Assembly element insertion hole pose detection method based on tactile feedback

Abstract

The application discloses an assembly element insertion hole pose detection method based on tactile feedback, which comprises the following steps: a bionic finger containing a tactile sensitive element is configured at the end of a robot, the voltage signals of the tactile sensitive elements are collected by sliding at a constant speed on the plane where the assembly hole is located, a tactile data matrix is generated through filtering, difference and weighting processing, and hole position detection is realized through threshold comparison; coordinates are calculated according to the motion trail of the robot and the hole position data, and the assembly pose is determined by fusing the geometric characteristics of the element, so that the robot is finally guided to adjust the pose and complete the high-precision insertion hole task. The application simulates the exploration and perception mode of a human hand in a weak light environment, detects and positions the scene where visual recognition is difficult to perform, and helps the robot to realize accurate assembly.

Description

Technical Field

[0001] This invention belongs to the field of robot intelligent perception and operation, specifically relating to a method for detecting the pose of assembly component sockets based on tactile feedback. Background Technology

[0002] Assembly is an essential step in the manufacturing of most electromechanical products, such as assembling leaded resistors and capacitors, inserting electronic connectors, and installing multi-pin housings. However, current product assembly is mostly done by assembly line workers, resulting in high labor costs, insufficient automation, and room for improvement in product consistency. Utilizing robots for product assembly has enormous market potential and application value. However, current robot movement largely relies on vision guidance, while the assembly process frequently involves contact and interaction between mechanical grippers, the object to be assembled, and the surrounding environment, leading to significant visual obstruction and hindering its application.

[0003] Therefore, in order to improve the adaptability of mechanical assembly to complex environments, it is necessary to develop a tactile feedback-based intelligent hole detection and positioning method for robots to solve the problem of not being able to complete precision assembly under conditions such as severe occlusion during visual guidance. Summary of the Invention

[0004] To address the aforementioned issues, this invention discloses a method for detecting the pose of assembly component sockets based on tactile feedback. This method equips robots with tactile bionic fingers, simulating the exploration and perception methods of human hands in low-light or dark environments, thereby enhancing their autonomous operation capabilities in complex and dynamic environments. By collecting sensitive component signals through multiple channels and employing threshold detection methods, this method improves the robot's work efficiency in actual production and adapts to diverse assembly tasks.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] A method for detecting the pose of assembly component sockets based on tactile feedback includes the following steps:

[0007] (1) Creating tactile bionic fingers

[0008] The tactile bionic finger is composed of a flexible skin layer and rigid phalanges, with tactile sensitive elements embedded in the flexible skin layer to sense static and dynamic forces.

[0009] (2) By tracing the surface to be assembled with a tactile bionic finger, the presence of mounting holes on the surface to be assembled is detected, and the electrical signals of multiple sensitive components collected simultaneously are fused and judged.

[0010] (3) Compare the hole positions and their distribution relationship detected in the previous step with the number of pins and geometric distribution characteristics of the component to be inserted, determine the hole position combination corresponding to the current component, determine its insertion position based on the hole position combination, and determine the orientation angle of the component in the assembly plane through the geometric distribution characteristics of the hole position combination.

[0011] Furthermore, in step (1), the static force tactile sensing element is made of piezoresistive material (Strain Gauge, abbreviated as SG), and the dynamic force tactile sensing element is made of piezoelectric sensing element polyvinylidene Difluoride film (PVDF, abbreviated as PVDF), which is used to convert external tactile signals into electrical signals.

[0012] The number of tactile sensitive elements in this tactile bionic finger is greater than or equal to 2. The more sensitive elements there are, the higher the recognition accuracy.

[0013] Furthermore, in step (2), the method for detecting the mounting hole positions is as follows:

[0014] S0: The component has pins with known quantity, size parameters and geometric distribution characteristics, and correspondingly distributed multiple sets of mounting holes with unknown position and orientation angle, but whose quantity, size parameters and geometric distribution characteristics match the above-mentioned component, on the surface to be assembled. The pins of the component and the mounting holes are precisely assembled at appropriate positions and orientation angles.

[0015] S1: Establish a Cartesian coordinate system within the surface to be assembled by selecting a suitable origin and orientation angle. Specifically, divide the surface along the Y-axis into m equal-spaced sliding paths parallel to the X-axis. Control the robot's tactile bionic finger to slide along these m sliding paths sequentially in ascending order of Y-coordinate, simultaneously acquiring tactile (voltage) signals from n sensitive element channels in the tactile bionic finger and storing these signals in the original signal matrix M1. Matrix element x ij This represents the tactile signal collected by the robot's tactile bionic finger on the i-th channel and the j-th sliding path, namely:

[0016]

[0017] S2: The acquired tactile signals are sequentially processed by differential, smoothing filtering, and normalization. The processing results of the n channels are then weighted and summed, where the summation result X of the j-th sliding path is... j as follows:

[0018] X j =w1x 1j +w2x 2j +…+w n x nj

[0019] In the above formula, w n This represents the weight coefficient of the nth channel. The magnitude of the weight coefficient is determined by the type of sensitive element and its spatial distribution in the finger. The summation result is stored in matrix M2, with element X... j This represents the tactile signal collected on the j-th sliding path after the above processing:

[0020] M2 = [X1, X2, ..., X j ,…,X m ]

[0021] The data from each channel of the tactile sensor are weighted and summed. The weighting coefficients are determined by the type of tactile sensor and its spatial distribution within the finger. Generally, the more sensitive the tactile sensor is to external tactile signals, the larger its corresponding weighting coefficient.

[0022] S3: In X j Signal segments collected when the tactile bionic finger slides above the assembly hole were detected and filtered from the signal.

[0023] Specifically, pore location detection is achieved through threshold comparison, where a selection threshold is used. Representing signal X j The mean of the signal X, σ represents the signal X. j The standard deviation of , where k is the threshold scaling factor, obtained from preliminary experiments.

[0024] In the preliminary experiment, the tactile fingertip needs to be repeatedly swiped over the assembly hole. It can be observed that the amplitude of the tactile signal will change significantly when the tactile fingertip slides over the assembly hole. By continuously adjusting the threshold coefficient k, the computer can accurately capture the corresponding signal changes. The corresponding k value is the appropriate threshold coefficient.

[0025] Let t be the tactile signal sequence X. j The index value of X j Compared with the threshold F, X satisfies j The signal segment with (t) ≤ F can be considered as the tactile signal collected when the tactile bionic finger slides above the assembly hole. The signal segments that meet the above comparison results are extracted and stored in matrix M3. The element a... ij This represents the i-th signal segment detected when sliding on the j-th sliding path, and c represents the maximum number of mounting holes detected under a single sliding path.

[0026]

[0027] Furthermore, in step (3), the method for achieving assembly hole positioning and autonomous component assembly is as follows:

[0028] S4: Determine the X coordinate of the assembly hole

[0029] Specifically, for the tactile signal acquired in a single sliding path, the minimum value within each effective signal segment in matrix M3 is considered as the tactile signal acquired when the tactile bionic finger slides across the center of the assembly hole. The index value corresponding to the minimum value is extracted, and the X-coordinate of each assembly hole center traversed by the current sliding path is calculated by combining the sliding speed of the tactile fingertip and the sampling rate of the tactile sensitive element. The above steps are repeated to process the tactile signal of each sliding path until the X-coordinates of the assembly hole centers under all m sliding paths are filtered and calculated.

[0030] S5: Considering that the tactile fingertip may repeatedly touch the same mounting hole under different sliding trajectories, thus calculating duplicate X-coordinate values, a threshold judgment method is needed to eliminate the calculation of duplicate X-coordinate values. Specifically, a threshold x0 is introduced. When the difference between the X-coordinates of the mounting hole calculated under different sliding paths is less than the threshold x0, these similar X-coordinates are regarded as the coordinate values ​​of the same mounting hole or mounting holes with the same X-coordinate. The weighted average of all the above similar X-coordinate values ​​is taken to replace all duplicate X-coordinate values. Furthermore, the X-coordinate values ​​of the eliminated duplicate data are stored in the array point_x in ascending order.

[0031] S6: Determine the Y-coordinate value of the assembly hole. Specifically, based on the X-position represented by each element of the array point_x, plan a path parallel to the Y-axis. Control the robot to slide across the plane to be measured along the direction of increasing Y-coordinate, and process the collected signals according to the methods of S1 to S4 to obtain the Y-coordinate point_y of the assembly hole.

[0032] S7: Determine component assembly information. Specifically, based on the X-coordinate array point_x and Y-coordinate array point_y collected above, the coordinate positions of the centers of each assembly hole on the plane are obtained sequentially. The assembly holes on the plane are grouped according to the number of pins, size parameters, and geometric distribution characteristics of the components. The geometric center inside each group of assembly holes and its orientation angle in the assembly plane are calculated. Based on the center position and geometric tilt angle of each group of assembly holes, the robotic arm is driven to assemble the components into each group of assembly holes at appropriate angles, thus completing the autonomous assembly task of the components.

[0033] This intelligent detection and positioning method is not limited to the insertion of single-pin, two-pin, and three-pin components, but is also applicable to the assembly of more pin components in the plane through the shaft insertion hole.

[0034] The beneficial effects of this invention are as follows:

[0035] This invention proposes a simple algorithm that uses filtering and noise reduction, threshold judgment and other methods to determine the position of the hole, which is suitable for the field of robot assembly.

[0036] This invention weights and sums the data from tactile sensing elements, fully leveraging the advantages of multi-channel information fusion to avoid the impact of data fluctuations from individual sensing elements on the results, thus improving assembly accuracy. Simultaneously, by estimating the minimum distance between holes, it eliminates instances where parts repeatedly slide past the same hole, resulting in more accurate results. This allows for precise identification of hole positions, ensuring a higher degree of matching between parts and holes, thereby significantly improving the accuracy of robot assembly and reducing product quality issues caused by assembly errors. For example, in the assembly of automotive parts, it enables precise insertion of bolts into matching threaded holes, preventing loosening or misalignment.

[0037] This invention enhances the autonomy and adaptability of robots. When faced with assembly tasks requiring different assembly environments, the tactile signal recognition algorithm can assist the robot in autonomously adjusting its assembly strategy to adapt to diverse production needs. For example, in determining the location of assembly holes on unknown flat plates, the collection of tactile signals overcomes the difficulties in observation and quantitative measurement. Attached Figure Description

[0038] Figure 1 This is a flowchart of the identification method of the present invention.

[0039] Figure 2 This is a schematic diagram of the tactile bionic finger structure and a flowchart of the hole detection process in this invention.

[0040] In the diagram: 10, flexible material layer; 20, rigid finger bone.

[0041] Figure 3 This is a schematic diagram of the testing device for an example of the present invention.

[0042] In the diagram: 1. Component to be assembled; 2. Plane to be tested; 3. Assembly hole to be tested; 4. Tactile fingertip; 5. Mechanical gripper; 6. Robotic arm; 7. Single assembly hole to be tested; 8. Double assembly hole to be tested; 9. Triple assembly hole to be tested.

[0043] a~d, four tactile fingertips sliding paths at equal intervals along the X-axis; e~h, four sliding paths along the Y-axis. Detailed Implementation

[0044] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0045] A method for detecting the pose of assembly component sockets based on tactile feedback includes the following steps:

[0046] (1) Creating tactile bionic fingers

[0047] The tactile bionic finger is composed of a flexible skin layer and rigid phalanges, with tactile sensitive elements embedded in the flexible skin layer to sense static and dynamic forces.

[0048] (2) The fingertip 4 of the tactile bionic finger is parallel to the test plane 2 to detect whether there are mounting holes on the test plane 2, and the electrical signals of multiple sensitive elements collected at the same time are fused and judged.

[0049] (3) Compare the hole positions and their distribution relationship detected in the previous step with the number of pins and geometric distribution characteristics of the component to be inserted, determine the hole position combination corresponding to the current component, determine its insertion position based on the hole position combination, and determine the orientation angle of the component in the assembly plane through the geometric distribution characteristics of the hole position combination.

[0050] Specifically, in step (1), as Figure 2 The robot's tactile fingertip shown employs a hybrid structure of a flexible skin layer, a tactile sensing layer, and a rigid phalanx. Internally, it integrates multiple dynamic force-sensitive elements made of polyvinylidene fluoride (PVDF) film and static force-sensitive elements made of resistive strain gauges (SG), used to convert external tactile signals into electrical signals.

[0051] Specifically, the outermost flexible skin layer is made of silicone, which can transmit external static force signals (pressure) and dynamic force signals (vibration) to the internal sensitive element relatively completely.

[0052] Specifically, in this design, one side of the static force tactile sensing element is embedded in the flexible skin layer, while the other side is tightly attached to the rigid phalanx. The dynamic force tactile sensing element has both sides in contact with the flexible layer, is entirely embedded in the flexible material, and does not contact the rigid phalanx. In this design, the static force tactile sensing element has a rigid phalanx as support on one side. The rigid phalanx has the advantage of smaller vibration amplitude compared to the flexible material, which reduces the impact of the vibration signal from the flexible material layer on the static force tactile sensing element, making it more conducive to the static force tactile sensing element acquiring static pressure signals. The dynamic force tactile sensing element, entirely embedded in the flexible material layer, has no rigid material for support, allowing the vibration signal from the flexible material to directly act on the dynamic force tactile sensing element, making it more conducive to the dynamic force tactile sensing element acquiring vibration signals.

[0053] Specifically, in step (2), the method for detecting the hole position is as follows:

[0054] S0: The component has pins with known quantity, size parameters and geometric distribution characteristics. On the assembly plane, there are correspondingly multiple sets of assembly holes with unknown position and orientation angle, but whose quantity, size parameters and geometric distribution characteristics match the above-mentioned component. The pins of the component can be precisely assembled with the assembly holes at appropriate position and orientation angle.

[0055] Specifically, the plate under test has a single assembly hole 7, two assembly holes 8 with different spacings of L1 and L2 respectively, and three assembly holes 9 with equal spacing. At the same time, the number, size parameters and geometric distribution characteristics of the components are consistent with the small holes under test.

[0056] S1: Establish a Cartesian coordinate system within the assembly plane by selecting a suitable origin and orientation angle. Specifically, divide the plane into m equal-spaced sliding paths parallel to the X-axis along the Y-axis. Control the robot's tactile bionic finger to slide along these m paths sequentially in ascending order of Y-coordinate, traversing the test plane 2. Simultaneously, collect tactile (voltage) signals from n sensitive element channels in the tactile bionic finger and store these signals in the original signal matrix M1. Matrix element x ij This represents the tactile signal collected by the robot's tactile bionic finger on the i-th channel and the j-th sliding path, namely:

[0057]

[0058] S2: The acquired tactile signals are sequentially processed by differential, smoothing filtering, and normalization. The processing results of the n channels are then weighted and summed, where the summation result X of the h-th sliding path is... j as follows:

[0059] X j =w1x 1j +w2x 2j +…+w n x nj

[0060] Specifically, in the above formula, w n This represents the weight coefficient of the nth channel. The magnitude of the weight coefficient is determined by the type of sensitive element and its spatial distribution in the finger. The summation result is stored in matrix M2, with element X... j This represents the tactile signal collected on the j-th sliding path after the above processing:

[0061] M2 = [X1, X2, ..., X j ,…,X m ]

[0062] Specifically, the more sensitive the tactile sensing element is to external tactile signals, the larger its corresponding weighting coefficient.

[0063] Specifically, in this experiment, the PVDF dynamic force-sensitive unit reacts more to the vibration signal of the flexible material during the tactile finger sliding process and has a larger weighting coefficient.

[0064] S3: Further, in X j The signal segment acquired when the tactile bionic finger slides above the assembly hole is detected and filtered from the signal. Specifically, the hole position is detected through threshold comparison, where the filtering threshold... Representing signal X j The mean of the signal X, σ represents the signal X. j The standard deviation of is given, and k is the threshold scaling factor, obtained from preliminary experiments. Let t be the tactile signal sequence X. j The index value of X j Compared with the threshold F, X satisfies j The signal segment with (t) ≤ F can be considered as the tactile signal collected when the tactile bionic finger slides above the assembly hole. The signal segments that meet the above comparison results are extracted and stored in matrix M3. The element a... ij This represents the i-th signal segment detected when sliding on the j-th sliding path, and c represents the maximum number of mounting holes detected under a single sliding path.

[0065]

[0066] Specifically, when the tactile fingertip 4 slides at a constant speed on the test plane 2, the surface layer of the flexible material directly slides relative to the plate, simultaneously subjected to static force (pressure) and dynamic force (vibration). During constant speed, the voltage signal remains stable. Upon encountering a small hole, the contact area between the flexible material surface and the plate decreases, causing a sudden change in the static force (pressure) and dynamic force (vibration). After differential processing in step S2, this results in a decreasing trend in the voltage signal. Therefore, X is satisfied. j The signal segment (t)≤F can be regarded as the tactile signal collected when the tactile bionic finger slides above the assembly hole, where the hole exists.

[0067] Specifically, the threshold scaling factor k needs to be obtained through preliminary experiments. In the preliminary experiments, the tactile fingertip needs to be repeatedly swiped across the top of the mounting hole. It can be observed that the amplitude of the tactile signal will change significantly when the tactile fingertip slides across the mounting hole. By continuously adjusting the threshold scaling factor k, the computer can accurately capture the corresponding signal changes, and the corresponding k value is the appropriate threshold scaling factor.

[0068] Specifically, in step (3), the method for achieving assembly hole positioning and autonomous component assembly is as follows:

[0069] S4: Determine the X coordinate of the assembly hole.

[0070] Specifically, for the tactile signal acquired in a single sliding path, the minimum value within each effective signal segment in matrix M3 is considered as the tactile signal acquired when the tactile bionic finger slides across the center of the assembly hole. The index value corresponding to the minimum value is extracted, and the X-coordinate of each assembly hole center traversed by the current sliding path is calculated by combining the sliding speed of the tactile fingertip and the sampling rate of the tactile sensitive element. The above steps are repeated to process the tactile signal of each sliding path until the X-coordinates of the assembly hole centers under all m sliding paths are filtered and calculated.

[0071] S5: Eliminate duplicate X-coordinate values. Specifically, a threshold x0 is introduced. When the difference between the calculated X-coordinates of the assembly hole under different sliding paths is less than the threshold x0, these similar X-coordinates are considered as the coordinate values ​​of the same assembly hole or assembly holes with the same X-coordinate. The weighted average of all the above similar X-coordinate values ​​is used to replace all duplicate X-coordinate values. Furthermore, the X-coordinate values ​​of the eliminated duplicate data are stored in the array point_x in ascending order.

[0072] Specifically, considering that the tactile fingertip may repeatedly touch the same assembly hole under different sliding trajectories, thus calculating repeated X-coordinate values, a threshold judgment method needs to be introduced to eliminate and merge the above redundant data.

[0073] Specifically, with Figure 3 In the sliding paths a to d, which are equally spaced along the Y-axis and parallel to the X-axis, when the fingertip moves along sliding paths c and d, the X-coordinate values ​​of the small holes will be very similar after judgment and calculation by methods S1 to S4. In order to ensure the accuracy of the assembly hole position information, a threshold x0 is introduced to eliminate the X-coordinate values ​​of duplicate hole positions. The threshold x0 is determined by the minimum inner diameter of the assembly hole in the experiment.

[0074] S6: Determine the Y-coordinate value of the assembly hole. Specifically, based on the X-position represented by each element of the array point_x, plan a path parallel to the Y-axis. Control the robot to slide across the plane to be measured along the direction of increasing Y-coordinate, and process the collected signals according to the methods of S1 to S4 to obtain the Y-coordinate point_y of the assembly hole.

[0075] Specifically, in the four tactile fingertip sliding paths a to d, after judgment and calculation in S1 to S5, the array point_x has recorded the X coordinate of the small hole in the process. Then, the tactile fingertip is controlled to slide along the path parallel to the Y axis at the position where the X coordinate exists. The collected signal is processed in the manner of S1 to S4 to obtain the Y coordinate point_y of the assembly hole.

[0076] S7: Determine component assembly information. Specifically, based on the X-coordinate array point_x and Y-coordinate array point_y collected above, the coordinate positions of the centers of each assembly hole on the plane are obtained sequentially. The assembly holes on the plane are grouped according to the number of pins, size parameters, and geometric distribution characteristics of the components. The geometric center inside each group of assembly holes and its orientation angle in the assembly plane are calculated. Based on the center position and geometric tilt angle of each group of assembly holes, the robotic arm is driven to assemble the components into each group of assembly holes at appropriate angles, thus completing the autonomous assembly task of the components.

[0077] It should be noted that the above content merely illustrates the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. For those skilled in the art, various improvements and modifications can be made without departing from the principle of the present invention, and all such improvements and modifications fall within the scope of protection of the claims of the present invention.

Claims

1. A method for detecting the pose of assembly component sockets based on tactile feedback, characterized in that: Includes the following steps: (1) Making tactile bionic fingers The tactile bionic finger is composed of a flexible skin layer and rigid phalanges, with tactile sensitive elements embedded in the flexible skin layer to sense static and dynamic forces. (2) By tracing the surface to be assembled with a tactile bionic finger, the presence of mounting holes on the surface to be assembled is detected, and the electrical signals of multiple sensitive components collected simultaneously are fused and judged. (3) Compare the hole positions and their distribution relationship detected in the previous step with the number of pins and geometric distribution characteristics of the component to be inserted, determine the hole position combination corresponding to the current component, determine its insertion position based on the hole position combination, and determine the orientation angle of the component in the assembly plane through the geometric distribution characteristics of the hole position combination. In step (2), the method for realizing hole position detection is as follows: S0: The component has pins with known quantity, size parameters and geometric distribution characteristics, and correspondingly distributed multiple sets of mounting holes with unknown position and orientation angle, but whose quantity, size parameters and geometric distribution characteristics match the above-mentioned component on the assembly plane. The pins of the component can be precisely assembled with the mounting holes at appropriate position and orientation angle. S1: Select a suitable origin and direction angle within the above-mentioned assembly plane to establish a plane rectangular coordinate system; Specifically, it is divided into equal intervals along the Y-axis direction. A sliding path parallel to the X-axis is used to control the robot's tactile bionic finger to slide along the path in sequence with increasing Y-coordinates. A sliding path glides across the assembly plane, simultaneously collecting tactile feedback from the bionic finger. The tactile signals from each sensitive element channel are collected and stored in the original signal matrix. In the middle; matrix elements The representative tactile bionic finger of the robot in the first The first channel, the first The tactile signals collected along the sliding path, namely: ； S2: The acquired tactile signals are sequentially processed by differential filtering, smoothing filtering, and normalization, and then... The processing results of each channel are weighted and summed, where the first channel's result is the sum of the results of the second channel's result. The summation result of the sliding paths as follows: ； In the above formula Indicates the first The weighting coefficients for each channel are determined by the type of the sensing element and its spatial distribution within the finger; the summation results are then stored in a matrix. In, elements This indicates that after the above processing, on the [number]th [day / month]... Tactile signals collected along the sliding path: ； S3: Furthermore, in The signal segments collected when the tactile bionic finger slides above the assembly hole were detected and filtered from the signal. Specifically, pore location detection is achieved through threshold comparison, where a selection threshold is used. , Representative signal The mean, Representative signal standard deviation Let be the threshold scaling factor, obtained from preliminary experiments; let be tactile signal sequence The index value will With threshold Comparison, meets the requirements The signal segments are considered as tactile signals collected when the tactile bionic finger slides above the assembly hole. Signal segments that match the above comparison results are extracted and stored in a matrix. In; among which elements Indicates the first The first detected when sliding on the sliding path A signal segment, This indicates the maximum number of mounting holes detected in a single sliding path: 。 2. The method for detecting the pose of assembly component sockets based on tactile feedback according to claim 1, characterized in that: Step (1) The outermost flexible material layer is made of silicone, the static force tactile sensitive element is made of piezoresistive material, and the dynamic force tactile sensitive element is made of piezoelectric sensitive element.

3. The method for detecting the pose of assembly component sockets based on tactile feedback according to claim 1, characterized in that: Step (1) One side of the static force tactile sensing element is embedded in the flexible skin layer, and the other side is in close contact with the rigid finger bone; both sides of the dynamic force tactile sensing element are in contact with the flexible layer, and are all embedded in the flexible material, without contact with the rigid finger bone.

4. The method for detecting the pose of assembly component sockets based on tactile feedback according to claim 1, characterized in that: In step (3), the method for achieving assembly hole positioning and autonomous component assembly is as follows: S4: Determine the X coordinate of the assembly hole Specifically, for tactile signals acquired along a single sliding path, the matrix... The minimum value within each effective signal segment is considered as the tactile signal collected when the tactile bionic finger slides across the center of the assembly hole. The index value corresponding to the minimum value is extracted, and the X coordinate of each assembly hole center traversed by the tactile fingertip and the sampling rate of the tactile sensitive element is calculated. The above steps are repeated to process the tactile signal of each sliding path until the X coordinates of the assembly hole centers under all m sliding paths are filtered and calculated. S5: Eliminate duplicate X-coordinate values Specifically, a threshold is introduced. When the difference in the X coordinates of the assembly hole calculated under different sliding paths is less than a threshold At this time, these similar X-coordinates are considered as the coordinate values ​​of the same assembly hole or assembly holes with the same X-coordinate; the weighted average of all the above similar X-coordinate values ​​is taken to replace all duplicate X-coordinate values; furthermore, the X-coordinate values ​​that have eliminated duplicate data are stored in an array from smallest to largest. middle; S6: Determine the Y-coordinate value of the assembly hole. Specifically, through array For each element representing the X position, plan a path parallel to the Y axis; control the robot to slide across the plane to be measured along the direction of increasing Y coordinate, and process the collected signals according to S1~S4 to obtain the Y coordinate of the assembly hole. ; S7: Determine component assembly information Specifically, based on the X-coordinate array collected above With Y coordinate array The coordinates of the centers of each assembly hole on the plane are obtained by sequentially combining them; the assembly holes on the plane are grouped according to the number of pins, size parameters and geometric distribution characteristics of the components, and the geometric center and orientation angle of each group of assembly holes are calculated. Based on the center position and geometric tilt angle of each group of assembly holes, the robotic arm is driven to assemble the components into each group of assembly holes at a suitable angle, thus completing the autonomous assembly task of the components.

5. The method for detecting the pose of assembly component sockets based on tactile feedback according to claim 1, characterized in that: In step S2, the data from each channel of the sensitive element are weighted and summed. The weighting coefficients are determined by the type of the sensitive element and its spatial distribution in the finger. The more sensitive a tactile element is to external tactile signals, the larger its corresponding weighting coefficient.

6. The method for detecting the pose of assembly component sockets based on tactile feedback according to claim 1, characterized in that: In step S3, the threshold scaling factor It needs to be obtained through preliminary experiments; In the preliminary experiment, the tactile fingertip was repeatedly swiped across the mounting hole. It was observed that the amplitude of the tactile signal changed significantly when the fingertip slid across the mounting hole. This was achieved by continuously adjusting the threshold coefficient. To enable the computer to accurately capture the corresponding signal changes, the corresponding The value is the appropriate threshold coefficient.

7. The method for detecting the pose of assembly component sockets based on tactile feedback according to claim 4, characterized in that: In step S5, considering that the tactile fingertip may repeatedly touch the same assembly hole under different sliding trajectories, thus calculating repeated X coordinate values, a threshold judgment method needs to be introduced to eliminate and merge redundant data.

Images