Cabin butt joint guide rail compensation method based on point cloud

Abstract

The invention discloses a cabin butt joint guide rail compensation method based on point cloud, and relates to the technical field of spacecraft cabin automatic butt joint, and the method comprises the following steps: building a system coordinate system, and carrying out initial calibration; carrying out guide rail error modeling; carrying out six-degree-of-freedom platform error correction modeling; point cloud data acquisition and pose measurement; calculating error compensation control; adjusting and closed-loop feedback are executed; according to the cabin butt joint guide rail compensation method based on the point cloud, modeling and real-time compensation control are carried out on pose errors in the cabin butt joint process under guide of the guide rail by fusing the three-dimensional point cloud sensing technology and six-degree-of-freedom platform control; the problems that in the prior art, guide rail errors cannot be automatically corrected, and the butt joint precision is insufficient are solved, and accurate butt joint, safety and reliability of the large cabin section in the automatic assembling process are guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of spacecraft cabin automatic docking, and particularly relates to a cabin docking guide rail compensation method based on point clouds. BACKGROUND

[0002] In the process of assembling an aerospace vehicle, the docking and assembly of huge cabins (such as rocket body segments, space station cabin segments, etc.) is a key link, and the docking accuracy directly affects the performance and safety of the spacecraft. The traditional cabin docking usually relies on mechanical guide rails and manual attitude adjustment. However, the traditional method has many shortcomings: (1) Limited accuracy of mechanical guide rails: It is difficult to achieve completely ideal straight lines and parallelism in the manufacture and installation of guide rails. In reality, guide rails often have straightness errors and installation errors, which will cause lateral deviation or attitude tilt of cabin B relative to cabin A when cabin B moves along the guide rail. If high-precision machining of the guide rail is relied on to ensure docking accuracy, not only is the cost high, but it is also difficult to eliminate deformation errors caused by other dynamic factors.

[0003] (2) Low efficiency of manual measurement and adjustment: In traditional assembly, in order to ensure docking accuracy, manual tools are often used to measure the positional and angular deviation between the two cabins multiple times, and the cabin attitude is repeatedly adjusted. This manual method is time-consuming and labor-intensive, has low automation, and human factors can cause measurement errors, which cannot meet the increasing requirements of modern aerospace assembly for efficiency and accuracy.

[0004] (3) Lack of real-time error feedback: In the docking process, if real-time monitoring by sensors is lacking, once a deviation occurs, it is often not discovered until the docking is difficult or stuck, and errors cannot be corrected in time, resulting in poor reliability of the docking process. Especially in large cabin docking, any small misalignment can cause the docking interface to fail to match, resulting in installation stress and even structural damage.

[0005] From the above analysis, it can be seen that the existing technology has the following problems and defects: it is difficult to overcome manufacturing, installation and environmental errors by relying on precise guide rail positioning, and docking accuracy cannot be guaranteed; there is a lack of real-time error sensing and compensation means during assembly, and deviations cannot be automatically corrected in time.

[0006] The difficulty in solving the above problems lies in the need for a scheme that combines high-precision measurement with flexible attitude adjustment. On the one hand, the measurement system must be able to obtain the relative pose error of the two cabins in real time during docking, and have sufficient accuracy and robustness; on the other hand, the actuator must have multi-degree-of-freedom active adjustment capability, and be able to implement fast and accurate compensation motion according to the measured error.

[0007] The significance of solving these problems lies in the fact that achieving automated and precise docking of modules will greatly improve the efficiency and quality of spacecraft assembly. If the docking process can be completed without human intervention, it will not only reduce human error and labor intensity but also decrease reliance on the absolute precision of hardware such as mechanical guide rails, thereby reducing costs. Maintaining docking accuracy in complex environments will ensure assembly reliability and structural safety, laying the foundation for intelligent aerospace assembly. Summary of the Invention

[0008] The purpose of this invention is to provide a point cloud-based method for compensating for docking guide rails in module docking, in order to solve the problems mentioned in the background art, such as the difficulty in overcoming manufacturing, installation and environmental errors by relying solely on precision guide rail positioning, which cannot guarantee docking accuracy; the lack of real-time error perception and compensation methods during assembly, which prevents timely automatic correction when deviations occur; the requirement that the measurement system can acquire the relative pose error of the two modules in real time during docking, and has sufficient accuracy and robustness; and the requirement that the actuator has multi-degree-of-freedom active adjustment capabilities, and can implement rapid and accurate compensation movements based on the measured errors.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a method for compensating for docking guide rails based on point clouds, comprising the following steps: Step (1) Establishing the system coordinate system and initial calibration: Provide module A and module B. Module B is mounted on a six-degree-of-freedom platform via guide rails and can move towards module A along the guide rail direction. Establish the coordinate system of module A, the coordinate system of module B, and the global coordinate system. Calibrate the initial alignment relationship of each coordinate system. Perform zero-position calibration on the six-degree-of-freedom platform. Install point cloud sensors to cover the three-dimensional measurement of the docking area between the two modules. Step (2) Guide rail error modeling: Before or during the movement of section B, the actual geometric parameters of the guide rail are obtained and a guide rail error model is established, including the straightness error and attitude error of the guide rail along its length direction; the guide rail error model gives the lateral offset error and angle deviation error in the form of a function of the position of section B on the guide rail. Step (3) Six-degree-of-freedom platform error correction modeling: The mechanism error of the six-degree-of-freedom platform is determined by offline calibration, and the output deviation of the platform under typical attitude is modeled to obtain the relationship between the command attitude and the actual attitude of the platform, which is used for error compensation of platform attitude control. Step (4) Point cloud data acquisition and pose measurement: During the movement of compartment B along the guide rail, the point cloud sensor is activated to acquire the three-dimensional point cloud data of the docking surface of compartment A and compartment B; through point cloud processing and registration calculation, the real-time pose error of compartment B relative to compartment A is obtained. Step (5) Error compensation control calculation: Based on the real-time pose error, combined with the guide rail error model obtained in step (2) and the platform error model obtained in step (3), calculate the required attitude adjustment amount for section B; the attitude adjustment amount is converted into the drive commands of each actuator of the six-degree-of-freedom platform through inverse kinematics solution; Step (6) Execute adjustment and closed-loop feedback: Send the drive command mentioned in step (5) to the six-degree-of-freedom platform to adjust the attitude of module B, and at the same time control the propulsion of module B along the guide rail; cyclically execute steps (4) and (5) during the movement of module B, update the attitude error in real time and superimpose new compensation commands to form closed-loop control until the two modules achieve docking and alignment. Step (7) Complete docking: When the positional error of segment B relative to segment A is detected to be within the predetermined threshold range, stop the movement of segment B and complete the segment connection and fixation.

[0010] Preferably, the establishment of the guide rail error model in step (2) includes using a point cloud sensor to collect actual position information along the guide rail reference path, comparing it with the ideal straight line of the guide rail to obtain the lateral offset curve and tilt angle curve of the guide rail, and compensating and correcting the thermal deformation error of the guide rail according to the ambient temperature data.

[0011] Preferably, in step (4), the pose measurement accuracy is improved by multi-sensor data fusion. The point cloud sensor includes a three-dimensional laser scanner and two industrial cameras. The laser point cloud is used to calculate the relative distance and initial attitude value of the two sections. The precise six-degree-of-freedom pose error is calculated by combining the feature image of the docking surface of the sections taken by the industrial camera with feature matching and registration algorithms.

[0012] Preferably, the error compensation control calculation in step (5) adopts a strategy combining feedforward and feedback. The feedforward control pre-compensates the expected deviation of section B at the current guide rail position based on the guide rail error model and the platform error model. The feedback control calculates the additional adjustment amount based on the residual error of the real-time point cloud measurement. The control calculation includes a proportional-derivative control law to smooth the error correction and prevent attitude adjustment overshoot.

[0013] Preferably, during the execution of step (6), when the error change rate of a certain degree of freedom exceeds a preset threshold or the error continues to fail to decrease, the control system reduces the moving speed of section B and increases the point cloud sampling frequency to enhance the stability and safety of the control; if abnormal interference is detected that causes the error to fail to converge, the system triggers an alarm and suspends the docking process.

[0014] Preferably, the module docking attitude control system includes a fixedly installed module A, a module B connected to a six-degree-of-freedom parallel platform via guide rails, a point cloud sensor module, and a central control system. The point cloud sensor module acquires the relative pose data of module A and module B. The central control system is configured to calculate the attitude error of module B based on the pose data and generate compensation control commands to drive the six-degree-of-freedom parallel platform to adjust the attitude of module B and control module B to move along the guide rails, thereby automatically completing the precise docking of the two modules.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. Significantly improve docking accuracy: High-precision point cloud sensing enables millimeter-level position and attitude measurement. Combined with the fine adjustment of the six-degree-of-freedom platform, it can correct deviations caused by guide rail machining errors and installation errors, ensuring that the center and attitude height of the docking surfaces of the two modules are highly consistent, which is significantly better than the accuracy level of simply relying on mechanical guidance.

[0016] 2. Real-time autonomous correction: This invention constructs an error model of the guide rail and the platform and implements closed-loop control, enabling the system to automatically sense and correct deviations in relative position and angle in real time during the docking process. Even if the guide rail is deformed by heat or there are slight deviations in the platform output, it can be adjusted immediately through sensor feedback to ensure that the docking path is always correct.

[0017] 3. Reduced Assembly Difficulty and Cost: Due to the active compensation mechanism, the guide rails do not need to achieve extremely high precision, and the machining and installation tolerances can be appropriately relaxed, reducing manufacturing costs. At the same time, there is no need to attach expensive positioning targets to the compartment surface or use manual alignment, reducing assembly complexity and human intervention. The docking process is automated, reducing human error and stress, and improving assembly efficiency.

[0018] 4. Enhanced System Robustness: Point cloud sensing and closed-loop control make the docking system less sensitive to environmental changes and disturbances. Even under non-ideal environments such as lighting and vibration, abundant 3D data and algorithms can filter out some interference, achieving stable measurements. The coordinated adjustment of the guide rail and the six-degree-of-freedom platform ensures smooth docking and prevents docking failure due to the accumulation of small deviations, giving the system strong fault tolerance and reliability.

[0019] In summary, this invention provides an automatic docking error compensation scheme for large spacecraft modules by combining point cloud measurement technology with parallel mechanism control, which can meet the requirements of high precision, automation and reliability for intelligent assembly of future spacecraft. Attached Figure Description

[0020] Figure 1 This is a flowchart of the point cloud guide rail error compensation control process of the present invention; Figure 2This is a schematic diagram of the guide rail and platform compensation model of the present invention; Figure 3 This is a schematic diagram of the six-degree-of-freedom platform structure in this invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figures 1-3 This invention provides a point cloud-based method for compensating for docking guide rails of modules, comprising the following steps: Step (1) Establishing the system coordinate system and initial calibration: Provide module A and module B. Module B is installed on a six-degree-of-freedom platform via a guide rail and can move towards module A along the guide rail direction. Establish the coordinate system of module A, the coordinate system of module B, and the global coordinate system. Calibrate the initial alignment relationship of each coordinate system. Perform zero-position calibration on the six-degree-of-freedom platform. Install point cloud sensors to cover the three-dimensional measurement of the docking area of ​​the two modules. Fix module A and module B to be moved are installed on the docking station respectively. Module A is fixed and module B is connected to the guide rail via the six-degree-of-freedom platform to realize movement and attitude adjustment relative to A along the guide rail axis. Establish the coordinate system of module A, the coordinate system of module B, and the global reference coordinate system. Determine the initial relationship between each coordinate system through calibration. The point cloud sensor is installed on the workshop fixed frame. Its field of view covers the docking area of ​​the two modules. During initialization, perform zero-position calibration on the six-degree-of-freedom platform to make the platform in a neutral position. The initial attitude of module B relative to the guide rail is known. Step (2) Guide rail error modeling: Before or during the movement of section B, the actual geometric parameters of the guide rail are obtained and a guide rail error model is established, including the straightness error and attitude error of the guide rail along its length direction; the guide rail error model gives the lateral offset error and angular deviation error in the form of a function of section B's position on the guide rail. Before or during docking, the geometric error of the guide rail is modeled. The ideal axis of the guide rail extends straight along the docking direction, but due to manufacturing and installation deviations, the actual guide rail may have slight offsets and tilts in the horizontal and vertical directions. This invention pre-measures the straightness error and attitude error of the guide rail, establishes the guide rail error model, and defines the guide rail coordinates with the starting point of the guide rail as a reference. Represent the displacement of section B on the guide rail; using a function Indicates the position of the slider Lateral offset relative to the ideal axis, using This indicates the tilt angle error of the guide rail at that point around the direction perpendicular to the axis. The error model can be obtained and corrected in real time through initial laser measurement or point cloud scanning. After the error model is established, it is stored in the control system for use in compensation calculations. Step (3) Six-DOF Platform Error Correction Modeling: The mechanism error of the six-DOF platform is determined through offline calibration. The output deviation of the platform under various typical attitudes is modeled to obtain the relationship between the commanded attitude and the actual attitude of the platform, which is used for error compensation of the platform attitude control. The kinematic model of the six-DOF platform for attitude adjustment of section B is used to calculate the relationship between the stroke of each actuator of the platform and the output pose of the platform. The mechanism error model of the platform is established through calibration experiments, considering the influence of the length deviation of each drive rod, hinge clearance, etc. on the output attitude. The mechanism error of the platform is expressed as the deviation function of the actual output pose of the platform relative to the ideal output under a given command, denoted as ,in Let the theoretical target pose parameter vector of the platform be, for example, when the platform ideally should achieve the pose. At that time, due to the mechanism error, the actual attitude was reached The model is obtained through offline calibration and can be used in control systems to compensate for the platform's own errors. Step (4) Point cloud data acquisition and pose measurement: During the movement of compartment B along the guide rail, the point cloud sensor is activated to acquire the three-dimensional point cloud data of the docking surface of compartment A and compartment B; through point cloud processing and registration calculation, the real-time pose error of compartment B relative to compartment A is obtained. During the movement of compartment B towards compartment A along the guide rail, the point cloud sensor is activated to perform real-time scanning and measurement of the relative position and attitude of the two compartments. The acquired data includes a large number of point clouds in the interface area of ​​compartment A and compartment B, containing rich depth information and three-dimensional geometric features. Through the processing of point cloud data, the corresponding docking features on compartment A and B (such as the contour of the compartment flange surface, the positioning point of the connecting hole, etc.) are extracted. Using the point cloud registration algorithm, the current pose error of compartment B relative to compartment A is calculated. This pose error can be represented as a rigid body transformation matrix. Or equivalent six-degree-of-freedom deviation parameters ,in This represents the translational deviation of B relative to A. Indicates the rotational deviation about each axis; Step (5) Error Compensation Control Calculation: Based on the real-time pose error, combined with the guide rail error model obtained in step (2) and the platform error model obtained in step (3), the required attitude adjustment amount for section B is calculated; the attitude adjustment amount is converted into the drive commands of each actuator of the six-degree-of-freedom platform through inverse kinematics solution, and the system receives the above-measured pose error. By combining the guide rail error model and the platform error model for comprehensive calculation, a compensating motion command is generated. Its core is based on the known current position of the guide rail. The expected guide rail offset and platform's own errors Under the premise that the solution requires the attitude adjustment amount applied by the six-degree-of-freedom platform to offset all errors and make the module B gradually approach the ideal docking attitude, satisfying: ; in This represents the actual pose change of current module B relative to module A. This represents the target pose of B relative to A during ideal docking. This is the compensation transformation that needs to be applied; the controller uses the platform's inverse kinematics calculations... The required platform-driven changes, through After making corrections, the final length adjustment of each actuator is obtained. ; Step (6) Execute adjustment and closed-loop feedback: Send the drive command mentioned in step (5) to the six-degree-of-freedom platform to adjust the attitude of module B, and at the same time control the advancement of module B along the guide rail; cyclically execute steps (4) and (5) during the movement of module B, update the pose error in real time and superimpose new compensation commands to form a closed-loop control until the two modules are docked and aligned, send the calculated platform attitude adjustment command to the six-degree-of-freedom platform controller, and at the same time control the advancement speed of the guide rail as needed. While the platform performs attitude adjustment, module B continues to move slowly along the guide rail. When B moves further, the point cloud sensor acquires the updated relative pose data again, enters the next round of measurement and compensation cycle, continuously monitors the attitude error of the module and corrects it in real time. After each iteration, the deviation of module B relative to A gradually decreases. Step (7) Complete docking: When the pose error of segment B relative to segment A is detected to be within the predetermined threshold range, stop the movement of segment B and complete the connection and fixation of the segments. When the point cloud measurement confirms that the errors of all six degrees of freedom of segment B relative to A are within the predetermined tolerance range, the system determines that precise docking and positioning have been achieved. At this time, the control platform keeps the attitude of segment B unchanged, and the guide rail pushes segment B to complete the final contact and positioning. Then, the mechanical locking mechanism can be used to connect and fix the two segments. After docking is completed, the final attitude data is recorded. Through the above steps, the present invention realizes the coordinated compensation control of the guide rail geometric error and the platform mechanism error during the segment docking process. During the movement of segment B, the known error of the guide rail is pre-corrected, the residual error is sensed in real time and dynamically adjusted to ensure that segment B always faces the correct direction and aligns with segment A, thereby achieving high-precision docking.

[0023] 1. System Structure and Working Principle The module docking compensation control system of the present invention includes a fixed module A, a module B to be docked, a guide rail mechanism, a six-degree-of-freedom adjustment platform, a point cloud measurement module, etc. Module A is securely mounted on the assembly platform via a support frame, with its docking interface fixed in the direction to be docked. Module B is mounted on the six-degree-of-freedom adjustment platform, the base of which is connected to the guide rail slide, allowing the platform and module B to move horizontally along the guide rail. The guide rail mechanism includes two parallel linear guide rails and a sliding trolley. The trolley is fixedly connected to the base of the six-degree-of-freedom platform, and a drive device controls the trolley to move along the guide rail axis, thereby achieving the feed movement of module B relative to module A.

[0024] The point cloud measurement module includes a 3D laser scanner and an industrial camera assembly mounted near the docking interface. The laser scanner is fixed to a frame on one side of section A, facing the end of section B, and continuously emits laser light and collects reflected point cloud data during section B's advancement. The industrial camera assembly consists of multiple cameras distributed around the docking area, capturing feature images of the end-to-end docking surfaces of the two sections from different angles. In this embodiment, the scanner acquires point clouds of the section surfaces for real-time distance measurement, while the cameras acquire images for feature recognition; the data from both are fused and calculated by the control system to improve the accuracy and reliability of pose measurement.

[0025] The central control system incorporates a high-performance processor and motion control unit for processing sensor data and driving actuators. The control system is connected to a point cloud measurement module, motor drivers, and a six-degree-of-freedom platform controller. On one hand, the control system acquires data from scanners and cameras, and calculates the pose error of section B relative to section A by executing feature extraction and point cloud registration algorithms. On the other hand, the control system calculates appropriate compensation motion commands based on the error, and adjusts these commands by driving the guide rail mechanism and the six-degree-of-freedom platform.

[0026] 2. Error Model Establishment and Correction Methods like Figure 1 As shown, the method flow of this invention includes alternating measurement and control phases. First, guide rail error calibration is performed: before formal docking, the control system drives the six-degree-of-freedom platform to slowly move section B along the guide rail, while a laser scanner collects the trajectory point cloud of the trolley reference point. By comparing this trajectory with an ideal straight line, the system fits the guide rail offset curves in the X and Y directions. And by comparing the attitude of the trolley, the deflection angle of the guide rail about the axis is obtained. The curve is used to establish and store the guide rail error model. Furthermore, repeating the above measurements at different ambient temperatures allows for the acquisition of the guide rail's thermal deformation characteristics, which can be used for error correction under temperature variations.

[0027] Next, platform error calibration is performed: the six-DOF platform is instructed to select several attitudes within the workspace, and the actual attitude of module B is measured using high-precision measurement methods. This measurement is then compared with the platform command values ​​to obtain the platform error under each typical attitude. The error is correlated with each drive stroke of the platform, and a platform mechanism error compensation table is established using the parameter separation calibration method. After calibration, the error model is integrated into the platform controller for real-time compensation.

[0028] 3. Real-time control during the docking process like Figure 2 As shown, after the docking officially begins, section B is slowly moved towards section A by the guide rail. Initially, the six-DOF platform maintains a reference attitude, ensuring that section B is roughly aligned with section A. The point cloud measurement module continues to operate: a laser scanner scans and acquires point clouds of the end faces of sections A and B, while an industrial camera captures pre-designed mating features on the docking surface. The control system fuses the acquired data: outliers in the point cloud are removed using the RANSAC algorithm, and the circular contours of the end faces of the two sections are fitted to accurately calculate the positions of the flange centers of section A and section B in the global coordinate system; simultaneously, the yaw and pitch angle differences of section B relative to section A are acquired through multi-view vision measurement. This yields the current real-time offset. and angle difference .

[0029] The control system will use the error vector Compare the values ​​with the guide rail error model predictions. For example, at this time, the slide of section B is located... The expected offset due to guide rail error is If measured and The discrepancy may be due to other factors causing additional deviations. The control algorithm then calculates the required platform compensation motion accordingly. The basic control law can be expressed as: ; in and To determine the position based on the guide rail model Expected translation and rotation errors, The control gain matrix (with different weights set according to different degrees of freedom) is used. This control law is equivalent to performing closed-loop correction on the residual error of the current position and considering the rate of change of the error to smooth the control. The calculated... To determine the pose increment that the platform should adjust in the next moment, the control system further uses the platform's inverse kinematics... Converted into the length increment of each leg The platform controller executes the length adjustment, and the attitude of section B is immediately corrected.

[0030] In actual operation, the above process is executed repeatedly at a high frequency. As section B approaches section A, the distance to the docking interface decreases, and the control system gradually increases the measurement resolution and decreases the feed per step to prevent missing minor errors. Finally, when section B is very close to the docking position, through high-precision fusion of multi-sensor data, the errors of all degrees of freedom are corrected to within the allowable range. At this point, the flatness and coaxiality of the flange surfaces of both sections meet the assembly requirements.

[0031] 4. Compensation strategies in special circumstances The method of this invention can also handle various special working conditions. For example, when a rapid change in ambient temperature causes significant thermal expansion deformation of the guide rail, the control system detects the change using a temperature sensor and corrects the guide rail error model to ensure that the compensation calculation is consistent with the actual situation. When uneven force on one side of the guide rail causes micro-twist, point cloud measurement will capture abnormal deflection of the compartment attitude. The control system can selectively increase the gain of the corresponding degree of freedom to quickly correct the resulting deviation and prevent further expansion. In addition, if repeated oscillations or failure to convergence of an error in a certain degree of freedom are detected during docking, the system determines that there may be a sensor abnormality or mechanism jamming, and can pause propulsion and issue an alarm for inspection, thereby improving safety. The establishment of the guide rail error model in step (2) includes using a point cloud sensor to collect actual position information along the guide rail reference path, comparing it with the ideal straight line of the guide rail to obtain the lateral offset curve and tilt angle curve of the guide rail, and compensating and correcting the thermal deformation error of the guide rail based on the ambient temperature data. Before docking, the position of a fixed reference object is collected along the guide rail sliding path using a point cloud sensor, the actual spatial straight line of the guide rail is fitted, and the error distribution is obtained by comparing it with the ideal straight line, which is used to construct the error model. This model can be dynamically updated according to temperature sensing data during the docking process to compensate for changes in guide rail length or curvature caused by other factors.

[0032] In step (4), the pose measurement accuracy is improved through multi-sensor data fusion. The point cloud sensor includes a 3D laser scanner and two industrial cameras. The laser point cloud is used to calculate the relative distance and initial attitude value of the two sections. Combined with the feature image of the docking surface of the sections captured by the industrial cameras, the accurate six-degree-of-freedom pose error is calculated through feature matching and registration algorithms. At the same time, the laser scanner and stereo vision camera are used to acquire point cloud and image information. A more reliable pose measurement result is obtained through sensor data fusion. In complex environments, filtering algorithms can be used to smooth the measured pose error, filter out noise interference, and ensure the stability of control commands.

[0033] In step (5), the error compensation control calculation adopts a strategy combining feedforward and feedback. Feedforward control pre-compensates for the expected deviation of section B at its current guide rail position based on the guide rail error model and platform error model. Feedback control calculates additional adjustments based on the residual error from real-time point cloud measurements. The control calculation includes a proportional-derivative control law to smooth error correction and prevent attitude adjustment overshoot. When calculating the platform adjustment, the platform's motion range and speed limitations are considered, and a smooth attitude adjustment path is planned to avoid exceeding tolerances or unusual positions of the mechanism. Simultaneously, it ensures stable movement of section B, preventing collisions or vibrations caused by drastic adjustments. The control system can predict the next necessary adjustment based on real-time error trends, coordinating the guide rail propulsion and platform movements in advance, making the entire docking process smooth and efficient.

[0034] During the execution of step (6), when the error change rate of a certain degree of freedom exceeds the preset threshold or the error cannot be reduced, the control system reduces the moving speed of section B and increases the point cloud sampling frequency to enhance the stability and safety of the control; if abnormal interference is detected that causes the error to fail to converge, the system triggers an alarm and suspends the docking process.

[0035] The module docking attitude control system includes a fixedly installed module A, a module B connected to a six-degree-of-freedom parallel platform via guide rails, a point cloud sensor module, and a central control system. The point cloud sensor module is used to acquire the relative pose data of module A and module B. The central control system is configured to calculate the attitude error of module B based on the pose data and generate compensation control commands to drive the six-degree-of-freedom parallel platform to adjust the attitude of module B and control module B to move along the guide rails, so as to automatically complete the precise docking of the two modules.

[0036] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A point cloud based cabin section docking guide rail compensation method, characterized by: The method comprises the following steps: Step (1) establishing a system coordinate system and initial calibration: including cabin section A and cabin section B, the cabin section B is installed on a six-degree-of-freedom platform through a guide rail and can move along the guide rail direction towards the cabin section A, a cabin section A coordinate system, a cabin section B coordinate system and a global coordinate system are established, and an initial alignment relationship of each coordinate system is calibrated; the six-degree-of-freedom platform is calibrated to zero position; a point cloud sensor is installed for covering three-dimensional measurement of the two-cabin-section docking area; Step (2) guide rail error modeling: before or during the movement of the cabin section B, the actual geometric parameters of the guide rail are obtained and a guide rail error model is established, including straightness error and attitude error of the guide rail along the length direction thereof; the guide rail error model gives the lateral offset error and the angle deviation error in the form of a function of the position of the cabin section B on the guide rail; Step (3) six-degree-of-freedom platform error correction modeling: the mechanism error of the six-degree-of-freedom platform is determined through offline calibration, the output deviation of the platform under a typical attitude is modeled, the relationship between the command attitude and the actual attitude of the platform is obtained, and error compensation for platform attitude control is used; Step (4) point cloud data acquisition and pose measurement: during the movement of the cabin section B along the guide rail, the point cloud sensor is started to obtain three-dimensional point cloud data of the docking surface of the cabin section A and the cabin section B; through point cloud processing and registration calculation, the real-time pose error of the cabin section B relative to the cabin section A is obtained; Step (5) error compensation control calculation: according to the real-time pose error, the guide rail error model obtained in step (2) and the platform error model obtained in step (3) are combined to calculate the required attitude adjustment amount of the cabin section B; the attitude adjustment amount is converted into driving instructions of each actuator of the six-degree-of-freedom platform through inverse kinematics; Step (6) executing adjustment and closed-loop feedback: the driving instructions of step (5) are sent to the six-degree-of-freedom platform to adjust the attitude of the cabin section B, while controlling the propulsion of the cabin section B along the guide rail; steps (4) and (5) are executed in a loop during the movement of the cabin section B, the pose error is updated in real time and new compensation instructions are superimposed, forming a closed-loop control until the two cabin sections are docked and aligned; Step (7) completing docking: when it is detected that the pose error of the cabin section B relative to the cabin section A is within a predetermined threshold range, the movement of the cabin section B is stopped and the cabin connection is fixed.

2. The method of claim 1, wherein: The establishment of the guide rail error model in step (2) includes collecting actual position information along the guide rail reference path using the point cloud sensor, comparing with the ideal straight line of the guide rail to obtain the lateral offset curve and the inclination angle curve of the guide rail, and compensating and correcting the thermal deformation error of the guide rail according to the environmental temperature data.

3. The method of claim 1, wherein: In step (4), the pose measurement accuracy is improved through multi-sensor data fusion, the point cloud sensor includes a three-dimensional laser scanner and two industrial cameras, the relative distance and attitude initial value of the two cabin sections are calculated using laser point cloud, and the accurate six-degree-of-freedom pose error is calculated through feature matching and registration algorithm based on the feature image of the cabin docking surface taken by the industrial camera.

4. The method of claim 1, wherein: The error compensation control calculation in step (5) adopts a strategy combining feedforward and feedback, wherein the feedforward control pre-compensates the predicted deviation of cabin section B at the current guide rail position according to the guide rail error model and the platform error model, and the feedback control calculates an additional adjustment amount according to the residual error measured by the real-time point cloud; the control calculation includes a proportional-differential control law to smooth the error correction and prevent attitude adjustment overshoot.

5. The method of claim 1, wherein: During the execution of step (6), when the error change rate of a certain degree of freedom exceeds a preset threshold or the error cannot be reduced for a long time, the control system reduces the moving speed of cabin section B and increases the point cloud sampling frequency to enhance the stability and safety of the control; if abnormal interference is detected that causes the error to be unable to converge, the system triggers an alarm and suspends the docking process.

6. A docking attitude control system for implementing the point cloud based docking guide compensation method according to any one of claims 1 to 5, characterized in that: The cabin docking attitude control system includes a fixedly installed cabin A, a cabin B connected on a six-degree-of-freedom parallel platform through a guide rail, a point cloud sensor module, and a central control system; wherein the point cloud sensor module is used to obtain the relative pose data of cabin A and cabin B, the central control system is configured to calculate the attitude error of cabin B according to the pose data and generate a compensation control instruction, drive the six-degree-of-freedom parallel platform to adjust the attitude of cabin B and control cabin B to move along the guide rail, so as to automatically complete the accurate docking of the two cabins.

Images