneous rotation angle change data of a motion control stick in a flight target, and vibration interference image data of a three-dimensional adjustable connecting piece;

[0048] an input module, configured to input the real-time stroke length and the instantaneous rotation angle change data into a pre-trained platform kinematics model, output displacement prediction parameters of the flight target, and generate three-dimensional coordinate point cloud data of the three-dimensional adjustable connecting piece based on the displacement prediction parameters;

[0049] a compensation module, configured to perform sub-pixel identification and optical distortion compensation on the vibration interference image data by using a marker point identification algorithm, and generate three-dimensional visual coordinate data;

[0050] a fusion module, configured to perform fusion processing on the three-dimensional coordinate point cloud data and the three-dimensional visual coordinate data by using a preset Kalman filter, and output a spatial pose offset of the three-dimensional adjustable connecting piece;

[0051] a generation module, configured to generate a displacement compensation instruction according to the spatial pose offset, input the displacement compensation instruction into a closed-loop position controller of the flight target, adjust the spatial pose of the three-dimensional adjustable connecting piece by using the closed-loop position controller, and realize accurate positioning of the three-dimensional adjustable connecting piece relative to a reference coordinate system of the flight target.

[0052] In a third aspect, the present application provides an electronic device, comprising:

[0053] a memory, configured to store a computer program;

[0054] a processor, configured to implement the steps of the accurate positioning method of a three-dimensional adjustable connecting piece of a flight simulator according to the first aspect when the computer program is executed.

[0055] In a fourth aspect, the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program can implement the steps of the accurate positioning method of a three-dimensional adjustable connecting piece of a flight simulator according to the first aspect when the computer program is executed by a processor.

[0056] The technical scheme provided by the present application has the following beneficial effects:

[0057] The application realizes the synchronous collection of motion control instructions and visual monitoring data, provides multi-source input guarantee for subsequent fusion processing. The mechanical motion parameters are converted into spatial prediction data, and a mathematical model of the motion trajectory of the connecting piece is established. The coordinate extraction accuracy of optical measurement in a vibration environment is improved, and the influence of lens distortion on positioning accuracy is eliminated. Through multi-source data complementary optimization, the error of a single sensing method is suppressed, and the reliability of pose estimation is improved. A complete control closed loop from detection to execution is formed to ensure that the connecting piece tracks the target trajectory in real time and maintains accurate positioning.

[0058] Further, the application also converts the point cloud data and visual coordinate data into observation vectors, inputs a Kalman filter with displacement prediction parameters as state constraints, and then performs state prediction in the time update stage and data fusion in the measurement update stage, and finally outputs the spatial pose offset of the connecting piece.

[0059] And, through the dynamic fusion of kinematic constraints and multi-source sensing data, the measurement noise in the vibration environment is effectively overcome, and the calculation accuracy of the pose offset and the system anti-interference ability are improved.

[0060] These aspects or other aspects of the application will be more apparent in the following description of the embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0061] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0062] Figure 1 A flowchart of a precise positioning method of a three-dimensional adjustable connecting piece of a flight simulator provided by the embodiment of the application;

[0063] Figure 2 A specific implementation schematic diagram of a precise positioning method of a three-dimensional adjustable connecting piece of a flight simulator provided by the embodiment of the application;

[0064] Figure 3 Another specific implementation schematic diagram of a precise positioning method of a three-dimensional adjustable connecting piece of a flight simulator provided by the embodiment of the application;

[0065] Figure 4 A structure schematic diagram of a precise positioning system of a three-dimensional adjustable connecting piece of a flight simulator provided by the embodiment of the application. DETAILED DESCRIPTION

[0066] The existing positioning scheme based on laser tracker has limitations in the vibration environment of flight simulator: the laser measurement is extremely sensitive to line-of-sight obstruction, the reflection signal of the target ball surface is prone to attenuation under severe vibration; there is inherent modeling error between the platform kinematic model and the measured data, resulting in systematic deviation between the predicted trajectory and the actual displacement; the single sensing mode has limited anti-interference ability in a complex vibration environment, ultimately affecting the accuracy and stability of the pose compensation. These defects are essentially due to the singleness of the sensing method and the fragmented processing between the model and the sensing data.

[0067] To solve the above problems, the present application provides a precise positioning method for a three-dimensional adjustable connecting piece of a flight simulator, which realizes precise positioning by cooperatively processing mechanical motion parameters and visual perception data.

[0068] In order for those skilled in the art to better understand the present application, the present application will be further described in detail below in conjunction with the drawings and specific embodiments. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0069] The core of the present application is to provide a precise positioning method for a three-dimensional adjustable connecting piece of a flight simulator, and a flowchart of a specific embodiment thereof is shown in Figure 1 The method comprises:

[0070] Step 101: Obtain real-time stroke length, instantaneous angle change data of the motion control lever in the flight target, and vibration interference image data of the three-dimensional adjustable connecting piece.

[0071] In step 101, the real-time stroke length represents the straight-line distance data of the motion control lever movement, reflecting the magnitude of the control command. The instantaneous angle change data represents the instantaneous change amount of the rotation angle of the motion control lever, representing the direction and rate of the control command. The three-dimensional adjustable connecting piece and the motion control lever are linked through the motion control system of the flight target. The stroke length and the angle change data of the motion control lever are used to calculate the displacement prediction parameters through the platform kinematic model, which are used to guide the pose adjustment of the three-dimensional adjustable connecting piece, so that the spatial position of the connecting piece can respond to the operation instructions input by the operator through the motion control lever in real time, so as to maintain the precise relative position relationship between the connecting piece and the reference coordinate system of the flight target in the dynamic vibration environment. The vibration interference image data represents the image containing visual marker points collected by the binocular vision system in the dynamic vibration environment, which is affected by mechanical vibration and optical distortion.

[0072] In the embodiments of the present application, the displacement sensor and the angle sensor installed on the motion control lever collect the stroke length and the rotation angle change data in real time, and the binocular vision system synchronously collects the vibration interference image data on the surface of the connecting member. The displacement sensor records the linear displacement of the control lever, the angle sensor records the rotation angle change of the control lever, and the binocular vision system synchronously exposes the left and right cameras to capture the marker point image. All data are transmitted to the processing unit through the high-speed data bus to ensure time synchronization.

[0073] For example, in a certain flight simulator, the displacement sensor of the motion control lever collects the real-time stroke length of 120 mm, and the angle sensor collects the instantaneous rotation angle change data of 15 degrees. At the same time, the left and right cameras of the binocular vision system synchronously collect the surface image of the connecting member at a frequency of 100 frames per second, and the image resolution is set to 1280x1024 pixels. Due to platform vibration, motion blur and distortion appear in the image, forming vibration interference image data. All data are transmitted to the central processing unit through the gigabit Ethernet, and the timestamp error is less than 1 millisecond.

[0074] Step 102: inputting the real-time stroke length and the instantaneous rotation angle change data into the pre-trained platform kinematics model, outputting the displacement prediction parameter of the flight target, and generating the three-dimensional coordinate point cloud data of the three-dimensional adjustable connecting member based on the displacement prediction parameter.

[0075] In step 102, the platform kinematics model represents a mathematical model describing the relationship between the input of the motion control lever and the change of the platform pose, which is obtained by training historical data. The displacement prediction parameter represents the expected displacement and rotation of the platform in the three-dimensional space. The three-dimensional coordinate point cloud data represents the set of spatial coordinates of the points on the surface of the connecting member obtained by the time difference ranging technology.

[0076] In the embodiments of the present application, the real-time collected stroke length and rotation angle change data are input into the platform kinematics model, and the model calculates the displacement prediction parameter of the flight target through forward kinematics. The parameter includes the expected displacement of the platform in X, Y and Z directions and the rotation around the axes. Based on the displacement prediction parameter, the time difference ranging base station transmits the ultra-wideband signal, the signal is reflected on the surface of the connecting member and then collected by the receiver, the signal propagation time difference is calculated and converted into distance information, and the three-dimensional coordinate point cloud data is obtained by combining the spatial geometric relationship of the base station.

[0077] For example, after inputting the kinematics model of the platform movement with a stroke length of 120 mm and a rotation angle of 15 degrees, the output displacement prediction parameters are an X-direction displacement of 200 mm, a Y-direction displacement of -150 mm, and a Z-direction displacement of 50 mm. According to the parameters, four time-difference ranging base stations are triggered to sequentially emit ultra-wideband signals, and the propagation time differences of the base stations A, B, C, and D to the connecting piece are measured to be 15 ns, 16 ns, 14 ns, and 17 ns, respectively. The distance is calculated by the formula d = c x (tA-tC) / (tA-tB-tC-tD) 8 where d represents the straight-line distance from the base station to the connecting piece, in meters (m), c is the speed of light, 3 x 10 The time difference of the ultra-wideband signal propagating from the base station to the connecting piece is represented by tA-tC, tA-tB, tC-tD, and tA-tD, and the distance values of the base stations are obtained to be 2.25 m, 2.40 m, 2.10 m, and 2.55 m, respectively. Combined with the base station coordinates (base station A (0, 0, 0), base station B (2, 0, 0), base station C (0, 2, 0), and base station D (0, 0, 2)), three-dimensional coordinate point cloud data containing 150 points is generated by a trilateration algorithm, and the point cloud range is X: 1.95-2.35 m, Y: -1.20 to -0.90 m, and Z: 0.15-0.30 m.

[0078] Step 103: Sub-pixel recognition and optical distortion compensation of the vibration interference image data are performed by using a marker point recognition algorithm to generate three-dimensional visual coordinate data.

[0079] In step 103, sub-pixel recognition refers to a technology for improving the pixel-level positioning accuracy to sub-pixel level by an interpolation algorithm. Optical distortion compensation refers to a process of correcting image coordinate errors caused by lens distortion. Three-dimensional visual coordinate data refers to the spatial coordinates of the marker points in the camera coordinate system calculated by stereo vision.

[0080] In the embodiments of the present application, the pixel area of the visual marker point is extracted from the vibration interference image, and the sub-pixel level coordinates are calculated by calculating the gravity center of the area gray scale distribution. A distortion compensation model is established by the camera calibration parameters to perform nonlinear transformation on the sub-pixel coordinates to eliminate distortion errors. The corrected coordinates of the left and right cameras are stereoscopically matched to calculate the parallax information, and finally the three-dimensional coordinates of the marker points in the camera coordinate system are obtained by the principle of triangulation.

[0081] For example, the pixel area of the marker point (coordinate range X: 620-660 pixels, Y: 470-510 pixels) is extracted from the left image, and the sub-pixel coordinates (640.5, 490.2) are calculated by calculating the gravity center of the gray scale distribution. The corrected coordinates (639.8, 489.5) are obtained by the distortion compensation model. The corrected coordinates of the marker point in the right image are (600.2, 489.7). The baseline distance is 0.5 m, the focal length is 1200 pixels, and the parallax is 39.6 pixels. The three-dimensional coordinates of the marker point in the camera coordinate system are calculated by the triangulation formula The calculation depth (wherein Z represents the depth coordinate of the marker point, B is the baseline, f is the focal length, and d is the parallax) is obtained Coordinate 15.15 meters, X coordinate wherein X represents the horizontal three-dimensional coordinate of the marker point, represents the corrected X direction pixel coordinate of the marker point in the left image, and the following is obtained Y coordinate wherein Y represents the vertical three-dimensional coordinate of the marker point, represents the corrected Y direction pixel coordinate of the marker point in the left image, and the following is obtained The three-dimensional visual coordinate data of the eight marker points is generated, and the coordinate variance is 0.12 mm.

[0082] Step 104: Based on the three-dimensional coordinate point cloud data and the three-dimensional visual coordinate data, a preset Kalman filter is used for fusion processing, and a spatial pose offset of the three-dimensional adjustable connecting piece is output.

[0083] In step 104, the Kalman filter represents an optimization estimation algorithm that fuses multi-source data through a recursive algorithm. The spatial pose offset represents the deviation amount of the actual pose of the connecting piece from the theoretical pose in position and attitude.

[0084] In the embodiments of the present application, the point cloud data is converted into a first observation vector, and the visual coordinate data is converted into a second observation vector. The two observation vectors are input into the Kalman filter, and the filter takes the displacement prediction parameter as the state transition constraint. In the time update stage, the state transition matrix is used to predict the state at the next time, and in the measurement update stage, the state estimation is updated combined with the observation vector and the noise characteristics, and finally the pose offset is extracted from the state estimation value.

[0085] For example, the point cloud data is converted into an observation vector , and the visual coordinate data is converted into an observation vector . The state transition matrix F is set as the unit matrix, and the process noise covariance wherein represents the process noise covariance matrix, represents the unit matrix, and the observation noise covariance R=diag(0.0004, 0.0004, 0.0004). The gain matrix is calculated by the Kalman gain formula wherein K represents the Kalman gain matrix, represents the state estimation covariance matrix at the kth time, H is the observation matrix, represents the transpose matrix of the observation matrix H, and R represents the observation noise covariance matrix), and K=[0.75, 0, 0; 0, 0.75, 0; 0, 0, 0.75] is obtained. After updating the state estimation value, the following is obtained . Compared with the theoretical pose The offset is obtained by comparison Meters.

[0086] Step 105: generating displacement compensation instructions according to the spatial pose offset, and inputting the displacement compensation instructions into a closed-loop position controller of the flight target, adjusting the spatial pose of the three-dimensional adjustable connecting piece through the closed-loop position controller to realize accurate positioning of the three-dimensional adjustable connecting piece relative to the reference coordinate system of the flight target.

[0087] In step 105, the displacement compensation instructions represent control signals for driving the actuator to perform pose correction. The closed-loop position controller represents a system that realizes accurate position control through feedback regulation. The spatial pose is obtained by fusing the binocular vision three-dimensional coordinate and the time difference ranging point cloud data, and through double-source Kalman filtering processing constrained by the flight target kinematic model, specifically representing the real-time position (X / Y / Z coordinates) and attitude (pitch / roll / roll angle) of the connecting piece in the flight target reference coordinate system, used to describe the accurate spatial relationship of the connecting piece relative to the flight target in three-dimensional space.

[0088] In the embodiments of the present application, the pose offset is decomposed into displacement and rotation of each motion axis, and the required pulse control signal is calculated according to the actuator characteristics. The pulse signal is input into a digital signal processing unit to drive the servo motor to run and detect the actual displacement through a position sensor in real time, and feedback to the controller to form a closed-loop control.

[0089] For example, the offset is converted to X-axis-8000 pulses, Y-axis-8000 pulses, and Z-axis-8000 pulses (pulse equivalent 5 microns / pulse). The pulse frequency is 1 kHz, and the pulse width is 20 microseconds. After the servo motor runs, the position sensor detects the actual displacement Meters, which is fed back to the controller for fine tuning. After 3 iterations, the displacement error is less than 0.001 meters, realizing accurate positioning.

[0090] The method realizes sub-millimeter positioning accuracy of the three-dimensional adjustable connecting piece in a vibration environment through effective combination of multi-source data synchronous acquisition, kinematic prediction, visual measurement, data fusion, and closed-loop control. It overcomes the limitations of single sensing method, improves the anti-interference ability and positioning stability of the system, and ensures the reliability and fidelity of flight simulation training.

[0091] In order to solve the problem of insufficient fusion accuracy of multi-source sensing data in a vibration environment, in some embodiments, step 104: based on the three-dimensional coordinate point cloud data and the three-dimensional visual coordinate data, a preset Kalman filter is used for fusion processing to output the spatial pose offset of the three-dimensional adjustable connecting piece, as shown in Figure 2 ​

[0092] Step 201: converting the three-dimensional coordinate point cloud data into a first observation vector, and converting the three-dimensional visual coordinate data into a second observation vector.

[0093] In step 201, the first observation vector is a spatial position feature vector obtained by averaging the three-dimensional coordinate point cloud data, representing the estimated coordinates of the connector center point measured by the time difference ranging system. The second observation vector is a visual measurement feature vector obtained by coordinate conversion and filtering processing of the three-dimensional visual coordinate data, representing the center coordinates of the connector surface marker calculated by the binocular vision system.

[0094] In the embodiments of the present application, the arithmetic mean of all point coordinates in the three-dimensional coordinate point cloud data generated by the time difference ranging system is calculated to obtain the three-dimensional coordinates of the point cloud center point, which is taken as the three components of the first observation vector. At the same time, the multiple sets of three-dimensional visual coordinate data generated by the binocular vision system are processed by a weighted average algorithm to obtain the visual measurement center point coordinates, which are taken as the three components of the second observation vector. The two observation vectors represent the spatial position information obtained by two different measurement methods respectively.

[0095] Step 202: synchronously inputting the first observation vector and the second observation vector into the Kalman filter, wherein the Kalman filter takes the displacement prediction parameter as a state transition constraint.

[0096] In step 202, the state transition constraint is a system state change rule constructed based on the displacement prediction parameter output by the platform kinematics model, which is used to constrain the evolution process of the state vector in the Kalman filter.

[0097] In the embodiments of the present application, the first observation vector and the second observation vector are simultaneously input into the pre-configured Kalman filter program, and the state transition matrix of the filter is dynamically adjusted according to the displacement prediction parameter output by the platform kinematics model. The displacement prediction parameter provides prior knowledge of system motion as a constraint condition in the prediction stage of the filter, ensuring that the state prediction conforms to the actual motion law.

[0098] Step 203: predicting the state estimation value at the next time based on the state transition constraint through the time update unit of the Kalman filter.

[0099] In step 203, the time update unit is a calculation module in the Kalman filter responsible for predicting the state at the next time according to the system model. The state estimation value is the optimal estimation of the current state of the system obtained by prediction calculation.

[0100] In the embodiments of the present application, the time update unit of the Kalman filter calculates the state estimation value at the next time by using the state transition constraint through the state prediction formula. The specific process is to perform matrix multiplication operation on the state transition matrix and the state estimation value at the last time, and then add the influence of the process noise to obtain the predicted state at the current time. This predicted state fuses the prior information provided by the kinematic model.

[0101] Step 204: generating the spatial pose offset of the three-dimensional adjustable connecting piece based on the state estimation value, the first observation vector and the second observation vector through the measurement update unit of the Kalman filter.

[0102] In step 204, the measurement update unit is a calculation module in the Kalman filter responsible for fusing observation data and updating state estimation.

[0103] In the embodiments of the present application, the measurement update unit of the Kalman filter fuses the state estimation value and the two observation vectors for calculation. First, the Kalman gain matrix is calculated, which determines the weight of the observation data in the state update. Then the gain matrix is used to weight and fuse the predicted state and the observation value to obtain the final state estimation value. Finally, the position and attitude deviation components are extracted from the state estimation value to form the spatial pose offset output.

[0104] The following is a specific example:

[0105] In the precise positioning process of the three-dimensional adjustable connection of the flight simulator, based on the three-dimensional coordinate point cloud data containing 150 points obtained from the foregoing embodiment, the point cloud range is 1.95-2.35 meters in the X direction, -1.20 to -0.90 meters in the Y direction, and 0.15-0.30 meters in the Z direction, the arithmetic mean of the coordinates of all points of the point cloud data is calculated to obtain the first observation vector in the form of transpose of 2.15 meters, -1.05 meters and 0.22 meters, and the three-dimensional visual coordinate data of the eight marker points generated based on the binocular vision system has a coordinate variance of 0.12 mm, and the weighted average of the visual coordinate data is calculated, and the weight is allocated according to the marker point confidence to obtain the second observation vector in the form of transpose of 2.18 meters, -1.02 meters and 0.20 meters. The first observation vector and the second observation vector are synchronously input into a preset Kalman filter, and the displacement prediction parameter 200 mm, -150 mm and 50 mm output by the platform kinematics model is used as the state transition constraint, wherein the displacement prediction parameter is converted into 0.20 m, -0.15 m and 0.05 m in meter unit, and is used to construct a state transition matrix. Through the time update unit of the Kalman filter, the state estimation value at the next time is predicted based on the state transition constraint, the state transition matrix is set to an identity matrix, the process noise covariance matrix is set to 0.01 times the identity matrix, and the state estimation value at the last time is initialized to the transpose form of the displacement prediction parameter 0.20 m, -0.15 m and 0.05 m, and the state prediction formula is used to calculate, wherein represents the current state estimation value, F represents the state transition matrix, and I is used to represent the identity matrix in the embodiment, represents the state estimation value at the last time, and the predicted state estimation value is . Through the measurement update unit of the Kalman filter, the spatial pose offset of the three-dimensional adjustable connection is generated based on the state estimation value, the first observation vector and the second observation vector. First, the Kalman gain matrix is calculated using the formula , wherein K represents the Kalman gain matrix without unit, represents the state prediction covariance matrix initial value 0.01 times the identity matrix without unit, H represents the observation matrix set to the identity matrix without unit, and R represents the observation noise covariance matrix set to the diagonal matrix with diagonal elements of 0.0004, 0.0004 and 0.0004 without unit. The numerical calculation is performed by substituting the values to obtain , and after simplification , so the gain matrix is . Then the state estimation value is updated using the formula , wherein represents the observation vector fusion value, represents the predicted state estimation value at the current time, K represents the Kalman gain matrix, and the example is , substitute to obtain, , . Compare the state estimate value with the theoretical pose , obtain , complete the Kalman filter fusion processing.

[0106] In the embodiments of the present application, through the collaborative fusion processing of multi-source observation data, the limitations of single sensing mode are effectively overcome, and the accuracy and reliability of pose measurement in the vibration environment are improved, providing a reliable data basis for subsequent precise control.

[0107] In order to solve the problem of insufficient fusion accuracy and stability of multi-source data in the vibration environment, in some embodiments, step 204: based on the state estimate value, the first observation vector and the second observation vector, the spatial pose offset of the three-dimensional adjustable connecting piece is generated, as shown in Figure 3 , including:

[0108] Step 301: based on the first observation vector and the second observation vector, the Kalman gain matrix is calculated through the adaptive weight distribution network in the measurement update unit.

[0109] In step 301, the adaptive weight distribution network is a neural network model that dynamically adjusts the calculation parameters according to the real-time environmental conditions and data quality. The Kalman gain matrix is a coefficient matrix used to balance the weights of the predicted value and the observed value in the state update.

[0110] In the embodiments of the present application, the adaptive weight distribution network in the measurement update unit receives the first observation vector and the second observation vector as input, analyzes the noise characteristics and confidence of the vector through the internal multi-layer perceptron structure, outputs the adaptive weight coefficients for each observation vector, then adjusts the observation noise covariance matrix according to these weight coefficients, and finally obtains the optimized Kalman gain matrix through the standard Kalman gain calculation formula.

[0111] Step 302: based on the Kalman gain matrix, the state estimate value is weighted and corrected to obtain a corrected state estimate value.

[0112] In step 302, the weighted correction is a process of weighting and adjusting the difference between the state prediction value and the observation value using the Kalman gain matrix. The corrected state estimate value is the optimal estimation value of the system state after data fusion optimization.

[0113] In the embodiments of the present application, the Kalman gain matrix is subjected to matrix multiplication operation with the difference between the observation vector and the state prediction value, to obtain a correction amount, and then the correction amount is added to the original state estimate value to obtain a corrected state estimate value closer to the true state. This process effectively integrates the data advantages of motion prediction and actual observation.

[0114] Step 303: Resolve the position deviation component and the angle deviation component from the corrected state estimate value.

[0115] In step 303, the position deviation component is the difference between the corrected state estimate value and the theoretical pose in the translational dimension. The angle deviation component is the difference between the corrected state estimate value and the theoretical pose in the rotational dimension.

[0116] In the embodiments of the present application, the first three elements of the corrected state estimate value are extracted as the position component, and the last three elements are extracted as the angle component. The corresponding components of the theoretical pose are subtracted to obtain the position deviation component and the angle deviation component, which directly reflect the pose error of the connecting member at the current time.

[0117] Step 304: Perform motion smoothing filtering on the position deviation component and the angle deviation component, respectively.

[0118] In step 304, motion smoothing filtering is a signal processing method using the sliding average principle to eliminate high-frequency noise and jitter in the measurement data. The filtered components are stable signal outputs after smoothing processing.

[0119] In the embodiments of the present application, a first-order low-pass filter is used to process the position deviation component and the angle deviation component, respectively. The smoothing degree is controlled by adjusting the time constant of the filter, the effective low-frequency trend signal is retained, and the high-frequency noise component is filtered out, so that the output deviation signal is more stable and reliable.

[0120] Step 305: Combine the filtered position deviation component and the filtered angle deviation component to generate the spatial pose offset of the three-dimensional adjustable connecting member.

[0121] In the embodiments of the present application, the filtered position deviation component and the filtered angle deviation component are spliced and combined in a predetermined order. The position component is in the first three dimensions, and the angle component is in the last three dimensions, forming a complete six-dimensional spatial pose offset. This offset is used as the final output for pose correction of the control system.

[0122] The following is a specific example:

[0123] In the precise positioning process of the three-dimensional adjustable connecting member of the flight simulator, based on the obtained final state estimate value meters and the theoretical pose meters, first calculate the Kalman gain matrix through the adaptive weight distribution network in the measurement update unit. The network receives the first observation vector meters and the second observation vector Using millimeters as input, the data quality is analyzed, and a weighted coefficient vector [0.7, 0.3] is output. The weight of the first observation vector (0.7) is derived from the stability of the point cloud data, and the weight of the second observation vector (0.3) is calculated based on the visual coordinate variance of 0.12 millimeters. This is then processed using the formula... It equals W multiplied by the Kalman gain matrix of the previous time step, where W represents the weight coefficient vector, and the Kalman gain matrix of the previous time step is... The new Kalman gain matrix is ​​calculated as follows: Based on this Kalman gain matrix, the state estimates are weighted and corrected using the formula... ,in This represents the corrected state estimate. This represents the new Kalman gain matrix. This represents the state estimate before correction. The theoretical pose coordinates are used to obtain the corrected state estimate. The positional deviation component is extracted from this estimate. The angular deviation component is calculated using the spatial geometric relationship of the marked points on the connector surface. The position deviation component is subjected to motion smoothing filtering using a first-order low-pass filter formula. ,in This represents the current filtered output value. Represents the filter coefficients. This represents the input value at the current moment. This represents the filtered output value at the previous moment, calculated as follows: The angular deviation component is also processed in the same way. Finally, the filtered position deviation components Meter and angular deviation components Degree combination generates spatial pose offset of three-dimensional adjustable connector. This completes the calculation of spatial pose offset.

[0124] In this embodiment, the accuracy and stability of pose offset calculation are effectively improved through the coordinated processing of adaptive weight allocation and motion smoothing filtering, ensuring that smooth and reliable pose compensation data can still be obtained in a vibration environment, providing an important guarantee for precise control.

[0125] To address the issue of unstable fusion accuracy caused by dynamic changes in the quality of observation data under vibration environments, in some embodiments, step 301: calculating the Kalman gain matrix based on the first observation vector and the second observation vector through an adaptive weight allocation network in the measurement update unit includes:

[0126] Step 401: input the first observation vector and the second observation vector into an adaptive weight allocation network, and generate an environmental interference coefficient through an environment perception module of the adaptive weight allocation network.

[0127] In step 401, the environment perception module is a calculation unit responsible for analyzing external environmental conditions in the adaptive weight allocation network. The environmental interference coefficient is a parameter that quantifies the degree of influence of the current environment on the measurement data, with a value range of zero to one. The larger the value, the more serious the environmental interference.

[0128] In the embodiments of the present application, the environment perception module receives environmental monitoring data from sensors in real time, including vibration intensity, light change and electromagnetic interference level. These data are processed through a multi-layer neural network to output a comprehensive environmental interference coefficient, which reflects the overall influence of the current environment on the reliability of the observation data.

[0129] Step 402: calculate the feature importance scores of the first observation vector and the second observation vector respectively through an attention mechanism module of the adaptive weight allocation network.

[0130] In step 402, the attention mechanism module is a calculation component in the adaptive weight allocation network for evaluating the importance of data features. The feature importance score is a score value indicating the reliability of each dimension of the observation vector.

[0131] In the embodiments of the present application, the attention mechanism module analyzes the first observation vector and the second observation vector respectively, calculates the data consistency, noise level and trend features of each vector, and calculates the feature importance scores of each vector through attention weights. The higher the score, the better the data quality of the observation vector.

[0132] Step 403: combine the environmental interference coefficient and the feature importance score to generate real-time confidence weights for the first observation vector and the second observation vector respectively.

[0133] In step 403, the real-time confidence weight is a weight value dynamically calculated according to the environmental conditions and data quality, which is used to determine the contribution of each observation vector in data fusion.

[0134] In the embodiments of the present application, the environmental interference coefficient and the feature importance score are comprehensively calculated. When the environmental interference is serious, the weight of the optical observation vector is appropriately reduced. When the data quality is high, the weight of the vector is correspondingly increased. Finally, the real-time confidence weights of the first observation vector and the second observation vector are generated.

[0135] Step 404: generate a Kalman gain matrix based on the real-time confidence weights through a Kalman gain calculation formula.

[0136] In step 404, the Kalman gain calculation formula is a mathematical expression for calculating the optimal estimation weight, which adjusts the traditional gain calculation process by incorporating real-time confidence weight.

[0137] In the embodiments of the present application, real-time confidence weight is introduced into the calculation process of the observation noise covariance matrix, and the adjusted noise covariance matrix is involved in the Kalman gain calculation, so that the gain matrix can dynamically adapt to environmental changes and data quality fluctuations, and finally output the optimized Kalman gain matrix.

[0138] The following is a specific example:

[0139] In the precise positioning process of the three-dimensional adjustable connector of the flight simulator, based on the obtained first observation vector m and the second observation vector m, first input the two observation vectors into the adaptive weight allocation network, generate the environmental disturbance coefficient through the environmental perception module of the network, collect vibration sensor data to obtain a vibration intensity of 3 levels, collect light sensor data to obtain a light change rate of 50 lux per second, and collect electromagnetic interference detector data to obtain an electromagnetic interference level of 10 decibels. Input these parameters into the three-layer neural network calculation module, wherein the vibration intensity weight is 0.5, the light change rate weight is 0.3, and the electromagnetic interference weight is 0.2. The calculation obtains , wherein dividing by 100 is to normalize to the range of 0-1; then calculate the feature importance scores of the two observation vectors through the attention mechanism module of the adaptive weight allocation network. The module first analyzes the data characteristics of the first observation vector, calculates its variance as 0.015, signal-to-noise ratio as 28 decibels, and stability index as 0.85, and calculates its feature importance score as through the attention weight formula. Similarly, the second observation vector is analyzed to obtain a variance of 0.012, a signal-to-noise ratio of 32 decibels, and a stability index of 0.9, and its feature importance score is calculated as ; then combine the environmental disturbance coefficient 0.185 and the feature importance score to generate the real-time confidence weight through the confidence weight calculation formula. The real-time confidence weight of the first observation vector is equal to , and the real-time confidence weight of the second observation vector is equal to , the normalized first observation vector weight is , and the second observation vector weight is ; finally, based on these real-time confidence weights, generate the Kalman gain matrix through the Kalman gain calculation formula.

[0140] In the embodiments of the present application, through the dual regulation of environmental perception and attention mechanism, the Kalman gain matrix can adapt to environmental changes and data quality fluctuations in real time, improving the accuracy and stability of data fusion in the vibration environment, and providing more reliable guarantee for pose estimation.

[0141] To solve the problem of insufficient visual measurement accuracy in the vibration environment, in some embodiments, step 103: the sub-pixel recognition and optical distortion compensation of the vibration interference image data are performed by using a marker point recognition algorithm, and three-dimensional visual coordinate data is generated, including:

[0142] Step 501: Obtain the camera internal parameters, lens distortion coefficients, spatial position relationship and attitude parameters of the binocular vision system in the flight target.

[0143] In step 501, the binocular vision system is fixedly installed around the motion environment of the flight target, and is used to collect image data of the three-dimensional adjustable connecting piece on the flight target in real time. The spatial position information of the connecting piece relative to the reference coordinate system of the flight target is obtained through the principle of stereo vision measurement. The camera internal parameters are a set of parameters describing the optical characteristics of the camera, including focal length, principal point coordinates and pixel size, etc. The lens distortion coefficient is a parameter quantifying the degree of lens optical distortion. The spatial position relationship is the relative position and distance relationship between the left and right cameras. The attitude parameter is the rotation angle of the camera coordinate system relative to the world coordinate system.

[0144] In the embodiments of the present application, these parameters are obtained through a pre-performed camera calibration process. The calibration process uses a high-precision calibration board to collect multiple groups of images, and calculates the internal parameters of the camera, the distortion coefficients, the baseline distance and the relative attitude angle between the two cameras through an optimization algorithm. All parameters are stored in a configuration file for subsequent processing and calling.

[0145] Step 502: Extract the pixel area of the visual marker point from the vibration interference image data, calculate the gray distribution barycenter of the pixel area, and take the gray distribution barycenter as the sub-pixel level coordinate position.

[0146] In step 502, the visual marker point is a specific pattern point formed by setting a high-reflectivity material on the surface of the three-dimensional adjustable connecting piece. These marker points have a pre-set geometric arrangement feature, which is used to provide stable recognition and positioning features in the image. The pixel area is a connected region in the image containing the visual marker point. The gray distribution barycenter is the weighted average position of all pixel gray values in the pixel area, which can provide sub-pixel level coordinate accuracy.

[0147] In the embodiment of the present application, first, threshold segmentation and connected region analysis are performed on the vibration interference image to identify all possible marker point regions, and then the weighted center position of the gray value of each region is calculated, which is accurate to a sub-pixel level and serves as the preliminary coordinate position of the marker point.

[0148] Step 503: An optical distortion compensation model is established based on the camera internal parameters and the lens distortion coefficients.

[0149] In step 503, the optical distortion compensation model is a mathematical model for correcting lens distortion, which describes the distortion characteristics by a polynomial function.

[0150] In the embodiment of the present application, a distortion correction model is constructed according to the camera internal parameters and the lens distortion coefficients, which contains compensation polynomials of radial distortion and tangential distortion, and the polynomial coefficients are determined by a calibration process and used to map the distorted coordinates to the undistorted coordinates.

[0151] Step 504: The sub-pixel level coordinate position is input into the optical distortion compensation model, the sub-pixel level coordinate position is nonlinearly transformed by the polynomial coefficients stored in the optical distortion compensation model, and the corrected coordinates after eliminating the lens distortion are output.

[0152] In step 504, the polynomial coefficients refer to the mathematical parameters in the optical distortion compensation model for describing the lens distortion characteristics, which are obtained by a camera calibration process and constitute the mathematical basis of the optical distortion compensation model together with the camera internal parameters and the lens distortion coefficients, but are not the camera internal parameters and the lens distortion coefficients themselves. The nonlinear transformation is a process of mapping and calculating the coordinates by a polynomial function. The corrected coordinates are the ideal coordinate positions after eliminating the distortion effects.

[0153] In the embodiment of the present application, the sub-pixel level coordinate position is input into the optical distortion compensation model, the stored polynomial coefficients are used for coordinate transformation calculation, and the accurate coordinate position after eliminating the distortion is obtained through iterative optimization, which is closer to the real physical projection relationship.

[0154] Step 505: According to the spatial position relationship and the attitude parameter, a horizontal position difference value of the corrected coordinates in the left and right images of the binocular vision system is calculated by using a stereo matching algorithm, and disparity information is generated based on the horizontal position difference value.

[0155] In step 505, the stereo matching algorithm is a method for finding corresponding points in left and right images. The left and right images are obtained by synchronously collecting left and right cameras installed in the binocular vision system around the flight target motion environment, corresponding to two image data containing the visual marker points collected from different angles at the same time. The horizontal position difference is the coordinate difference of the same marker point in the horizontal direction of the left and right images. The disparity information is the result of the horizontal position difference after standardization.

[0156] In the embodiment of the present application, according to the spatial position relationship and attitude parameters of the dual cameras, an epipolar constraint relationship is established, matching point pairs are found in the corrected coordinates of the left and right images, and the coordinate difference of the horizontal direction is calculated. This difference reflects the depth information of the object.

[0157] Step 506: Based on the disparity information, three-dimensional visual coordinate data of the visual marker points in the camera coordinate system is generated by the principle of triangulation.

[0158] In step 506, the principle of triangulation refers to observing the same target point by two cameras with known positions and attitudes at the same time, calculating the horizontal position difference of the target point in the left and right camera images, i.e. the parallax, combining the baseline distance and optical parameters between the cameras, and based on the similar triangle geometric relationship, the mathematical method for solving the three-dimensional space coordinates of the target point relative to the camera coordinate system.

[0159] In the embodiment of the present application, based on the disparity information and the camera internal parameters, the three-dimensional coordinates of the marker points are calculated by the triangulation formula. This calculation process uses the principle of similar triangles to convert two-dimensional image coordinates to three-dimensional space coordinates, and finally obtains accurate three-dimensional position information.

[0160] The following is a specific example:

[0161] In the precise positioning process of the three-dimensional adjustable connecting piece of the flight simulator, based on the vibration interference image data collected by the binocular vision system, the pre-calibrated camera internal parameters including focal length 1200 pixels, principal point coordinates 640 pixels and 480 pixels, lens distortion coefficients including radial distortion coefficients and , tangential distortion coefficients and are obtained., the spatial position relationship is that the baseline distance of left and right cameras is 0.5 meters, and the attitude parameter is that the rotation angle of the right camera relative to the left camera is zero degrees; a pixel region of the visual marker point is extracted from the vibration interference image data, the coordinate range of the region in the left image is 620 to 660 pixels and 470 to 510 pixels, and the corresponding region in the right image is 580 to 620 pixels and 470 to 510 pixels, the gray distribution barycenter of the pixel region of the left image is calculated to obtain the sub-pixel level coordinate positions 640.5 pixels and 490.2 pixels, and the gray distribution barycenter of the pixel region of the right image is calculated to obtain the sub-pixel level coordinate positions 600.2 pixels and 489.7 pixels; an optical distortion compensation model is established based on the camera internal parameters and the lens distortion coefficient, the sub-pixel level coordinate positions are input into the optical distortion compensation model, the left image coordinates 640.5 pixels and 490.2 pixels are corrected to obtain the left image correction coordinates 639.8 pixels and 489.5 pixels, and the right image coordinates are also processed to obtain 600.2 pixels and 489.7 pixels; according to the spatial position relationship and the attitude parameter, a stereo matching algorithm is used to calculate the horizontal position difference value of the correction coordinates in the left and right images, the difference value between the left image x coordinate 639.8 pixels and the right image x coordinate 600.2 pixels is , the difference value is the parallax information; based on the parallax information 39.6 pixels, three-dimensional visual coordinate data of the visual marker point in the camera coordinate system is generated by the triangulation principle, and the calculation result is , the calculation result is , the calculation result is , and finally the three-dimensional visual coordinate data of the visual marker point is 8.08 meters, 6.18 meters and 15.15 meters, and the conversion process from image data to three-dimensional coordinates is completed.

[0162] In the embodiments of the present application, through the complete image processing and three-dimensional reconstruction process, the influence of vibration interference and optical distortion is effectively overcome, and high-precision three-dimensional visual coordinate data is obtained, which provides a reliable visual measurement basis for subsequent data fusion and pose control.

[0163] In order to solve the problem of three-dimensional coordinate measurement precision and efficiency in a vibration environment, in some embodiments, step 102: generating three-dimensional coordinate point cloud data of the three-dimensional adjustable connecting piece based on the displacement prediction parameter, comprising:

[0164] Step 601: converting the displacement prediction parameter into a space scanning trigger instruction of a plurality of time difference ranging base stations, and controlling the plurality of time difference ranging base stations to emit ultra-wideband signals according to a preset time sequence by using the space scanning trigger instruction.

[0165] In step 601, the time difference of arrival base station is fixedly installed around the moving environment of the flight target, and the spatial position relationship of the connecting piece relative to the reference coordinate system of the flight target is determined in real time by measuring the time difference of arrival of the ultra-wideband signal from the base station to the three-dimensional adjustable connecting piece on the flight target, so as to provide positioning reference for the flight target. The space scanning trigger instruction is a time sequence control signal generated according to the displacement prediction parameter, which is used to coordinate the transmission time sequence of multiple time difference of arrival base stations. The ultra-wideband signal is a radio wave signal with a nanosecond-level pulse width, which is suitable for high-precision distance measurement.

[0166] In the embodiment of the present application, the spatial coordinate components in the displacement prediction parameter are converted into time delay signals, each coordinate component corresponds to the trigger delay of a base station, and the trigger instructions are sent to each time difference of arrival base station through the control bus. The base station transmits the ultra-wideband signal according to the received instruction in a predetermined time sequence.

[0167] Step 602: receiving the echo signal of the ultra-wideband signal reflected by the surface of the three-dimensional adjustable connecting piece, and measuring the signal propagation time difference of each time difference of arrival base station based on the echo signal.

[0168] In step 602, the echo signal is the electromagnetic wave signal returned after the ultra-wideband signal is reflected by the surface of the connecting piece. The signal propagation time difference is the time difference experienced by the signal from transmission to reception.

[0169] In the embodiment of the present application, the receiver of the time difference of arrival base station continuously monitors the reflected echo signal, accurately measures the arrival time of each signal through a correlation detection algorithm, calculates the difference between the transmission time and the reception time, and obtains accurate signal propagation time difference data.

[0170] Step 603: calculating the signal propagation path length according to the signal propagation time difference.

[0171] In step 603, the signal propagation path length is the actual distance traveled by the radio wave from transmission to reception.

[0172] In the embodiment of the present application, according to the principle that the propagation speed of electromagnetic wave is constant, the measured signal propagation time difference is multiplied by the speed of light and then divided by two to obtain the one-way propagation path length, which represents the straight-line distance from the base station to the surface of the connecting piece.

[0173] Step 604: generating a three-dimensional coordinate set of the surface reflection point of the three-dimensional adjustable connecting piece based on the signal propagation path length and the spatial geometric distribution relationship of each time difference of arrival base station.

[0174] In step 604, the geometric distribution relationship is obtained through accurate calibration of the base station installation position, which refers to the fixed coordinate position of multiple base stations in three-dimensional space and their relative orientation relationship, and is used to provide spatial geometric constraints when calculating signal propagation path length based on signal arrival time difference. The three-dimensional coordinate set is a set of surface point coordinate data calculated by space intersection.

[0175] In the embodiments of the present application, the path length data measured by multiple base stations is combined with the known base station coordinate position, and the spatial coordinates of the surface reflection points of the connecting member are calculated by a trilateration algorithm. Each base station measurement value determines a sphere, and the intersection of multiple spheres is the three-dimensional coordinates of the reflection points.

[0176] Step 605: based on the spatial position change amount in the displacement prediction parameter, performing spatial clustering processing on the three-dimensional coordinate set to form three-dimensional coordinate point cloud data.

[0177] In step 605, the spatial position change amount is calculated by kinematics forward solution according to the real-time stroke length and instantaneous angle change data of the motion control rod through the kinematics model, which reflects the position change trend of the flight target in three-dimensional space. Spatial clustering processing is a method of grouping coordinate data according to spatial proximity.

[0178] In the embodiments of the present application, according to the spatial position change trend provided by the displacement prediction parameter, the original coordinate set is subjected to density clustering analysis, points with similar spatial positions are classified into the same category, isolated noise points are removed, and point cloud data representing the surface features of the connecting member are extracted.

[0179] The following is a specific example:

[0180] In the precise positioning process of the three-dimensional adjustable connecting member of the flight simulator, based on the displacement prediction parameters X direction displacement 200 mm, Y direction displacement negative 150 mm, and Z direction displacement 50 mm output by the platform kinematics model, the displacement prediction parameters are first converted into spatial scanning trigger instructions of four time difference ranging base stations, wherein the X coordinate 200 mm corresponds to the base station A trigger delay 0 ms, the Y coordinate negative 150 mm corresponds to the base station B trigger delay 2 ms, the Z coordinate 50 mm corresponds to the base station C trigger delay 4 ms, and the base station D is used as a reference base station with a trigger delay of 6 ms. The spatial scanning trigger instructions are used to control the base stations to emit ultra-wideband signals in sequence; the echo signals of the received ultra-wideband signals reflected by the surface of the three-dimensional adjustable connecting member are used to measure the signal propagation time difference of each time difference ranging base station, and the base station A propagation time difference is 15 nanoseconds, the base station B propagation time difference is 16 nanoseconds, the base station C propagation time difference is 14 nanoseconds, and the base station D propagation time difference is 17 nanoseconds; the signal propagation path length is calculated according to the signal propagation time difference, and the formula The distance is calculated, wherein represents the first signal propagation path length of the first base station to the surface reflection point of the three-dimensional adjustable connector, represents the speed of light , represents the signal propagation time difference of the first base station, the path length of base station A is calculated , the path length of base station B is , the path length of base station C is , and the path length of base station D is ; based on the signal propagation path length and the spatial geometric distribution relationship of each time difference ranging base station, the base station coordinates are base station A (0, 0, 0), base station B (2, 0, 0), base station C (0, 2, 0), and base station D (0, 0, 2), respectively. A three-dimensional coordinate set of the surface reflection point of the three-dimensional adjustable connector is generated by a trilateration algorithm, and the formula is used in the algorithm, where is the coordinate to be solved, is the base station coordinate, and the equation set is solved to obtain 150 surface point coordinates, the coordinate range is X direction 1.95 to 2.35 meters, Y direction negative 1.20 to negative 0.90 meters, and Z direction 0.15 to 0.30 meters; based on the spatial position change amount in the displacement prediction parameter X direction 0.02 meters, Y direction 0.01 meters, and Z direction 0.03 meters, the three-dimensional coordinate set is processed by spatial clustering, a density clustering algorithm is adopted, a neighborhood radius of 0.05 meters is set, and the minimum number of points is 10. 150 points are clustered into 3 main cluster groups, the centroid coordinates of each cluster group are taken, and the final three-dimensional coordinate point cloud data is formed, including three core point coordinates , , , the point cloud data generation process is completed.

[0181] In the embodiments of the present application, through the time sequence control of the multi-base station cooperative measurement and the intelligent data processing, the accuracy and reliability of the three-dimensional coordinate measurement in the vibration environment are effectively improved, and a high-quality point cloud data basis is provided for subsequent pose calculation.

[0182] In order to solve the problem of accurate conversion of the pose offset to the control instruction of the actuator, in some embodiments, step 105: generating a displacement compensation instruction according to the spatial pose offset, and inputting the displacement compensation instruction to a closed-loop position controller of the flight target, adjusting the spatial pose of the three-dimensional adjustable connector through the closed-loop position controller, comprising:

[0183] Step 701: decompose the spatial pose offset into three orthogonal linear displacement components and three rotation angle components.

[0184] In step 701, the linear displacement component is the translation of the spatial pose offset in the directions of the three orthogonal coordinate axes. The rotation angle component is the rotation of the spatial pose offset around the three coordinate axes.

[0185] In the embodiments of the present application, the six-dimensional spatial pose offset is decomposed into the first three elements as linear displacement components, corresponding to the translation amounts of the X, Y, and Z axes, respectively, and the last three elements as rotation angle components, corresponding to the rotation angles around the X, Y, and Z axes, respectively. This decomposition facilitates subsequent independent control of each motion axis.

[0186] Step 702: According to the linear displacement component and the rotation angle component, and in combination with the motion characteristics of the electrically controlled displacement mechanism, the displacement compensation amount required for each motion axis is calculated.

[0187] In step 702, the electrically controlled displacement mechanism refers to an electromechanical execution device installed in the flight target structure for accurately adjusting the spatial pose of the three-dimensional adjustable connecting piece. By receiving the instructions of the closed-loop position controller, the servo motor is driven to generate precise displacement, and the pose deviation of the connecting piece relative to the reference coordinate system of the flight target is corrected in real time, ensuring that the connecting piece maintains precise synchronization with the motion trajectory of the flight target in a dynamic vibration environment. The motion characteristics of the electrically controlled displacement mechanism refer to the mechanical motion performance parameters exhibited by the mechanism under electrical signal control, including but not limited to the torque-speed characteristics of the servo motor, the transmission ratio and precision of the reduction mechanism, the stiffness and friction characteristics of the guide rail, the maximum stroke and resolution of each motion axis, the dynamic response characteristics of the mechanism, and the nonlinear error compensation parameters during the motion process, which are key physical characteristics affecting displacement accuracy and control effect. The displacement compensation amount is the actual distance or rotation angle that each motion axis needs to move.

[0188] In the embodiments of the present application, the displacement amount required for each linear motion axis is calculated based on the linear displacement component, and the rotation amount required for each rotary motion axis is calculated based on the rotation angle component in combination with the kinematic model of the mechanism, while considering the mechanical characteristics of the electrically controlled displacement mechanism such as lead screw pitch and reduction ratio, to ensure that the calculated displacement compensation amount is within the executable range of the mechanism.

[0189] Step 703: Convert the displacement compensation amount into a displacement compensation instruction in the form of a pulse control signal.

[0190] In step 703, the pulse control signal is a digital signal that represents the displacement amount through pulse quantity and frequency.

[0191] In the embodiments of the present application, the displacement compensation amount is converted into corresponding pulse number and pulse frequency. The pulse number is obtained by dividing the displacement amount by the pulse equivalent, and the pulse frequency is determined according to the motion speed requirement, generating a standardized pulse control signal as the displacement compensation instruction.

[0192] Step 704: input the displacement compensation instruction into the digital signal processing unit of the closed-loop position controller, adjust the displacement compensation instruction in real time through the digital signal processing unit, and drive the servo motor of the electric control displacement mechanism to adjust the spatial pose of the three-dimensional adjustable connecting piece based on the adjusted displacement compensation instruction.

[0193] In step 704, the digital signal processing unit is the core computing component of the closed-loop position controller. The servo motor is an electric mechanism that performs displacement operations.

[0194] In the embodiments of the present application, the pulse control signal is input into the digital signal processing unit, which monitors the execution state in real time and adjusts the pulse output according to the feedback signal to drive the servo motor to operate according to the instruction requirements, and drives the connecting piece to produce corresponding pose adjustment through the mechanical transmission mechanism.

[0195] The following is a specific example:

[0196] In the precise positioning process of the three-dimensional adjustable connecting piece of the flight simulator, based on the obtained spatial pose offset , first, the spatial pose offset is decomposed into three orthogonal linear displacement components and three rotation angle components, where the linear displacement components are X-axis negative 0.04 meters, Y-axis negative 0.04 meters, and Z-axis negative 0.04 meters, and the rotation angle components are calculated through the spatial geometric relationship of the connecting piece surface marker points to be 0.5 degrees around the X-axis, negative 0.3 degrees around the Y-axis, and 0.2 degrees around the Z-axis; according to the linear displacement components and the rotation angle components, the displacement compensation amount required by each motion axis is calculated in combination with the motion characteristics of the electric control displacement mechanism, where the electric control displacement mechanism adopts ball screw transmission, the lead of the screw is 5mm, which is equal to 0.005m, the encoder resolution of the servo motor is 10000 pulses per revolution, the X-axis displacement compensation amount is calculated, the Y-axis needs , , the Z-axis needs , the reduction ratio of the rotation mechanism is 10 to 1, the corresponding pulse number of 0.5 degrees around the X-axis rotation is calculated , the corresponding pulse number of negative 0.3 degrees around the Y-axis rotation is , and the corresponding pulse number of 0.2 degrees around the Z-axis rotation is ; the displacement compensation amount is converted into a displacement compensation instruction in the form of a pulse control signal, the pulse frequency is set to 1000 Hz, the pulse width is set to 20 microseconds, X-axis instruction negative 8000 pulses, Y-axis instruction negative 8000 pulses, Z-axis instruction negative 8000 pulses, rotation axis instruction 139 pulses, negative 83 pulses, 56 pulses are generated; the displacement compensation instruction is input to the digital signal processing unit of the closed-loop position controller, the unit uses a 32-bit processor, the sampling period is 1 millisecond, the displacement compensation instruction is adjusted in real time through the digital signal processing unit, fine adjustment is performed according to the actual displacement amount fed back by the position sensor, and the spatial pose of the three-dimensional adjustable connecting piece is adjusted based on the adjusted displacement compensation instruction. The servo motor of the electric control displacement mechanism drives the three-dimensional adjustable connecting piece, the rated speed of the servo motor is 3000 revolutions per minute, the torque is 5 newton-meters, first, X-axis negative 8000 pulses correspond to a displacement of negative 0.04 meters, Y-axis negative 8000 pulses correspond to a displacement of negative 0.04 meters, Z-axis negative 8000 pulses correspond to a displacement of negative 0.04 meters, then 139 pulses of the rotation axis correspond to a rotation of 0.5 degrees around the X-axis, negative 83 pulses correspond to a rotation of negative 0.3 degrees around the Y-axis, and 56 pulses correspond to a rotation of 0.2 degrees around the Z-axis. The position sensor detects that the actual displacement is negative 0.039 meters, negative 0.039 meters, and negative 0.039 meters, and the actual rotation angle is 0.49 degrees, negative 0.29 degrees, and 0.19 degrees. Feedback to the closed-loop position controller generates an adjustment instruction. After 3 iterations of adjustment, the displacement error is less than 0.001 meters, and the rotation angle error is less than 0.02 degrees. Finally, the precise positioning of the spatial pose of the three-dimensional adjustable connecting piece is realized.

[0197] In the embodiments of the present application, accurate instruction conversion and closed-loop control are used to realize accurate mapping from the pose offset to mechanical execution, ensure that the three-dimensional adjustable connecting piece can still be quickly and accurately adjusted to the target pose in a vibrating environment, and improve the positioning accuracy and stability of the system.

[0198] Figure 4 The structure diagram of the precise positioning system of the three-dimensional adjustable connecting piece of the flight simulator provided in the embodiments of the present application is described in the specific implementation part:

[0199] The acquisition module 41 is configured to acquire real-time stroke length, instantaneous angle change data of the motion control lever in the flight target, and vibration interference image data of the three-dimensional adjustable connecting piece.

[0200] The input module 42 is configured to input the real-time stroke length and the instantaneous angle change data into a pre-trained platform kinematics model, output displacement prediction parameters of the flight target, and generate three-dimensional coordinate point cloud data of the three-dimensional adjustable connecting piece based on the displacement prediction parameters.

[0201] The compensation module 43 is configured to perform sub-pixel identification and optical distortion compensation on the vibration interference image data by using a mark point identification algorithm, and generate three-dimensional visual coordinate data.

[0202] The fusion module 44 is configured to perform fusion processing on the three-dimensional coordinate point cloud data and the three-dimensional visual coordinate data by using a preset Kalman filter, and output a spatial pose offset of the three-dimensional adjustable connecting piece.

[0203] The generation module 45 is configured to generate a displacement compensation instruction according to the spatial pose offset, and input the displacement compensation instruction to a closed-loop position controller of the flight target, so as to adjust the spatial pose of the three-dimensional adjustable connecting piece by using the closed-loop position controller, and realize accurate positioning of the three-dimensional adjustable connecting piece relative to a reference coordinate system of the flight target.

[0204] The accurate positioning system of the flight simulator three-dimensional adjustable connecting piece according to the embodiments of the present application is used to realize the accurate positioning method of the flight simulator three-dimensional adjustable connecting piece as described above, and therefore the specific embodiments of the accurate positioning system of the flight simulator three-dimensional adjustable connecting piece can be seen from the foregoing embodiments of the accurate positioning method of the flight simulator three-dimensional adjustable connecting piece, and the specific embodiments can be referred to the descriptions of the corresponding embodiments of each part, which will not be described herein again.

[0205] The present application also provides an electronic device, which comprises a memory configured to store a computer program, and a processor configured to execute the computer program to implement the steps of the accurate positioning method of the flight simulator three-dimensional adjustable connecting piece.

[0206] The present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the accurate positioning method of the flight simulator three-dimensional adjustable connecting piece.

[0207] In an exemplary embodiment, the computer readable storage medium can include, but is not limited to, a U disk, a read-only memory, a random access memory, a mobile hard disk, a magnetic disk or an optical disk, and various media that can store computer programs.

[0208] The embodiments of the present application also provide a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the steps in the embodiments of the accurate positioning method of the flight simulator three-dimensional adjustable connecting piece.

[0209] Those skilled in the art will further realize that the mere concepts, teachings, and embodiments described herein are merely meant to provide an enabling description of embodiments of the present application and are not intended to limit the scope of the present application. Accordingly, embodiments as described herein contemplate all modifications that come within the scope of the present application as recited by the claims set forth below and any equivalents thereto, with the scope of the present application being measured by the broadest interpretation of those claims set forth below.

[0210] The above has carried out the detailed introduction to the precision positioning method, system, equipment and storage medium of the three-dimensional adjustable connecting piece of the flight simulator provided by the application. The principle and implementation mode of the application are described in the specific examples in this paper, and the above example description is only used to help understand the method and core idea of the application. It should be pointed out that for ordinary skilled in the art, without departing from the principle of the application, the application can be improved and modified, and these improvements and modifications also fall within the protection scope of the application.

Claims

1. A method for precision positioning of a three-dimensional adjustable connection of a flight simulator, characterized in that The method comprises the following steps: acquiring real-time stroke length, instantaneous rotation angle change data of a motion control lever in a flight target, and vibration interference image data of a three-dimensional adjustable connecting piece; inputting the real-time stroke length and the instantaneous rotation angle change data into a pre-trained platform kinematics model to output displacement prediction parameters of the flight target, and generating three-dimensional coordinate point cloud data of the three-dimensional adjustable connecting piece based on the displacement prediction parameters; performing sub-pixel recognition and optical distortion compensation on the vibration interference image data by using a marker point recognition algorithm to generate three-dimensional visual coordinate data; based on the three-dimensional coordinate point cloud data and the three-dimensional visual coordinate data, performing fusion processing by using a preset Kalman filter to output a spatial pose offset of the three-dimensional adjustable connecting piece; generating displacement compensation instructions according to the spatial pose offset, and inputting the displacement compensation instructions into a closed-loop position controller of the flight target to adjust the spatial pose of the three-dimensional adjustable connecting piece, so as to realize accurate positioning of the three-dimensional adjustable connecting piece relative to a reference coordinate system of the flight target.

2. The method of claim 1, wherein, The method of performing fusion processing by using a preset Kalman filter based on the three-dimensional coordinate point cloud data and the three-dimensional visual coordinate data to output a spatial pose offset of the three-dimensional adjustable connecting piece comprises the following steps: converting the three-dimensional coordinate point cloud data into a first observation vector, and converting the three-dimensional visual coordinate data into a second observation vector; synchronously inputting the first observation vector and the second observation vector into the Kalman filter, wherein the Kalman filter takes the displacement prediction parameters as a state transition constraint; predicting a state estimation value at a next time point based on the state transition constraint by using a time update unit of the Kalman filter; generating the spatial pose offset of the three-dimensional adjustable connecting piece based on the state estimation value, the first observation vector and the second observation vector by using a measurement update unit of the Kalman filter.

3. The method of claim 2, wherein, The method of generating the spatial pose offset of the three-dimensional adjustable connecting piece based on the state estimation value, the first observation vector and the second observation vector comprises the following steps: calculating a Kalman gain matrix by using an adaptive weight distribution network in the measurement update unit based on the first observation vector and the second observation vector; performing weighted correction on the state estimation value based on the Kalman gain matrix to obtain a corrected state estimation value; analyzing a position deviation component and an angle deviation component from the corrected state estimation value; respectively performing motion smoothing filtering processing on the position deviation component and the angle deviation component; combining the filtered position deviation component and the filtered angle deviation component to generate the spatial pose offset of the three-dimensional adjustable connecting piece.

4. The method of claim 3, wherein, The method of calculating a Kalman gain matrix by using an adaptive weight distribution network based on the first observation vector and the second observation vector comprises the following steps: inputting the first observation vector and the second observation vector into the adaptive weight distribution network, and generating an environmental interference coefficient by using an environment perception module of the adaptive weight distribution network. The attention mechanism module of the adaptive weight distribution network is used to calculate feature importance scores of the first observation vector and the second observation vector, respectively; The real-time confidence weights of the first observation vector and the second observation vector are generated by combining the environmental interference coefficients and the feature importance scores, respectively; Based on the real-time confidence weights, a Kalman gain matrix is generated through a Kalman gain calculation formula.

5. The method of claim 1, wherein, The sub-pixel recognition and optical distortion compensation of the vibration interference image data are performed by using the marker point recognition algorithm to generate three-dimensional visual coordinate data, including: Obtaining the camera internal parameters, lens distortion coefficients, spatial position relationship and attitude parameters of the binocular vision system in the flight target; Extracting the pixel area of the visual marker point from the vibration interference image data, calculating the gray distribution barycenter of the pixel area, and taking the gray distribution barycenter as the sub-pixel level coordinate position; Based on the camera internal parameters and the lens distortion coefficients, an optical distortion compensation model is established; The sub-pixel level coordinate position is input into the optical distortion compensation model, and the sub-pixel level coordinate position is nonlinearly transformed by the polynomial coefficients stored in the optical distortion compensation model to output the corrected coordinates after eliminating the lens distortion; According to the spatial position relationship and the attitude parameters, the horizontal position difference of the corrected coordinates in the left and right images of the binocular vision system is calculated by using a stereo matching algorithm, and the disparity information is generated based on the horizontal position difference; Based on the disparity information, the three-dimensional visual coordinate data of the visual marker point in the camera coordinate system is generated by the principle of triangulation.

6. The method of claim 1, wherein, The three-dimensional coordinate point cloud data of the three-dimensional adjustable connecting piece is generated based on the displacement prediction parameters, including: The displacement prediction parameters are converted into spatial scanning trigger instructions of multiple time difference ranging base stations, and the spatial scanning trigger instructions are used to control multiple time difference ranging base stations to emit ultra-wideband signals according to a preset time sequence; The echo signals of the ultra-wideband signals reflected by the surface of the three-dimensional adjustable connecting piece are received, and the signal propagation time differences of each time difference ranging base station are measured based on the echo signals; According to the signal propagation time differences, the signal propagation path lengths are calculated; Based on the signal propagation path lengths, the three-dimensional coordinate set of the surface reflection points of the three-dimensional adjustable connecting piece is generated in combination with the spatial geometric distribution relationship of each time difference ranging base station; Based on the spatial position change amount in the displacement prediction parameters, the three-dimensional coordinate set is subjected to spatial clustering processing to form three-dimensional coordinate point cloud data.

7. The method of claim 1, wherein, The displacement compensation instructions are generated according to the spatial pose offset amount, and the displacement compensation instructions are input into the closed-loop position controller of the flight target to adjust the spatial pose of the three-dimensional adjustable connecting piece, including: The spatial pose offset amount is decomposed into three orthogonal linear displacement components and three rotation angle components; According to the linear displacement components and the rotation angle components, the displacement compensation amounts required by each motion axis are calculated in combination with the motion characteristics of the electrically controlled displacement mechanism; The displacement compensation amounts are converted into displacement compensation instructions in the form of pulse control signals. The displacement compensation instruction is input to a digital signal processing unit of the closed-loop position controller, the displacement compensation instruction is adjusted in real time by the digital signal processing unit, and based on the adjusted displacement compensation instruction, a servo motor of the electric control displacement mechanism is driven to adjust the spatial pose of the three-dimensional adjustable connecting piece.

8. A precision positioning system for a three-dimensional adjustable connection of a flight simulator, characterized in that The method comprises the steps of: An acquisition module is configured to acquire real-time stroke length, instantaneous angle change data of a motion control lever in a flight target, and vibration interference image data of a three-dimensional adjustable connecting piece; An input module is configured to input the real-time stroke length and the instantaneous angle change data to a pre-trained platform kinematics model, output displacement prediction parameters of the flight target, and generate three-dimensional coordinate point cloud data of the three-dimensional adjustable connecting piece based on the displacement prediction parameters; A compensation module is configured to perform sub-pixel identification and optical distortion compensation on the vibration interference image data by using a marker point recognition algorithm, and generate three-dimensional visual coordinate data; A fusion module is configured to perform fusion processing on the three-dimensional coordinate point cloud data and the three-dimensional visual coordinate data by using a preset Kalman filter, and output a spatial pose offset of the three-dimensional adjustable connecting piece; A generation module is configured to generate a displacement compensation instruction according to the spatial pose offset, input the displacement compensation instruction to a closed-loop position controller of the flight target, adjust the spatial pose of the three-dimensional adjustable connecting piece by the closed-loop position controller, and achieve accurate positioning of the three-dimensional adjustable connecting piece relative to a reference coordinate system of the flight target.

9. A computing device, comprising: The method comprises the steps of: A memory is configured to store a computer program; A processor is configured to execute the computer program to implement the steps of the accurate positioning method of the three-dimensional adjustable connecting piece of the flight simulator according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer program is stored in the computer-readable storage medium, and when executed by the processor, the computer program can implement the accurate positioning method of the three-dimensional adjustable connecting piece of the flight simulator according to any one of claims 1 to 7.

Images