Method, device, storage device and program product for assembling flexible shaft holes with large deflection based on tactile perception

Abstract

The method, device, storage device and program product for assembling flexible shaft holes with large deflection deformation based on tactile perception belong to the field of robot assembly technology. The problem that the existing assembly method does not consider the influence of the nonlinear large deflection deformation of the flexible shaft on the shaft hole assembly accuracy is solved. Based on the pressure distribution data fed back by the tactile sensor, the present invention can estimate and adjust the deformation of the shaft, so that the flexible shaft workpiece and the rigid plane where the target hole is located are effectively contacted, and whether the shaft passes through the hole in real time during the blind hole search process, and after finding the hole, the robot arm is controlled to automatically align the shaft hole to complete the shaft insertion action. The present invention is mainly used for shaft hole assembly of flexible shaft workpieces with large deflection deformation.

Description

Technical Field

[0001] The invention belongs to the technical field of robot assembly, and in particular relates to a shaft hole assembly method for a flexible shaft with large deflection deformation based on tactile perception. Background Art

[0002] With the improvement of robotic control theory and the development of various skill learning methods, robots have become able to autonomously manipulate rigid objects, but they still struggle to handle non-rigid flexible materials. Manipulating flexible objects presents significant challenges in sensing, modeling, planning, and controlling them due to their strong nonlinearity and strong deformation, coupled with the high-precision, unstructured environment requirements. Shaft-to-hole assembly is a typical task in robotic assembly technology. Conventional rigid shaft-to-hole assembly methods generally employ robotic force feedback control, making it difficult to directly apply to flexible shafts with large deflection characteristics. Shaft-to-hole assembly technology for flexible shafts is generally based on visual methods, such as using depth and color images, and model-based tracking to achieve real-time perception of flexible objects, estimate the elastic modulus and position of the grasped object, and predict deformation. However, this approach has not been validated under complex force interactions. Using visual servoing, the deformation of the end of a deformable slender shaft is modeled, and the differential kinematic relationship between image features and the manipulator joints is derived. When the image features converge to the desired features, shaft-hole alignment is achieved. However, vision-based methods are limited by factors such as image quality and occlusion. However, this method is applied to specific tasks and is not easily generalizable. Subsequently, a model-free robust adaptive control (MFRAC) strategy was proposed to control a robot with unknown dynamics to manipulate a flexible rubber object with unknown elasticity. They used a Gaussian mixture model (GMM) based on a distributional similarity measure (DSM) to model the contact state of the rubber rod. However, the flexible rubber rod they used had a small aspect ratio, meaning it was not susceptible to large deflections, and could not be transferred to more general flexible slender shafts.

[0003] With advances in sensing technology, tactile sensors are widely used in robotic manipulation tasks such as object attribute recognition, grasping, and human-machine interaction. Robotic tactile sensing technologies generally include resistive tactile sensors, capacitive tactile sensors, and visual tactile sensors. These sensors have one-dimensional or multi-dimensional force sensing capabilities, providing robots with the ability to perceive the local touch of the end effector.

[0004] Tactile sensors have been successfully applied to numerous assembly tasks. Tactile-aware robots can estimate the posture of clamped objects through touch, enabling precise assembly without relying on vision or external force sensors. However, the posture of flexible shafts is more challenging to estimate. To complete shaft-to-hole assembly tasks, existing tactile-aware robots fail to consider the impact of large, nonlinear deflections on shaft-to-hole assembly accuracy. Therefore, a method is needed to account for these large, nonlinear deformations. Summary of the Invention

[0005] The purpose of the present invention is to solve the problem that the existing assembly method does not consider the influence of the nonlinear large deflection deformation of the flexible shaft on the shaft hole assembly accuracy. The present invention provides a large deflection deformation flexible shaft hole assembly method, device, storage device and program product based on tactile perception.

[0006] A large deflection flexible shaft-hole assembly method based on tactile perception includes:

[0007] S1. Preparation phase: After installing a thin film pressure tactile sensor array on the end gripper of the robot arm, the end gripper of the robot arm is controlled to grip the flexible shaft workpiece and keep the flexible shaft workpiece perpendicular to the rigid plane where the target hole is located, preparing for assembly operation; wherein the thin film pressure tactile sensor array is used to sense the distribution data of the gripping force;

[0008] S2, operation close to the target hole: the tactile sensing deformation control algorithm controls the flexible shaft workpiece and the rigid plane to form effective contact according to the sensed clamping force distribution data; wherein, the effective contact is the height z of the flexible shaft workpiece after deformation. est In the effective deformation range [z min , z max ] and keep the flexible shaft workpiece in contact with the upper surface of the rigid plane; the height z of the flexible shaft workpiece after deformation est z is the vertical height between the clamping jaw and the bottom end of the flexible shaft workpiece; min The minimum height value to prevent the shaft hole from getting stuck, z max The maximum height value of the axis and the rigid surface to produce stable contact, and z max Smaller than the length of the flexible shaft workpiece in the free state;

[0009] S3. Search target hole operation: Control the robotic arm to move horizontally along the set hole search trajectory, so that the bottom end of the flexible shaft workpiece moves back and forth linearly along the preset path and sweeps the rigid surface. The target hole is searched based on the change in the gripping force distribution data sensed at the current moment and the previous moment. When it is determined that the flexible shaft workpiece passes through the target hole at the current moment, the robotic arm stops moving.

[0010] S4, jack operation: According to the deviation between the current clamping force distribution data and the clamping force distribution data sensed in the initial preparation stage, the relative position error between the axis and the hole is calculated. According to the relative position error between the axis and the hole, the control amount of the horizontal movement direction and movement distance of the robot arm is determined to control the movement of the robot arm to achieve the alignment of the flexible axis workpiece and the target hole. Then, the robot arm is controlled to move downward to a given depth h. step , complete the jack.

[0011] Preferably, in step S2, the tactile sensing deformation control algorithm controls the flexible shaft workpiece to form effective contact with the rigid plane according to the sensed clamping force distribution data, including the following steps:

[0012] S21, controlling the flexible shaft workpiece to move toward the rigid plane according to the spatial position of the rigid plane, and making the bottom end of the flexible shaft workpiece contact with the upper surface of the rigid plane;

[0013] S22, the pre-trained CNN network predicts the height z of the flexible shaft workpiece after deformation based on the numerical matrix M formed by the clamping force distribution data sensed by the thin film pressure tactile sensor array est ;

[0014] S23, determine whether z is satisfied est In the effective deformation range [z min , z max ], if the result is yes, it is determined that the flexible shaft workpiece forms effective contact with the rigid plane, and step S25 is executed; if the result is no, step S24 is executed;

[0015] S24, controlling the robotic arm to move upward in the vertical direction, and returning to step S22;

[0016] S25, maintain the height z of the flexible shaft workpiece after deformation est Constant and unchanging.

[0017] Preferably, in step S3, searching for the target hole according to the change in the clamping force distribution data sensed at the current moment and the previous moment is implemented as follows:

[0018] When the change between the numerical matrix formed by the clamping force distribution data sensed by the thin film pressure tactile sensor array at the current moment and the previous moment is greater than the preset threshold, it is determined that the flexible shaft workpiece passes through the target hole at the current moment and the target hole search is completed.

[0019] Preferably, in step S4, the calculation of the relative position error e between the shaft and the hole is implemented as follows:

[0020] e=mean(M t -M Target );

[0021] Mt is the numerical matrix of the clamping force distribution data sensed by the pressure tactile sensor array at the current moment t, M Target is the numerical matrix of the clamping force distribution data sensed by the pressure tactile sensor array when the flexible shaft workpiece is not subjected to any external force except the clamping force, and mean() is the column-wise average of the vector.

[0022] Preferably, in step S4, the method for determining the control amount of the horizontal movement direction and movement distance of the robot arm according to the relative position error between the shaft and the hole includes:

[0023] When the relative position error between the axis and the hole is greater than the preset relative position tolerance, the PID control generates a control variable for controlling the horizontal movement direction and movement distance of the robot arm based on the relative position error between the axis and the hole.

[0024] Preferably, the gripper is implemented as an adaptive two-finger gripper.

[0025] Preferably, a thin film pressure tactile sensor array is mounted on one or both sides of the adaptive two-finger gripper.

[0026] A flexible shaft-hole assembly device based on tactile perception includes a storage device, a processor, and a computer program stored in the storage device and runnable on the processor. The processor executes the computer program to implement the large deflection deformation flexible shaft-hole assembly method based on tactile perception.

[0027] A computer-readable storage device stores a computer program, and when the computer program is executed, the flexible shaft-hole assembly method with tactile perception is implemented.

[0028] A computer program product comprises a computer program, which implements the flexible shaft-hole assembly method with tactile perception when the computer program is executed by a processor.

[0029] The beneficial effects brought by the present invention are:

[0030] In view of the fact that a flexible shaft may undergo large deflection deformation when interacting with the environment, the present invention innovatively applies tactile sensing technology and designs effective solutions for each major process of shaft-hole assembly.

[0031] First, to control the deformation of the flexible shaft during operation, traditional vision technology requires complex image processing algorithms and cannot cope with the operating requirements under harsh working conditions such as obstacle obstruction and insufficient light. A single wrist force sensor cannot accurately sense the deformation of the flexible shaft. The tactile technology used in the present invention can accurately and efficiently sense the deformation of the flexible shaft to facilitate control.

[0032] Secondly, the present invention does not need to estimate the specific position of the end of the flexible shaft workpiece, and can quickly enable the shaft to find the hole based on the tactile response.

[0033] Finally, the present invention does not require additional force control and relies on tactile feedback to align the hole axis and complete the insertion.

[0034] Compared with the existing technology, the present invention only uses a single dimension of tactile perception, enabling the robot to automatically and highly precisely complete the shaft hole assembly task of the flexible workpiece. The cost of the visual and force sensors used in the existing technology is generally high, while the thin film tactile sensor used in the present invention is low in cost and has a very high cost-effectiveness in completing the same task, greatly reducing the cost and complexity of the assembly system.

[0035] The present invention establishes a shaft-hole assembly strategy for flexible shaft workpieces with large deflection deformation based on tactile perception. The application of the present invention can realize dexterous and precise assembly operations on flexible shaft workpieces, which has a certain inspiration and optimization effect on the industrial assembly of flexible parts. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a flow chart of the method for assembling a flexible shaft hole with large deflection based on tactile perception according to the present invention;

[0037] Figure 2 Flowchart for controlling the flexible shaft workpiece to form effective contact with the rigid plane;

[0038] Figure 3 Schematic diagram of the hole-searching process;

[0039] Figure 4 A schematic diagram of the interaction between a flexible shaft workpiece and a rigid plane is given;

[0040] Figure 4 a is a schematic diagram of the state when the flexible shaft workpiece and the rigid plane produce contact deformation. Figure 4 b is a schematic diagram of the state when the flexible shaft workpiece meets the target hole on the rigid plane. Figure 4 c is a schematic diagram of the state when the shaft hole is blocked. Figure 4 d is a schematic diagram of the state when the shaft hole is stuck and the large deformation is caused by inserting the hole downward;

[0041] Figure 5 This is a graph showing the measurement value change trend of a sensing unit in the thin film pressure tactile sensor array during the shaft hole assembly process. DETAILED DESCRIPTION

[0042] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0043] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0044] This invention designs a method for guiding a robot during shaft-hole assembly operations using tactile sensor measurement feedback. This method enables deformation control, blind hole search, and insertion of flexible shaft workpieces. An array of thin-film pressure tactile sensors is installed within the gripper jaws as a tactile sensing device, capable of detecting the distribution of normal clamping force. The robot is not equipped with any additional six-dimensional force sensors or visual sensors for additional force or visual feedback. The object being manipulated is a slender shaft made of soft material, capable of elastically deforming to a large degree under external force.

[0045] Based on the pressure distribution data fed back by the tactile sensor, the deformation of the shaft can be estimated and adjusted to ensure effective contact between the flexible shaft workpiece and the rigid plane where the target hole is located. During the blind hole search process, the shaft is monitored in real time to see if it passes through the hole. After finding the hole, the control robot automatically aligns the shaft and hole to complete the shaft insertion. The specific implementation method provided is as follows:

[0046] Specific implementation method 1: Referring to 1, this implementation method is described. The method of assembling a flexible shaft hole with large deflection deformation based on tactile perception in this implementation method includes:

[0047] S1. Preparation phase: After installing a thin film pressure tactile sensor array on the end gripper of the robot arm, the end gripper of the robot arm is controlled to grip the flexible shaft workpiece and keep the flexible shaft workpiece perpendicular to the rigid plane where the target hole is located, preparing for assembly operation; wherein the thin film pressure tactile sensor array is used to sense the distribution data of the gripping force;

[0048] S2, operation close to the target hole: the tactile sensing deformation control algorithm controls the flexible shaft workpiece and the rigid plane to form effective contact according to the sensed clamping force distribution data; wherein, the effective contact is the height z of the flexible shaft workpiece after deformation. est In the effective deformation range [z min , z max ] and keep the flexible shaft workpiece in contact with the upper surface of the rigid plane; the height z of the flexible shaft workpiece after deformation est z is the vertical height between the clamping jaw and the bottom end of the flexible shaft workpiece;min The minimum height value to prevent the shaft hole from getting stuck, z max The maximum height value of the axis and the rigid surface to produce stable contact, and z max Smaller than the length of the flexible shaft workpiece in the free state;

[0049] S3. Search target hole operation: Control the robot arm to move horizontally according to the set hole search trajectory, so that the bottom end of the flexible shaft workpiece moves back and forth along the preset path to slide across the rigid plane. Search for the target hole based on the change in the clamping force distribution data sensed at the current moment and the previous moment. When it is determined that the flexible shaft workpiece passes through the target hole at the current moment, the robot arm stops moving. For details, see Figure 3 ;

[0050] S4, jack operation: According to the deviation between the current clamping force distribution data and the clamping force distribution data sensed in the initial preparation stage, the relative position error between the axis and the hole is calculated. According to the relative position error between the axis and the hole, the control amount of the horizontal movement direction and movement distance of the robot arm is determined to control the movement of the robot arm to achieve the alignment of the flexible axis workpiece and the target hole. Then, the robot arm is controlled to move downward to a given depth h. step , complete the jack.

[0051] This specific embodiment utilizes a simple, low-cost array of thin-film pressure tactile sensors, enabling the robotic arm to perform tactile sensing during operation. The pressure distribution sensed by the tactile unit can reflect the large deflection deformation of the flexible shaft. A deformation control algorithm based on tactile sensing is used to adjust the degree of deformation of the flexible shaft during the process of approaching the target hole. Tactile sensing of changes in the clamping force distribution is used to detect the intersection of the hole and the shaft during the search for the target hole. Finally, tactile feedback is used to guide the robotic arm to eliminate the relative position error between the shaft and the hole, completing the insertion. Compared to existing technologies, this invention utilizes single-dimensional tactile feedback, which is low-cost and highly efficient. This not only reduces the complexity of the assembly system, but also ensures that the robotic arm automatically controls shaft deformation, accurately finds and inserts holes, and is easy to use.

[0052] The implementation basis of shaft-hole assembly is to establish effective force contact between the shaft and the plane. Traditional rigid shaft-hole assembly often adopts force-position hybrid control or impedance control. These methods are difficult to apply directly to shafts with large inherent compliance. For flexible shaft workpieces, deformation control is the key to controlling the interaction between the shaft and the environment. Due to the uncertainty of the rigid plane where the target hole is located in the z direction, it is necessary to determine the contact between the shaft and the plane and the hole by detecting the deformation of the shaft. Figure 4As shown in the figure, when the shaft is bent from the free state, it means that the shaft has come into contact with the plane, that is, it has found the plane. According to the deformation theory of flexible shaft workpieces, greater deformation will cause more obvious changes in sensor values ​​and is conducive to detection. However, excessive deformation may sometimes cause some unexpected situations, such as shaft hole blocking, such as Figure 4 c. The horizontal movement of the gripper prevents the shaft from falling into the hole. When the deformation caused by the interaction between the shaft and the plane is too large, the shaft will become stuck when passing over the hole. In this case, no matter how the shaft's horizontal position is adjusted, the error in the shaft-hole posture cannot be eliminated. If the shaft is controlled to be inserted downward at this time, not only will the shaft not fall into the hole, but it will also be further deformed. Therefore, it is necessary to design a reasonable deformation control strategy to ensure effective contact between the shaft and the plane while avoiding excessive deformation.

[0053] Therefore, it is necessary to limit the effective deformation range of the flexible shaft workpiece [z min , z max ], so that the height of the flexible shaft workpiece after deformation is in the effective deformation range [z min , z max ] to achieve accurate shaft-hole assembly, see Figure 2 , specifically give

[0054] In step S2, the tactile sensing deformation control algorithm controls the flexible shaft workpiece and the rigid plane to form effective contact based on the sensed clamping force distribution data, including the following steps:

[0055] S21, controlling the flexible shaft workpiece to move toward the rigid plane according to the spatial position of the rigid plane, and making the bottom end of the flexible shaft workpiece contact with the upper surface of the rigid plane;

[0056] S22, the pre-trained CNN network predicts the height z of the flexible shaft workpiece after deformation based on the numerical matrix M formed by the clamping force distribution data sensed by the thin film pressure tactile sensor array est ;

[0057] S23, determine whether z is satisfied est In the effective deformation range [z min , z max ], if the result is yes, it is determined that the flexible shaft workpiece forms effective contact with the rigid plane, and step S25 is executed; if the result is no, step S24 is executed;

[0058] S24, controlling the robotic arm to move upward in the vertical direction, and returning to step S22;

[0059] S25, maintain the height z of the flexible shaft workpiece after deformation est Constant and unchanging.

[0060] This preferred embodiment is based on the tactile pressure distribution characteristics caused by large deflection of the flexible shaft. It not only considers the necessary conditions for the flexible shaft to contact the plane, but also considers the situation where the flexible shaft will be stuck in the hole and cause assembly failure when the deformation is too large. It gives the deformation range for controlling the effective contact between the flexible shaft workpiece and the rigid plane, and designs a CNN network to establish the relationship between the deformation of the flexible shaft and the height z. est The mapping between the two makes it possible to manipulate the robot arm to adjust the shaft deformation in the vertical direction, thereby keeping the shaft within the effective deformation range. It is precisely because the deformation of the flexible shaft is effectively controlled that the shaft hole jamming is avoided.

[0061] Specifically, in step S3, the method for searching for the target hole based on the change in the clamping force distribution data sensed at the current moment and the previous moment is as follows:

[0062] When the change between the numerical matrix formed by the clamping force distribution data sensed by the thin film pressure tactile sensor array at the current moment and the previous moment is greater than the preset threshold, it is determined that the flexible shaft workpiece passes through the target hole at the current moment and the target hole search is completed.

[0063] Furthermore, in step S4, the relative position error e between the shaft and the hole is calculated as follows:

[0064] e=mean(M t -M Target );

[0065] M t is the numerical matrix of the clamping force distribution data sensed by the pressure tactile sensor array at the current moment t, M Target is the numerical matrix of the clamping force distribution data sensed by the pressure tactile sensor array when the flexible shaft workpiece is not subjected to any external force except the clamping force, and mean() is the column-wise average of the vector.

[0066] Furthermore, in step S4, the implementation method of determining the control amount of the horizontal movement direction and movement distance of the robot arm according to the relative position error between the shaft and the hole includes:

[0067] When the relative position error between the axis and the hole is greater than the preset relative position tolerance, the PID control generates a control variable for controlling the horizontal movement direction and movement distance of the robot arm based on the relative position error between the axis and the hole.

[0068] In this preferred embodiment, PID control can be implemented through existing technology to generate control quantities for the horizontal movement direction and movement distance of the robot arm based on the relative position error between the shaft and the hole, and the advantage of this setting method for the jacking operation is that there is no need to consider the precise position of the end of the flexible shaft workpiece and the hole in the robot coordinate system. The tactile pressure distribution characteristics of the flexible shaft that does not undergo large deflection deformation when not subjected to any force other than the clamping force are used as the control target, so that the robot can complete the shaft-hole alignment and jacking by only adjusting the relative position of the flexible shaft and the hole without knowing the precise position of the hole, which greatly reduces the complexity of the task.

[0069] Furthermore, the gripper is implemented by an adaptive two-finger gripper. Specifically, the thin film pressure tactile sensor array can be installed on one side or both sides of the adaptive two-finger gripper.

[0070] Specific embodiment 2: The flexible shaft hole assembly device based on tactile perception described in this embodiment includes a storage device, a processor, and a computer program stored in the storage device and runnable on the processor. The processor executes the computer program to implement the large deflection deformation flexible shaft hole assembly method based on tactile perception as described in specific embodiment 1.

[0071] Specific embodiment three: This embodiment describes a computer-readable storage device, wherein the storage device stores a computer program, and when the computer program is executed, the tactile-perceived flexible shaft-hole assembly method described in specific embodiment one is implemented.

[0072] Specific embodiment 4: A computer program product described in this embodiment includes a computer program. When the computer program is executed by a processor, the flexible shaft-hole assembly method with tactile perception as described in specific embodiment 1 is implemented.

[0073] Verification test:

[0074] The effectiveness of the present invention is demonstrated through the following verification test. Specifically, the present invention method has been verified through the test, without using the optical camera and six-dimensional force sensor commonly used in traditional assembly methods, only equipped with an array of thin film pressure distribution tactile sensors. During the test, the pressure change of a certain contact point in the tactile sensor array is as follows: Figure 5 As shown, Figure 5 The horizontal axis is the sampling point, and the vertical axis is the sensitivity (dimensionless), and its display range is 0-4095;

[0075] Figure 5It is reflected in the figure that, in the stage close to the target hole, the flexible shaft has not yet contacted the plane, and the pressure signal of the tactile unit is basically 0. After the flexible shaft establishes effective contact with the plane, the pressure signal of the tactile unit rises to a larger value. After estimating the deformation amount and confirming that the shaft is in the effective deformation range, the search begins. Since the deformation remains constant, the tactile signal also remains constant at this time. When the shaft passes over the hole, the pressure signal of the contact shows obvious jitter. In the process of eliminating the relative position error of the shaft and hole, each time the robot arm is adjusted, the pressure value of the tactile unit decreases once, until it finally approaches 0, indicating that the shaft and hole are aligned. At this time, the robot arm is controlled to insert the hole downward to complete the operation. Experimental verification shows that the tactile information used by the present invention can effectively and regularly guide the robot arm to complete the shaft hole assembly task of the flexible shaft with large deflection deformation.

[0076] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely illustrative of the principles and applications of the invention. It should be understood that many modifications may be made to the illustrative embodiments, and that other arrangements may be devised, without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in ways other than those described in the original claims. It should also be understood that features described in conjunction with individual embodiments may be employed in conjunction with other described embodiments.

Claims

1. A large deflection flexible shaft hole assembly method based on tactile perception, characterized in that: Methods include: S1. Preparation phase: After installing a thin film pressure tactile sensor array on the end gripper of the robot arm, the end gripper of the robot arm is controlled to grip the flexible shaft workpiece and keep the flexible shaft workpiece perpendicular to the rigid plane where the target hole is located, preparing for assembly operation; wherein the thin film pressure tactile sensor array is used to sense the distribution data of the gripping force; S2, operation close to the target hole: the tactile sensing deformation control algorithm controls the flexible shaft workpiece and the rigid plane to form effective contact according to the sensed clamping force distribution data; wherein, the effective contact is the height z of the flexible shaft workpiece after deformation. est In the effective deformation range [z min , z max ] and keep the flexible shaft workpiece in contact with the upper surface of the rigid plane; the height z of the flexible shaft workpiece after deformation est z is the vertical height between the clamping jaw and the bottom end of the flexible shaft workpiece; min The minimum height value to prevent the shaft hole from getting stuck, z max The maximum height value of the axis and the rigid surface to produce stable contact, and z max Smaller than the length of the flexible shaft workpiece in the free state; S3. Search target hole operation: Control the robotic arm to move horizontally along the set hole search trajectory, so that the bottom end of the flexible shaft workpiece moves back and forth linearly along the preset path and sweeps the rigid surface. The target hole is searched based on the change in the gripping force distribution data sensed at the current moment and the previous moment. When it is determined that the flexible shaft workpiece passes through the target hole at the current moment, the robotic arm stops moving. S4, jack operation: According to the deviation between the current clamping force distribution data and the clamping force distribution data sensed in the initial preparation stage, the relative position error between the axis and the hole is calculated. According to the relative position error between the axis and the hole, the control amount of the horizontal movement direction and movement distance of the robot arm is determined to control the movement of the robot arm to achieve the alignment of the flexible axis workpiece and the target hole. Then, the robot arm is controlled to move downward to a given depth h. step , complete the jack.

2. The method for assembling a flexible shaft hole with large deflection based on tactile perception according to claim 1 is characterized in that: In step S2, the tactile sensing deformation control algorithm controls the flexible shaft workpiece and the rigid plane to form effective contact based on the sensed clamping force distribution data, including the following steps: S21, controlling the flexible shaft workpiece to move toward the rigid plane according to the spatial position of the rigid plane, and making the bottom end of the flexible shaft workpiece contact with the upper surface of the rigid plane; S22, the pre-trained CNN network predicts the height z of the flexible shaft workpiece after deformation based on the numerical matrix M formed by the clamping force distribution data sensed by the thin film pressure tactile sensor array est ; S23, determine whether z is satisfied est In the effective deformation range [z min , z max ], if the result is yes, it is determined that the flexible shaft workpiece forms effective contact with the rigid plane, and step S25 is executed; if the result is no, step S24 is executed; S24, controlling the robotic arm to move upward in the vertical direction, and returning to step S22; S25, maintain the height z of the flexible shaft workpiece after deformation est Constant and unchanging.

3. The method for assembling a flexible shaft hole with large deflection based on tactile perception according to claim 1, characterized in that: In step S3, the target hole is searched based on the change in the clamping force distribution data sensed at the current moment and the previous moment as follows: When the change between the numerical matrix formed by the clamping force distribution data sensed by the thin film pressure tactile sensor array at the current moment and the previous moment is greater than the preset threshold, it is determined that the flexible shaft workpiece passes through the target hole at the current moment and the target hole search is completed.

4. The method for assembling a flexible shaft hole with large deflection based on tactile perception according to claim 1, characterized in that: In step S4, the relative position error e between the shaft and the hole is calculated as follows: e=mean(M t -M Target ); M t is the numerical matrix of the clamping force distribution data sensed by the pressure tactile sensor array at the current moment t, M Target is the numerical matrix of the clamping force distribution data sensed by the pressure tactile sensor array when the flexible shaft workpiece is not subjected to any external force except the clamping force, and mean() is the column-wise average of the vector.

5. The method for assembling a flexible shaft hole with large deflection based on tactile perception according to claim 1, characterized in that: In step S4, the implementation method of determining the control amount of the horizontal movement direction and movement distance of the robot arm according to the relative position error between the shaft and the hole includes: When the relative position error between the axis and the hole is greater than the preset relative position tolerance, the PID control generates a control variable for controlling the horizontal movement direction and movement distance of the robot arm based on the relative position error between the axis and the hole.

6. The method for assembling a flexible shaft hole with large deflection based on tactile perception according to claim 1, characterized in that: The gripper is realized with an adaptive two-finger gripper.

7. The method for assembling a flexible shaft hole with large deflection based on tactile perception according to claim 1, characterized in that: An array of thin-film pressure tactile sensors is mounted on one or both sides of the adaptive two-finger gripper.

8. A flexible shaft-hole assembly device based on tactile perception, comprising a storage device, a processor, and a computer program stored in the storage device and executable on the processor, characterized in that: The processor executes the computer program to implement the large deflection deformation flexible shaft hole assembly method based on tactile perception as described in any one of claims 1 to 7.

9. A computer-readable storage device, characterized in that: The storage device stores a computer program, wherein the computer program is configured to implement the tactilely perceived flexible shaft-hole assembly method according to any one of claims 1 to 7 when executed.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the tactilely perceived flexible shaft-hole assembly method according to claims 1 to 7 is implemented.

Images