Robot screw hole assembling method and system based on force feedback and displacement detection

Abstract

The invention discloses a robot screw hole assembling method and system based on force feedback and displacement detection, and belongs to the field of robot automatic assembling. The method comprises the steps that the tail end of the robot is controlled to move to a screw hole coarse positioning position; the tail end is controlled to execute contact movement in the assembling direction; the force component of the tail end six-dimensional force sensor in the assembling direction and the tail end actual displacement are obtained in real time; judging the contact state of the assembly tool and the screw hole based on the logic combination of the force component and the actual displacement; if it is judged that the tool only makes contact with the surface of the workpiece and does not enter the hole, the tail end is controlled to retreat and reposition; if it is judged that the tool enters the screw hole, the tool is controlled to execute tightening operation. The method does not depend on high-precision visual positioning, the contact state is intelligently judged through force and displacement information, the problem of assembly failure caused by robot positioning errors and visual errors is solved, and the method has the advantages of being high in robustness, good in universality, easy to implement and low in cost and is particularly suitable for precision assembly tasks of humanoid robots and industrial mechanical arms.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of robot automatic assembly, and relates to a robot screw hole assembly method and system suitable for a positioning error scene, and particularly suitable for a shaft hole precision assembly task of a humanoid robot or an industrial robot arm. BACKGROUND

[0002] It is a challenging work to apply a humanoid robot or a high degree of freedom robot arm to an industrial assembly field to complete a shaft hole assembly task such as screwing a screw into a screw hole. The successful execution of such a task highly depends on the accurate alignment between an end effector (assembly tool) and a target hole on a workpiece. However, in actual applications, the repeated positioning error of the robot body, the link flexibility, and the recognition error of the vision system (especially for small size and low contrast screw holes) will cumulatively cause an unavoidable deviation (usually in millimeters) in position and angle between the robot end and the hole center. If a pure position-based control strategy is directly used for insertion, hard contact between the screw and the hole edge is likely to occur, which will cause assembly failure, screw or workpiece thread damage, and even damage to the robot joints.

[0003] In order to solve the above problems, the prior art proposes a variety of solutions. One idea is to seek higher precision visual positioning and robot calibration, but this will significantly increase the system cost and complexity, and in some complex lighting conditions or featureless scenes, the effect is limited. Another idea is to introduce force sensing and force control technology.

[0004] For example, the Chinese invention patent application with publication number CN120663319A proposes a "mechanical arm shaft hole assembly method and system based on AER-DDPG algorithm". This method uses a deep reinforcement learning (DDPG) algorithm to let the mechanical arm learn the assembly strategy through a large number of simulations and actual training. Although this method has certain adaptive ability, its defects are: first, deep reinforcement learning requires a large amount of training data and a long training period, and the learning cost is high; second, the learned strategy has poor interpretability ("black box" problem), and may fail when encountering new scenes not covered by the training data, and the reliability is difficult to guarantee; finally, this method assumes a continuous task from grasping to assembly, but the mechanism for how to use force feedback for real-time and robust fine-tuning obstacle avoidance during assembly is not directly and explicitly.

[0005] For example, Chinese invention patent application CN120697043A proposes a "control method and shaft-hole assembly system for over-constrained robots based on six-dimensional force feedback calculation of branch forces." This solution focuses on high-precision force control of a specific configuration—over-constrained parallel robots. It calculates the branch internal forces through a complex mechanical model, thereby achieving precise force-position hybrid control. The limitations of this method are: first, its core is a dedicated force control algorithm for the special structure of over-constrained parallel mechanisms, which has poor versatility and is difficult to directly apply to widely used serial humanoid robots or industrial robotic arms; second, the system is complex, requiring special over-constrained six-dimensional force sensors and corresponding over-constrained robot platforms, resulting in high implementation thresholds and high costs; third, the control algorithm involves complex matrix operations and model calculations, requiring high computing power from the controller.

[0006] In summary, existing technologies either have stringent requirements for perception accuracy and lack fault tolerance, rely on data-driven approaches which carry stability risks, or are too complex to be widely adopted. Therefore, the industry urgently needs a robotic screw hole assembly method that is versatile, easy to implement, does not rely on high-precision prior information, and can handle contact uncertainties during assembly. Summary of the Invention

[0007] This invention aims to overcome the shortcomings of existing technologies and provide a robot screw hole assembly method and system based on force feedback and displacement detection. The core idea of ​​this invention is to not pursue absolutely precise initial visual positioning, but rather allow for a certain degree of coarse positioning error. Instead, it utilizes a six-dimensional force sensor at the robot's end effector and the robot's own displacement information, through a simple yet effective logical judgment rule, to identify the contact state between the screw and the workpiece / screw hole in real time, and intelligently switch control strategies accordingly, thereby achieving reliable automatic assembly. To achieve the above objective, this invention adopts the following technical solution.

[0008] In a first aspect, the present invention provides a robot screw hole assembly method based on force feedback and displacement detection, characterized by comprising the following steps: S1: Control the robot end effector to move the assembly tool to the preset coarse positioning position above the target screw hole; S2: Control the robot end effector to perform contact movement along a direction perpendicular to the workpiece surface toward the location of the screw hole; S3: During the contact movement, the first directional force component detected by the six-dimensional force sensor installed at the end of the robot and the actual displacement of the end of the robot are acquired in real time; Wherein, the first direction is the direction of the contact movement; S4: Based on the first directional force component and the actual displacement, determine the contact state between the assembly tool and the screw hole; S5: Based on the determination result of the contact state, execute the corresponding assembly control strategy; If it is determined that the assembly tool has entered the screw hole, the assembly tool is controlled to perform a tightening operation; if it is determined that the assembly tool is only in contact with the workpiece surface and has not entered the screw hole, the robot end effector is controlled to retract and step S1 is re-executed.

[0009] The method provided by this invention replaces the reliance on high-precision visual positioning by jointly judging force and displacement information, thus solving the assembly problem under initial positioning error.

[0010] Furthermore, in step S1, the preset coarse positioning position is obtained through a visual positioning system, which includes a head camera for initial positioning and / or a wrist camera for auxiliary positioning.

[0011] Furthermore, in step S4, determining the contact state between the assembly tool and the screw hole specifically involves: When the absolute value of the first directional force component is detected to be greater than the first threshold and the absolute value of the actual displacement is less than the second threshold, it is determined that the assembly tool is only in contact with the workpiece surface and has not entered the screw hole. When the absolute value of the first directional force component is detected to be greater than the first threshold, and the absolute value of the actual displacement is greater than or equal to the second threshold, it is determined that the assembly tool has entered the screw hole.

[0012] Furthermore, the first direction is a vertically downward direction, and the force component in the first direction is the force in the Z-axis direction of the six-dimensional force sensor. When the assembly tool comes into contact with the workpiece, this force is negative.

[0013] Furthermore, in step S2, the robot employs Cartesian spatial impedance control, and the relationship between the end-effector force and the end-effector position error satisfies: , in, M The inertia matrix, B Here is the damping matrix. K Here is the stiffness matrix. x For positional error, F ext The external force acting on the end.

[0014] Furthermore, during the contact movement in step S2, a constant displacement or speed command is continuously applied in the direction of the screw hole; when the assembly tool contacts the workpiece surface, due to the impedance control effect, the actual displacement will be less than the commanded displacement, and the value of the actual displacement is calculated in real time by the impedance control algorithm based on the force component in the first direction.

[0015] Furthermore, the robot is a humanoid robot or a serially articulated robotic arm, and the assembly tool is an electric screwdriver or a dexterous hand.

[0016] Secondly, the present invention provides a robot screw hole assembly system based on force feedback and displacement detection for implementing the above method, the system comprising: The robot body has assembly tools installed at its end; A six-dimensional force sensor is installed at the end of the robot to detect the force and torque acting on the end. The controller, which is communicatively connected to the robot body and the six-dimensional force sensor, is configured as follows: Follow the steps described above.

[0017] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that the program, when executed by a processor, implements the steps of the above-described method.

[0018] Fourthly, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the program to implement the steps of the method as described in any one of claims 1-7.

[0019] Compared with the prior art, the present invention has the following beneficial technical effects.

[0020] (1) High robustness and fault tolerance: This invention does not rely on high-precision absolute positioning and allows for initial positioning errors at the millimeter level. Through the joint sensing and intelligent judgment of force and displacement, it can effectively handle the "misalignment" problem caused by various errors. When the screw contacts the edge of the hole, it can automatically identify and trigger repositioning, avoiding failure and damage caused by hard insertion, and greatly improving the success rate of assembly tasks.

[0021] (2) High versatility and simple implementation: The six-dimensional force sensor and impedance control technology used in this invention are mature technologies in the field of robotics and can be widely applied to various serial robots (including humanoid robots and industrial robotic arms). Its core judgment logic is clear and simple, without the need for complex mechanical model calculations (unlike CN120697043A) or large amounts of data training (unlike CN120663319A), making it easy to integrate and deploy in existing robot control systems with low implementation costs and barriers.

[0022] (3) Direct response and good real-time performance: Decisions are made based on clear physical quantities (force, displacement) and logical rules, avoiding the inference delay and uncertainty of machine learning methods. The control response is more direct, fast and stable, which is especially suitable for precision assembly scenarios with high real-time requirements.

[0023] (4) Achieving a fully automated process: This invention seamlessly connects "visual coarse positioning", "force contact search and status judgment" and "decision execution (tightening or retrying)" to form a complete automated screw hole assembly closed-loop process that does not require manual intervention, which significantly improves production efficiency. Attached Figure Description

[0024] Figure 1 This is a schematic diagram showing the relationship between the Z-axis force (Fz) detected by the six-dimensional force sensor of this invention and time.

[0025] Figure 2 This is a partial structural diagram of the screw hole assembly of the present invention.

[0026] Figure 3 The flowchart of the screw hole assembly control method based on force feedback and displacement detection provided by the present invention is shown. Detailed Implementation

[0027] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0028] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance, quantity, or position.

[0029] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0030] Example 1: System Composition This embodiment provides a robot screw hole assembly system based on force feedback and displacement detection, which mainly includes the following parts: Robot body: In this embodiment, it can be a serially articulated humanoid robot arm with multiple degrees of freedom or an industrial six-axis robotic arm. It has the ability to perform precise motion control in Cartesian space.

[0031] End-effector assembly tool: mounted on the end flange of the robot, it can be an electric screwdriver (for tightening screws) or a dexterous hand / gripper that holds screws, depending on the task requirements.

[0032] Six-dimensional force sensor: Installed between the robot's end flange and the assembly tool. Used for high-precision measurement of the force (Fx, Fy, Fz) and torque (Mx, My, Mz) at the tool tip in three orthogonal directions (X, Y, Z). In this embodiment, the sensor coordinate system is defined as follows: the Z-axis pointing vertically downwards is the positive direction.

[0033] The vision system includes a head camera for large-scale coarse positioning and a wrist camera for close-range assisted positioning. The head camera acquires a global image, and visual algorithms (such as template matching and feature recognition) are used to roughly estimate the two-dimensional image position of the screw hole on the worktable. Combined with depth information or a known workpiece model, the rough three-dimensional position of the screw hole relative to the robot can be calculated.

[0034] Robot controller: Electrically or communicatively connected to the robot body, six-dimensional force sensor, vision system, and assembly tools. The controller is responsible for receiving sensor data, performing logical judgments, generating control commands, and sending them to the robot's actuators. When the program within the controller is executed, it implements the steps of the method described in this invention.

[0035] Example 2: Detailed Description of the Method Flow Combination Figure 3 The control flow of the method of the present invention will be described in detail. This example uses the most common scenario of tightening a screw vertically downwards.

[0036] (1) Initial coarse positioning The controller acquires images of the worktable via a head-mounted camera, identifying the approximate area of ​​the workpiece and the target screw hole. Based on this visual information, a safe path is planned, controlling the robot's end effector, carrying the screw (assembly tool), to move to a preset position directly above the screw hole. This position does not require precise alignment with the hole's center, allowing for a horizontal error of a few millimeters. After this step, the screw tip is positioned a short distance above the screw hole, such as... Figure 2 As shown.

[0037] (2) Contact movement and data acquisition The controller directs the robot's end effector to perform "contact movement" in a vertically downward direction (i.e., the -Z direction, the first direction) at a low speed or a constant position increment (e.g., 0.1 mm per control cycle). Crucially, the robot operates in Cartesian impedance control mode during this phase. Its dynamic model is as follows:

[0038] in, x This is the error between the commanded position and the actual position at the end of the line. This can be addressed by setting an appropriate stiffness matrix. K (Usually set to a lower value) and damping matrix B The robot can track commands while moving in free space, but once the end effector comes into contact with the environment... F ext (≠0), it will "yield" like a spring, the actual displacement x actual The displacement is smaller than the command displacement under pure position control. x command .

[0039] During this process, the controller executes the following sub-steps in real time (e.g., 1000 times per second): Read the Z-force at the current moment from the six-dimensional force sensor. F z (First directional force component). When not in contact, Fz ≈0 (may include minute values ​​after gravity compensation). When the screw tip contacts the workpiece surface, the workpiece exerts an upward vertical reaction force on the screw, which is detected by the sensor. F z It will become a negative value, such as Figure 1 As shown after time t1.

[0040] Calculate the actual displacement of the robot's end effector in the current contact movement direction (Z direction). D z This displacement is not a simple accumulation of commanded displacements, but rather, within an impedance control framework, it is determined based on the current applied force. F ext The actual end-effector position change is calculated in real time from the controller parameters. Alternatively, the actual end-effector pose can be calculated directly from the robot's forward kinematics, based on feedback from the joint encoders, and then subtracted from the pose at the start of the initial contact movement to obtain the final position. D z .

[0041] (3) Contact state logic judgment The controller will collect data in real time. F z and D zThe screw's exact condition is determined by comparing it with a preset threshold.

[0042] Case A (surface contact only): | was detected at time t1 F z |> F th (The first threshold, e.g., 5N), indicates that contact has occurred. However, continuous monitoring of | D z ∣< D th (Second threshold, e.g., 0.5mm). This means that although an upward force is applied, the robot undergoes almost no downward yield displacement under impedance control. Physically, this corresponds to the screw tip just touching the workpiece plane around the screw hole, or getting stuck at the edge of the hole, where the end effector cannot continue to move downward due to the lack of clearance at the contact point (except for slight material deformation). In this case, the status flag is set to "Fail".

[0043] Case B (already slid into the hole): | was detected at time t1. F z |> F th Subsequently, at time t2 and beyond, | D z |≥ D th And it continues to increase. This indicates that after contact occurs, the robot end effector produces a significant downward displacement under impedance control. This typically corresponds to the screw tip contacting the chamfer at the screw hole entrance. Since the chamfer is an inclined plane, when the screw is subjected to a downward command force, the chamfer surface generates a lateral component force on the screw. This component force guides the screw tip to "slide" towards the center of the hole, while allowing the end effector to displace in the Z direction. At this time, the status flag is set to "Success".

[0044] (4) Decision-making and execution The branching strategy is executed based on the judgment result of the contact state logic judgment in step (3): If the result is "Fail" (surface contact only): the controller immediately stops the downward contact movement command. The robot end effector is slightly raised along the original path (e.g., moved up 2-3 mm) to disengage from contact. Then, the end effector position can be fine-tuned in the horizontal plane based on a certain search strategy (e.g., spiral search, grid search), and then the system returns to step (1) for initial coarse positioning or step (2) to retry the contact movement. This mechanism ensures that when the initial positioning deviation is too large to utilize chamfer self-guidance, the system can automatically retry instead of forcibly pressing and causing damage.

[0045] If the result is "Success" (screw already in hole): The controller determines that the screw has been initially aligned and entered the screw hole using the chamfer. At this point, the controller can maintain the impedance control mode and continue to move down a preset distance to ensure that the screw thread is also in the screw hole. Subsequently, the controller sends a command to the electric screwdriver to start its rotation and perform the tightening operation until the preset torque or number of turns is reached, completing the assembly task.

[0046] Example 3: Threshold Selection and Parameter Setting First threshold F th The setting should be based on the sensor noise, gravity compensation residual, and desired contact sensitivity, and is usually set to a value slightly higher than the system noise level, such as 2N~10N.

[0047] Second threshold D th This is a key parameter, and its selection is related to the size of the chamfer on the threaded hole and the desired discrimination sensitivity. For standard threaded holes, the chamfer depth is typically between 0.2mm and 1mm. Therefore, D th It can be set to a value slightly smaller than the typical chamfer depth, such as 0.3mm to 0.8mm. This ensures that when the screw slides on the chamfer and causes a small displacement, it is promptly recognized as successful, while avoiding false judgments caused by sensor noise or minor structural deformation.

[0048] Impedance control parameters ( M , B , K The stiffness needs to be adjusted based on the robot's characteristics, assembly speed requirements, and contact stability. Generally speaking, in assembly tasks, stiffness... K It should be set relatively low to provide sufficient flexibility and prevent excessive contact force.

[0049] This invention creatively combines the contact force detection of a six-dimensional force sensor with the actual end displacement information generated by impedance control, and designs an extremely simple and efficient binary logic judgment rule. This rule cleverly utilizes two completely different physical phenomena, "force without displacement = pressing against the surface" and "force with displacement = sliding into the chamfer / hole", transforming the complex shaft-hole alignment problem into a binary state machine problem that can be monitored and judged in real time.

[0050] Compared to CN120663319A, this invention eliminates the uncertainty of data-driven approaches and adopts rule-based control based on a defined physical model, resulting in greater reliability and efficiency. Compared to CN120697043A, this invention avoids complex mechanisms and obscure calculations, directly utilizing commercial sensors and standard robot interfaces, significantly improving versatility, ease of use, and cost-effectiveness. This invention provides a novel, practical, and efficient path to solving the "last millimeter" alignment problem in robot shaft and hole assembly.

[0051] Although embodiments of the present invention have been shown and described above, it is understood that these embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and alterations to the above embodiments within the scope of the present invention without departing from its principles and spirit. The scope of protection of the present invention is defined by the claims and their equivalents.

Claims

1. A robot screw hole assembly method based on force feedback and displacement detection, characterized in that, Includes the following steps: S1: Control the robot end effector to move the assembly tool to the preset coarse positioning position above the target screw hole; S2: Control the robot end effector to perform contact movement along a direction perpendicular to the workpiece surface toward the location of the screw hole; S3: During the contact movement, the first directional force component detected by the six-dimensional force sensor installed at the end of the robot and the actual displacement of the end of the robot are acquired in real time; Wherein, the first direction is the direction of the contact movement; S4: Based on the first directional force component and the actual displacement, determine the contact state between the assembly tool and the screw hole; S5: Based on the determination result of the contact state, execute the corresponding assembly control strategy; If it is determined that the assembly tool has entered the screw hole, the assembly tool is controlled to perform a tightening operation; if it is determined that the assembly tool is only in contact with the workpiece surface and has not entered the screw hole, the robot end effector is controlled to retract and step S1 is re-executed.

2. The method according to claim 1, characterized in that, In step S1, the preset coarse positioning position is obtained through a visual positioning system, which includes a head camera for initial positioning and / or a wrist camera for auxiliary positioning.

3. The method according to claim 1 or 2, characterized in that, In step S4, determining the contact state between the assembly tool and the screw hole specifically involves: When the absolute value of the first directional force component is detected to be greater than the first threshold and the absolute value of the actual displacement is less than the second threshold, it is determined that the assembly tool is only in contact with the workpiece surface and has not entered the screw hole. When the absolute value of the first directional force component is detected to be greater than the first threshold, and the absolute value of the actual displacement is greater than or equal to the second threshold, it is determined that the assembly tool has entered the screw hole.

4. The method according to claim 3, characterized in that, The first direction is the vertically downward direction, and the force component in the first direction is the force in the Z-axis direction of the six-dimensional force sensor. When the assembly tool comes into contact with the workpiece, this force is negative.

5. The method according to claim 1, characterized in that, In step S2, the robot employs Cartesian spatial impedance control, and the relationship between the end effector force and the end effector position error satisfies: ， in, M The inertia matrix, B Here is the damping matrix. K Here is the stiffness matrix. x For positional error, F ext The external force acting on the end.

6. The method according to claim 5, characterized in that, During the contact movement in step S2, a constant displacement or speed command is continuously applied in the direction of the screw hole. When the assembly tool contacts the workpiece surface, due to the impedance control effect, the actual displacement will be less than the commanded displacement. The value of the actual displacement is calculated in real time by the impedance control algorithm based on the force component in the first direction.

7. The method according to claim 1, characterized in that, The robot is a humanoid robot or a serially articulated robotic arm, and the assembly tool is an electric screwdriver or a dexterous hand.

8. A robotic screw hole assembly system based on force feedback and displacement detection, characterized in that, The system for implementing the method as described in any one of claims 1-7 comprises: The robot body has assembly tools installed at its end; A six-dimensional force sensor is installed at the end of the robot to detect the force and torque acting on the end. The controller, which is communicatively connected to the robot body and the six-dimensional force sensor, is configured as follows: Perform the method steps as described in any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method as described in any one of claims 1-7.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method as described in any one of claims 1-7.

Images