A method and system for high-precision control of clamping force of parts
By mining and analyzing historical clamping data, high-frequency clamping patterns and optimal parameter combinations are identified, solving the problem of rigid forklift clamping control and achieving high-precision and adaptive clamping control.
Patent Information
- Application Number
- CN202511309662.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-15
- Publication Date
- 2025-12-02
- Estimated Expiration
- 2045-09-15
AI Technical Summary
In existing technologies, the forklift clamping process cannot achieve high-precision automated control, resulting in rigid clamping solutions that cannot adapt to different working conditions and may lead to parts becoming loose or surface damage.
By mining and analyzing historical clamping operation data, we can identify high-frequency clamping patterns and optimal parameter combinations for different part types. We can also use a multi-level support threshold evaluation mechanism to dynamically recommend the most reliable clamping position and clamping force.
It improves the accuracy and adaptability of clamping parameter decisions, ensures high precision and stability in the clamping process, avoids parts from loosening or surface damage, and achieves adaptive optimization control.
Smart Images
Figure CN120793808B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of clamping parameter control technology, specifically to a high-precision control method and system for clamping force of parts. Background Technology
[0002] In automated forklift clamping operations, the stability and precision of the part clamping process are crucial. Current technologies generally rely on operator experience or static presets based on finite rules to control the clamping force. The basic approach involves pre-setting fixed clamping positions and theoretical clamping force values. However, such methods face significant limitations in practical applications: due to a lack of in-depth analysis and learning capabilities of historical clamping operation data, they cannot perceive the dynamic changes in optimal clamping strategies under different actual working conditions, such as different handling paths and different workpieces. This leads to rigid selection of clamping schemes, resulting in either insufficient clamping force causing part loosening during movement or excessive clamping force causing surface damage or deformation of the part. Consequently, control precision and efficiency are low, failing to meet the high precision and adaptability requirements of automated forklift clamping. Summary of the Invention
[0003] This application provides a high-precision control method and system for clamping force of parts, which is used to solve the technical problem that high-precision automated control of forklift clamping cannot be achieved in the prior art.
[0004] In view of the above problems, this application provides a high-precision control method and system for clamping force of parts.
[0005] In a first aspect, this application provides a high-precision control method for clamping force of a part, the method comprising:
[0006] Using part model as a constraint, frequent pattern analysis is performed on clamping patterns to obtain pattern support, wherein the clamping pattern represents the clamping type;
[0007] When the pattern support is greater than or equal to the first support threshold, based on the surface space coordinates of the part, and with the part model and the clamping pattern, a frequent pattern analysis is performed on the clamping scheme to obtain the selected clamping position and selected clamping force when the scheme support is greater than or equal to the second support threshold.
[0008] The clamping assembly is controlled based on the selected clamping position and the selected clamping force.
[0009] Secondly, this application provides a high-precision control system for part clamping force, comprising:
[0010] The clamping pattern analysis module is used to perform frequent pattern analysis on clamping patterns with part model as a constraint to obtain pattern support, wherein the clamping pattern represents the clamping type.
[0011] The clamping parameter selection module is used to perform frequent pattern analysis on the clamping scheme based on the spatial coordinates of the part surface, the part model and the clamping mode, when the pattern support is greater than or equal to the first support threshold, to obtain the selected clamping position and selected clamping force when the scheme support is greater than or equal to the second support threshold.
[0012] The clamping component control module is used to control the clamping component based on the selected clamping position and the selected clamping force.
[0013] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0014] This application proposes a high-precision control method and system for part clamping force. By mining and analyzing historical clamping operation data based on part type and movement trajectory constraints, the accuracy and adaptability of clamping parameter decisions are significantly improved. Compared with traditional methods, the technical solution provided in this application significantly overcomes the rigidity problem of forklift clamping control caused by reliance on manual experience or static parameters, achieving adaptive optimization and high-precision control of forklift clamping operations under diverse working conditions. Attached Figure Description
[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 This is a flowchart illustrating a high-precision control method for clamping force of a part provided in an embodiment of this application.
[0017] Figure 2 This is a schematic diagram of a high-precision control system for clamping force of a part provided in an embodiment of this application.
[0018] The components represented by each number in the attached diagram are explained below:
[0019] Clamping mode analysis module 100, clamping parameter selection module 200, clamping component control module 300. Detailed Implementation
[0020] This application provides a high-precision control method and system for clamping force of parts, which addresses the technical problem that high-precision automated control of forklift clamping cannot be achieved in the prior art.
[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.
[0022] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.
[0023] Example 1, as Figure 1 As shown, this application provides a high-precision control method for the clamping force of a part, wherein the method includes:
[0024] S10: Using part model as a constraint, perform frequent pattern analysis on clamping patterns to obtain pattern support, wherein the clamping pattern represents the clamping type.
[0025] In the process of part clamping control, the primary challenge is how to quickly and accurately identify the optimal clamping mode for a specific part model from historical operational data. Traditional methods rely on manual experience or fixed rules, lacking quantitative analysis capabilities, resulting in high arbitrariness and poor adaptability in clamping mode selection. Especially in multi-product clamping scenarios, different part models may correspond to multiple potential clamping modes, but existing technologies cannot accurately identify which clamping modes have high reliability, leading to low efficiency in clamping scheme decision-making and even the risk of clamping failure or part damage due to improper mode selection.
[0026] Step S10 in the method provided in this application embodiment includes:
[0027] Obtain the spatial coordinates of the movable area, combine the starting point and ending point of the movement to plan the movement path, and obtain the target movement trajectory coordinate sequence;
[0028] Using the target movement trajectory coordinate sequence as the movement trajectory constraint, retrieve the clamping movement log of the part model;
[0029] The process of retrieving the clamping movement log of the part model, using the target movement trajectory coordinate sequence as the movement trajectory constraint, includes:
[0030] Obtain the clamping movement log to be analyzed, wherein the clamping movement log to be analyzed has a label that identifies the coordinate sequence of the movement trajectory to be analyzed;
[0031] The movement type discriminator processes the coordinate sequence of the movement trajectory to be analyzed to obtain the movement type sequence and the movement distance sequence to be analyzed.
[0032] The target movement trajectory coordinate sequence is processed by a movement type discriminator to obtain a target movement type sequence and a target movement distance sequence.
[0033] The construction steps of the movement type discriminator include:
[0034] Extract the coordinates of the i-th and (i+1)-th sequences from the movement trajectory coordinate record sequence;
[0035] The movement type and movement distance are identified for the i-th and (i+1)-th coordinates to obtain labels for the movement type and the movement distance.
[0036] Until the entire sequence of movement trajectory coordinate records has been traversed, output the label identifying the initial movement type sequence and the label identifying the initial movement distance sequence;
[0037] Specifically, the movement type and movement distance are identified for the i-th and (i+1)-th coordinates to obtain labels indicating the movement type and movement distance, including:
[0038] Based on the identification rules, the movement type and movement distance of the i-th and (i+1)-th coordinates are identified, resulting in labels indicating the movement type and movement distance. The identification rules include rule 1 and rule 2, which can be triggered simultaneously.
[0039] Rule 1: When the first ordinate of the (i+1)th index is greater than the second ordinate of the i-th index, and the ordinate distance between the first ordinate and the second ordinate is greater than or equal to a distance threshold, generate a label indicating the type of lifting movement and a label indicating the lifting distance as the ordinate distance for the distance from the i-th index to the (i+1)th index.
[0040] When the first ordinate of the (i+1)th index coordinate is less than the second ordinate of the i-th index coordinate, and the ordinate distance between the first ordinate and the second ordinate is greater than or equal to a distance threshold, a label indicating the descent movement type and a label indicating the descent distance as the ordinate distance are generated for the distance from the i-th index coordinate to the (i+1)th index coordinate.
[0041] Rule 2: When the horizontal distance between the i-th coordinate and the (i+1)-th coordinate is greater than or equal to the distance threshold, generate a label indicating the horizontal movement type and a label indicating the horizontal movement distance as horizontal distance for the distance from the i-th coordinate to the (i+1)-th coordinate.
[0042] Based on the labels identifying the initial movement type sequence and the labels identifying the initial movement distance sequence, delete the initial movement type and initial movement distance whose sum of movement distances is less than or equal to the movement distance threshold, and obtain the labels identifying the movement type sequence and the labels identifying the movement distance sequence;
[0043] Using the labels of the movement type sequence and the labels of the movement distance sequence as supervision, and the movement trajectory coordinate record sequence as input, the movement type discriminator is trained through machine learning;
[0044] When the first edit distance between the movement type sequence to be analyzed and the target movement type sequence is less than or equal to the first edit distance threshold, and the second edit distance between the movement distance sequence to be analyzed and the movement distance sequence to be analyzed is less than or equal to the second edit distance threshold, the movement log to be analyzed is considered to satisfy the movement trajectory constraint.
[0045] Add the clamping movement log to be analyzed to the clamping movement log;
[0046] The percentage of log entries in the clamping movement log that use the clamping mode to move the part model is calculated and set as the mode support rate.
[0047] In this embodiment, using the part model as a constraint, frequent pattern analysis is performed by traversing the first clamping mode up to the Nth clamping mode when the same part model is clamped, to obtain the support of the first mode up to the Nth mode. The clamping modes include various clamping modes, such as vacuum adsorption and hydraulic clamping.
[0048] Using part model number as a constraint, frequent pattern analysis is performed on clamping patterns to obtain pattern support.
[0049] Specifically, the spatial coordinates of the movable area are obtained. Combining the starting and ending points of the movement, a movement path is planned to obtain the target movement trajectory coordinate sequence. The movable area refers to the spatial coordinates of the area gripped by the forklift. The starting point is the coordinate point where the forklift begins gripping the part, and the ending point is the coordinate point where the forklift lowers the part. The spatial coordinates are obtained by constructing a Cartesian coordinate system for the area, where the origin can be the starting point, and the X and Y axes can be set in the due east and due north directions, respectively, with a unit length of 10 centimeters. Combining the starting and ending points, a fast random tree algorithm is used to plan the movement path and obtain the target movement trajectory coordinate sequence. This target movement trajectory sequence includes multiple coordinates, with each coordinate spaced 10 seconds apart.
[0050] Using the target movement trajectory coordinate sequence as the movement trajectory constraint, retrieve the clamping movement logs of the part model to obtain clamping movement logs with the same target movement trajectory coordinate sequence and the same clamping part model.
[0051] Obtain the clamp movement log to be analyzed, wherein the clamp movement log to be analyzed has a label that identifies the coordinate sequence of the movement trajectory to be analyzed.
[0052] The movement type discriminator processes the coordinate sequence of the movement trajectory to be analyzed, and obtains the movement type sequence and the movement distance sequence to be analyzed.
[0053] The construction of the movement type discriminator includes: extracting the i-th and i+1-th index coordinates from the movement trajectory coordinate record sequence; wherein the index coordinates are obtained by marking the coordinates during the movement process according to the time sequence.
[0054] The i-th and (i+1)-th coordinates are labeled with movement type and movement distance, respectively, to obtain labels for the movement type and movement distance. Movement type refers to the direction of movement, such as descending or ascending, while movement distance refers to the distance between the i-th and (i+1)-th coordinates. Specifically, based on labeling rules, the i-th and (i+1)-th coordinates are labeled with movement type and movement distance, respectively, to obtain labels for the movement type and movement distance. The labeling rules include Rule 1 and Rule 2, which can be triggered simultaneously.
[0055] Rule 1: When the first ordinate of the (i+1)th index is greater than the second ordinate of the i-th index, and the ordinate distance between the first and second ordinates is greater than or equal to the distance threshold, generate a label indicating the type of lift movement and a label indicating the lift distance as the ordinate distance for the distance from the i-th index to the (i+1)-th index.
[0056] When the first ordinate of the (i+1)th index is less than the second ordinate of the i-th index, and the ordinate distance between the first and second ordinates is greater than or equal to the distance threshold, generate a label indicating the descent movement type and a label indicating the descent distance as the ordinate distance for the distance from the i-th index to the (i+1)-th index.
[0057] Rule 2: When the horizontal distance between the i-th coordinate and the (i+1)-th coordinate is greater than or equal to the distance threshold, generate a label indicating the horizontal movement type and a label indicating the horizontal movement distance as horizontal distance for the distance from the i-th coordinate to the (i+1)-th coordinate.
[0058] Until the entire sequence of movement trajectory coordinate records has been traversed, output the label identifying the initial movement type sequence and the label identifying the initial movement distance sequence.
[0059] Based on the labels identifying the initial movement type sequence and the initial movement distance sequence, initial movement types and initial movement distances whose sum of movement distances is less than or equal to a movement distance threshold are deleted, thus obtaining labels identifying the movement type sequence and sequences identifying the movement distance. For example, the movement distance threshold can be set to 10 centimeters. Since initial movement distances less than the threshold are short and have little impact on the overall movement status, and too many initial movement types and initial movement distances would lead to data redundancy and consume a large amount of computational resources, initial movement types and initial movement distances whose sum of movement distances is less than or equal to the movement distance threshold are deleted.
[0060] Using labels identifying movement type sequences and movement distance sequences as supervision, and movement trajectory coordinate record sequences as input, a movement type discriminator is trained through machine learning. For example, a movement type discriminator is constructed based on a random forest, with an ensemble model of 100 trees, a maximum depth of 5, and a minimum sample size of 2. For movement type, Gini impurity is used as the criterion for splitting and optimization; for movement distance, mean squared error is used. The constructed movement type discriminator is trained, taking movement trajectory coordinate records as input and outputting the discriminated movement type and movement distance, as well as the difference between the judgment and the labeled movement sequence. Using a chain rule, starting from the output layer, the contribution of each parameter to the final loss value is calculated layer by layer. The direction of the gradient indicates the direction in which the parameters should be adjusted; for example, if the movement type is incorrectly judged, the parameter should be reduced. The magnitude of the gradient indicates the magnitude of the parameter adjustment; for example, if the movement distance deviation is large, the parameter adjustment magnitude should be increased. The model is adjusted and optimized based on the gradient until convergence. That is, when the accuracy of the input movement trajectory coordinate record and the output movement type and movement distance is above 90%, the movement type discriminator training is complete.
[0061] The movement type discriminator processes the coordinate sequence of the movement trajectory to be analyzed, and obtains the movement type sequence and the movement distance sequence to be analyzed.
[0062] The target movement trajectory coordinate sequence is processed by the movement type discriminator to obtain the target movement type sequence and the target movement distance sequence.
[0063] The motion log to be analyzed is considered to satisfy the motion trajectory constraint when the first edit distance between the motion type sequence to be analyzed and the target motion type sequence is less than or equal to the first edit distance threshold, and the second edit distance between the motion distance sequences to be analyzed and the target motion type sequence is less than or equal to the second edit distance threshold. The edit distance threshold is the minimum number of operations required to transform one sequence into another. Operations on the sequences include insertion, deletion, and character replacement. A smaller edit distance indicates greater similarity between the two sequences, meaning their corresponding motion trajectories are also more similar. For example, the edit distance threshold can be set to 10.
[0064] Add the clamping movement logs that meet the movement trajectory constraints to the clamping movement log.
[0065] The percentage of log entries using clamping modes to move parts is recorded in the clamping and movement logs, and this percentage is defined as the mode support. For example, in the clamping and movement logs, if vacuum adsorption is used 10 times and hydraulic clamping is used 40 times when clamping the same model of cargo box, with no other clamping modes, then the mode support for vacuum adsorption is 10 ÷ (10 + 40) = 0.2, and the mode support for hydraulic clamping is 40 ÷ (10 + 40) = 0.8.
[0066] By using part model as a constraint, frequent pattern analysis of historical clamping patterns is performed, and pattern support is calculated. This enables intelligent screening and evaluation of clamping types, allowing for the automatic identification of repeatedly verified high-frequency clamping patterns for specific part models. Each pattern is assigned a support value, providing a scientific basis for subsequent clamping parameter decisions and significantly improving the accuracy and efficiency of clamping pattern selection.
[0067] S20: When the pattern support is greater than or equal to the first support threshold, based on the surface space coordinates of the part, and with the part model and the clamping pattern, perform frequent pattern analysis on the clamping scheme to obtain the selected clamping position and selected clamping force when the scheme support is greater than or equal to the second support threshold.
[0068] After determining the clamping mode, the key is to further determine specific clamping parameters such as position and clamping force. Traditional methods usually use fixed parameters, without considering the correlation between the characteristics of the part in actual clamping operations and historical successful cases, resulting in parameter settings that are out of touch with actual needs. For example, the same clamping mode may require different clamping force distributions under different surface space coordinates. If it is not possible to extract verified and efficient parameter combinations from historical data, it can easily lead to problems such as uneven clamping force distribution, decreased positioning accuracy, or part deformation.
[0069] Step S20 in the method provided in this application embodiment includes:
[0070] Retrieve the clamping scheme to be analyzed for clamping the part model in the clamping mode, wherein the clamping scheme to be analyzed includes clamping record space coordinates, part surface record space coordinates, and record clamping force;
[0071] Align the spatial coordinates of the part surface with the spatial coordinates of the part surface recording, locate the spatial coordinates of the clamping recording, and obtain the clamping recording position;
[0072] The clamping record position and the record clamping force are stored as a first clamping scheme and added to the clamping scheme set;
[0073] When the number of clamping schemes in the clamping scheme set is greater than or equal to the number threshold of clamping schemes, based on the clamping record position and the record clamping force, the clamping scheme set is traversed, and the number of clamping schemes with clamping position Euclidean distance less than or equal to the position Euclidean distance threshold and clamping force deviation less than or equal to the clamping force deviation threshold is counted and set as the first clamping scheme support.
[0074] When the support of the first clamping scheme is greater than or equal to the second support threshold, the clamping record position and the record clamping force are set as the selected clamping position and the selected clamping force.
[0075] In this embodiment of the application, a clamping scheme to be analyzed is retrieved that clamps a part model in a clamping mode. The clamping scheme to be analyzed includes clamping record space coordinates, part surface record space coordinates, and record clamping force.
[0076] Align the spatial coordinates of the part surface with the spatial coordinates of the part surface record, and locate the spatial coordinates of the clamping record to obtain the clamping record position. For example, a coordinate alignment algorithm such as the Iterative Closest Point (ICP) algorithm is invoked. Using the part surface spatial coordinates as the target, the spatial coordinates of the part surface record in the historical records are rotated and translated to calculate an optimal spatial transformation matrix. Then, this same transformation matrix is applied to the corresponding clamping record spatial coordinates, thereby uniformly locating these historical clamping point coordinates to the current part's coordinate system. The located coordinates are the clamping record position, indicating where the clamp should be positioned if a historical operation occurred on the current part.
[0077] Record the clamping position and clamping force, store them as the first clamping scheme, and add them to the clamping scheme set.
[0078] When the number of clamping schemes in the clamping scheme set is greater than or equal to a threshold, based on the clamping record position and the recorded clamping force, the clamping scheme set is traversed, and the number of clamping schemes whose clamping position Euclidean distance is less than or equal to a position Euclidean distance threshold and whose clamping force deviation is less than or equal to a clamping force deviation threshold is counted. This number is set as the first clamping scheme support. Here, the clamping position Euclidean distance refers to the Euclidean distance between the target clamping position and the clamping record position. When the Euclidean distance is less than or equal to the position Euclidean distance threshold, it indicates that the target clamping position and the schemes in the clamping scheme set have high similarity, and the schemes in the clamping scheme set have strong reference value. The clamping force deviation threshold refers to the threshold value of the deviation between the target clamping force and the recorded clamping force. A deviation value less than or equal to the clamping force deviation threshold indicates that the clamping force and the schemes in the clamping scheme set have high similarity. The number of clamping schemes whose clamping position Euclidean distance is less than or equal to the position Euclidean distance threshold and whose clamping force deviation is less than or equal to the clamping force deviation threshold is counted and set as the first clamping scheme support.
[0079] When the support of the first clamping scheme is greater than or equal to the second support threshold, the clamping recording position and clamping force of the first clamping scheme are set as the selected clamping position and selected clamping force. The second support threshold is a threshold characterizing the degree to which the scheme supports the current clamping requirements. The higher the support, the more suitable the first clamping scheme is for the current clamping scenario, indicating that it has received high support from historical data and is a reliable scheme that has been repeatedly verified. Preferably, there may be multiple first clamping schemes. The set of clamping recording positions and clamping forces with the lowest energy consumption among the first clamping schemes is selected as the selected clamping position and selected clamping force.
[0080] By combining the spatial coordinates of the part surface, the part model, and the clamping mode, a secondary frequent pattern analysis is performed on historical clamping schemes. Based on the scheme support threshold, the optimal clamping position and clamping force are selected, realizing the refinement and adaptive recommendation of clamping parameters. This ensures that the adopted schemes not only meet the clamping mode requirements but also have high statistical significance and practical reliability. Thus, the stability and accuracy of the clamping process can still be maintained in complex and ever-changing application scenarios, effectively avoiding performance defects caused by unreasonable parameter settings.
[0081] S30: Control the clamping assembly based on the selected clamping position and the selected clamping force.
[0082] In this embodiment, the clamping component is controlled based on the selected clamping position and the selected clamping force, achieving a seamless connection from parameter decision-making to physical execution. This ensures that the clamping component can strictly follow the historically proven high-reliability scheme to perform operations, which not only improves the accuracy and repeatability of the control process, but also enhances the adaptability under different working conditions, ultimately achieving the goal of high-precision and high-stability clamping control.
[0083] Example 2, as Figure 2 As shown, based on the same inventive concept as the high-precision control method for part clamping force provided in Embodiment 1, this embodiment of the invention also provides a high-precision control system for part clamping force, including:
[0084] The clamping pattern analysis module 100 is used to perform frequent pattern analysis on clamping patterns with part model as a constraint to obtain pattern support, wherein the clamping pattern represents the clamping type.
[0085] The clamping parameter selection module 200 is used to perform frequent pattern analysis on the clamping scheme based on the spatial coordinates of the part surface, the part model, and the clamping pattern when the pattern support is greater than or equal to a first support threshold, to obtain the selected clamping position and selected clamping force when the scheme support is greater than or equal to a second support threshold.
[0086] The clamping component control module 300 is used to control the clamping component based on the selected clamping position and the selected clamping force.
[0087] In one embodiment, the clamping pattern analysis module 100 is further configured to:
[0088] Obtain the spatial coordinates of the movable area, combine the starting point and ending point of the movement to plan the movement path, and obtain the target movement trajectory coordinate sequence;
[0089] Using the target movement trajectory coordinate sequence as the movement trajectory constraint, retrieve the clamping movement log of the part model;
[0090] The process of retrieving the clamping movement log of the part model, using the target movement trajectory coordinate sequence as the movement trajectory constraint, includes:
[0091] Obtain the clamping movement log to be analyzed, wherein the clamping movement log to be analyzed has a label that identifies the coordinate sequence of the movement trajectory to be analyzed;
[0092] The movement type discriminator processes the coordinate sequence of the movement trajectory to be analyzed to obtain the movement type sequence and the movement distance sequence to be analyzed.
[0093] The target movement trajectory coordinate sequence is processed by a movement type discriminator to obtain a target movement type sequence and a target movement distance sequence.
[0094] The construction steps of the movement type discriminator include:
[0095] Extract the coordinates of the i-th and (i+1)-th sequences from the movement trajectory coordinate record sequence;
[0096] The movement type and movement distance are identified for the i-th and (i+1)-th coordinates to obtain labels for the movement type and the movement distance.
[0097] Until the entire sequence of movement trajectory coordinate records has been traversed, output the label identifying the initial movement type sequence and the label identifying the initial movement distance sequence;
[0098] Specifically, the movement type and movement distance are identified for the i-th and (i+1)-th coordinates to obtain labels indicating the movement type and movement distance, including:
[0099] Based on the identification rules, the movement type and movement distance of the i-th and (i+1)-th coordinates are identified, resulting in labels indicating the movement type and movement distance. The identification rules include rule 1 and rule 2, which can be triggered simultaneously.
[0100] Rule 1: When the first ordinate of the (i+1)th index is greater than the second ordinate of the i-th index, and the ordinate distance between the first ordinate and the second ordinate is greater than or equal to a distance threshold, generate a label indicating the type of lifting movement and a label indicating the lifting distance as the ordinate distance for the distance from the i-th index to the (i+1)th index.
[0101] When the first ordinate of the (i+1)th index coordinate is less than the second ordinate of the i-th index coordinate, and the ordinate distance between the first ordinate and the second ordinate is greater than or equal to a distance threshold, a label indicating the descent movement type and a label indicating the descent distance as the ordinate distance are generated for the distance from the i-th index coordinate to the (i+1)th index coordinate.
[0102] Rule 2: When the horizontal distance between the i-th coordinate and the (i+1)-th coordinate is greater than or equal to the distance threshold, generate a label indicating the horizontal movement type and a label indicating the horizontal movement distance as horizontal distance for the distance from the i-th coordinate to the (i+1)-th coordinate.
[0103] Based on the labels identifying the initial movement type sequence and the labels identifying the initial movement distance sequence, delete the initial movement type and initial movement distance whose sum of movement distances is less than or equal to the movement distance threshold, and obtain the labels identifying the movement type sequence and the labels identifying the movement distance sequence;
[0104] Using the labels of the movement type sequence and the labels of the movement distance sequence as supervision, and the movement trajectory coordinate record sequence as input, the movement type discriminator is trained through machine learning;
[0105] When the first edit distance between the movement type sequence to be analyzed and the target movement type sequence is less than or equal to the first edit distance threshold, and the second edit distance between the movement distance sequence to be analyzed and the movement distance sequence to be analyzed is less than or equal to the second edit distance threshold, the movement log to be analyzed is considered to satisfy the movement trajectory constraint.
[0106] Add the clamping movement log to be analyzed to the clamping movement log;
[0107] The percentage of log entries in the clamping movement log that use the clamping mode to move the part model is calculated and set as the mode support rate.
[0108] In one embodiment, the clamping parameter selection module 200 is further configured to:
[0109] Retrieve the clamping scheme to be analyzed for clamping the part model in the clamping mode, wherein the clamping scheme to be analyzed includes clamping record space coordinates, part surface record space coordinates, and record clamping force;
[0110] Align the spatial coordinates of the part surface with the spatial coordinates of the part surface recording, locate the spatial coordinates of the clamping recording, and obtain the clamping recording position;
[0111] The clamping record position and the record clamping force are stored as a first clamping scheme and added to the clamping scheme set;
[0112] When the number of clamping schemes in the clamping scheme set is greater than or equal to the number threshold of clamping schemes, based on the clamping record position and the record clamping force, the clamping scheme set is traversed, and the number of clamping schemes with clamping position Euclidean distance less than or equal to the position Euclidean distance threshold and clamping force deviation less than or equal to the clamping force deviation threshold is counted and set as the first clamping scheme support.
[0113] When the support of the first clamping scheme is greater than or equal to the second support threshold, the clamping record position and the record clamping force are set as the selected clamping position and the selected clamping force.
[0114] In summary, the embodiments of this application have at least the following technical effects:
[0115] This application proposes a high-precision control method and system for part clamping force. By mining and analyzing historical clamping operation data based on part type and movement trajectory constraints, the accuracy and adaptability of clamping parameter decisions are significantly improved. Specifically, the method provided in this application can identify repeatedly verified efficient clamping patterns and optimal parameter combinations from massive historical operation logs, and dynamically recommend the most reliable clamping position and clamping force according to actual working conditions. By introducing a multi-level support threshold evaluation mechanism, the adopted clamping scheme is ensured to have high statistical significance and operational reliability, effectively avoiding problems such as part loosening or surface damage caused by improper clamping force. At the same time, through refined matching and recognition of movement trajectory sequences, the correlation between clamping decisions and specific handling processes is further enhanced, making clamping control more in line with actual application scenarios. Compared with traditional methods, the technical solution provided in this application significantly overcomes the rigidity of forklift clamping control caused by reliance on manual experience or static parameters, achieving adaptive optimization and high-precision control of forklift clamping operations under diverse working conditions.
[0116] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.
[0117] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.
[0118] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.
Claims
1. A high-precision control method for clamping force of a part, characterized in that, Applied to clamping components, including: Using part model as a constraint, frequent pattern analysis is performed on clamping patterns to obtain pattern support, wherein the clamping pattern represents the clamping type; When the pattern support is greater than or equal to the first support threshold, based on the surface space coordinates of the part, and with the part model and the clamping pattern, a frequent pattern analysis is performed on the clamping scheme to obtain the selected clamping position and selected clamping force when the scheme support is greater than or equal to the second support threshold. The clamping assembly is controlled based on the selected clamping position and the selected clamping force. Among them, frequent pattern analysis is performed on the clamping patterns using part model as a constraint to obtain pattern support, including: Obtain the spatial coordinates of the movable area, combine the starting point and ending point of the movement to plan the movement path, and obtain the target movement trajectory coordinate sequence; Using the target movement trajectory coordinate sequence as the movement trajectory constraint, retrieve the clamping movement log of the part model; The percentage of log entries in the clamping movement log that use the clamping mode to move the part model is defined as the mode support rate. The process of retrieving the clamping movement log of the part model, using the target movement trajectory coordinate sequence as the movement trajectory constraint, includes: Obtain the clamping movement log to be analyzed, wherein the clamping movement log to be analyzed has a label that identifies the coordinate sequence of the movement trajectory to be analyzed; The movement type discriminator processes the coordinate sequence of the movement trajectory to be analyzed to obtain the movement type sequence and the movement distance sequence to be analyzed. The target movement trajectory coordinate sequence is processed by a movement type discriminator to obtain a target movement type sequence and a target movement distance sequence. When the first edit distance between the movement type sequence to be analyzed and the target movement type sequence is less than or equal to the first edit distance threshold, and the second edit distance between the movement distance sequence to be analyzed and the movement distance sequence to be analyzed is less than or equal to the second edit distance threshold, the movement log to be analyzed is considered to satisfy the movement trajectory constraint. Add the clamp movement log to be analyzed into the clamp movement log.
2. The method as described in claim 1, characterized in that, The clamping patterns include a first clamping pattern up to the Nth clamping pattern. Using the part number as a constraint, frequent pattern analysis is performed on the clamping patterns to obtain pattern support, including: Using the part model as a constraint, frequent pattern analysis is performed by traversing the first clamping pattern up to the Nth clamping pattern to obtain the support of the first pattern up to the support of the Nth pattern.
3. The method as described in claim 1, characterized in that, The steps for constructing the movement type discriminator include: Extract the coordinates of the i-th and (i+1)-th sequences from the movement trajectory coordinate record sequence; The movement type and movement distance are identified for the i-th and (i+1)-th coordinates to obtain labels for the movement type and the movement distance. Until the entire sequence of movement trajectory coordinate records has been traversed, output the label identifying the initial movement type sequence and the label identifying the initial movement distance sequence; Based on the labels identifying the initial movement type sequence and the labels identifying the initial movement distance sequence, delete the initial movement type and initial movement distance whose sum of movement distances is less than or equal to the movement distance threshold, and obtain the labels identifying the movement type sequence and the labels identifying the movement distance sequence; Using the labels of the movement type sequence and the labels of the movement distance sequence as supervision, and the movement trajectory coordinate record sequence as input, the movement type discriminator is trained through machine learning.
4. The method as described in claim 3, characterized in that, The movement type and movement distance are identified for the i-th and (i+1)-th coordinates, respectively, to obtain labels indicating the movement type and the movement distance, including: Based on the identification rules, the movement type and movement distance of the i-th and (i+1)-th coordinates are identified, resulting in labels indicating the movement type and movement distance. The identification rules include rule 1 and rule 2, which can be triggered simultaneously. Rule 1: When the first ordinate of the (i+1)th index is greater than the second ordinate of the i-th index, and the ordinate distance between the first ordinate and the second ordinate is greater than or equal to a distance threshold, generate a label indicating the type of lifting movement and a label indicating the lifting distance as the ordinate distance for the distance from the i-th index to the (i+1)th index. When the first ordinate of the (i+1)th index coordinate is less than the second ordinate of the i-th index coordinate, and the ordinate distance between the first ordinate and the second ordinate is greater than or equal to a distance threshold, a label indicating the descent movement type and a label indicating the descent distance as the ordinate distance are generated for the distance from the i-th index coordinate to the (i+1)th index coordinate. Rule 2: When the horizontal distance between the i-th coordinate and the (i+1)-th coordinate is greater than or equal to the distance threshold, generate a label indicating the horizontal movement type and a label indicating the horizontal movement distance as horizontal distance for the distance from the i-th coordinate to the (i+1)-th coordinate.
5. The method as described in claim 1, characterized in that, When the pattern support is greater than or equal to a first support threshold, based on the surface space coordinates of the part, and using the part model and the clamping pattern, frequent pattern analysis is performed on the clamping scheme to obtain the selected clamping position and selected clamping force where the scheme support is greater than or equal to a second support threshold, including: Retrieve the clamping scheme to be analyzed for clamping the part model in the clamping mode, wherein the clamping scheme to be analyzed includes clamping record space coordinates, part surface record space coordinates, and record clamping force; Align the spatial coordinates of the part surface with the spatial coordinates of the part surface recording, locate the spatial coordinates of the clamping recording, and obtain the clamping recording position; The clamping record position and the record clamping force are stored as a first clamping scheme and added to the clamping scheme set; When the number of clamping schemes in the clamping scheme set is greater than or equal to the number threshold of clamping schemes, based on the clamping record position and the record clamping force, the clamping scheme set is traversed, and the number of clamping schemes with clamping position Euclidean distance less than or equal to the position Euclidean distance threshold and clamping force deviation less than or equal to the clamping force deviation threshold is counted and set as the first clamping scheme support. When the support of the first clamping scheme is greater than or equal to the second support threshold, the clamping record position and the record clamping force are set as the selected clamping position and the selected clamping force.
6. A high-precision control system for clamping force of parts, characterized in that, A system for implementing a high-precision control method for clamping force of a part according to any one of claims 1 to 5, the system comprising: The clamping pattern analysis module is used to perform frequent pattern analysis on clamping patterns with part model as a constraint to obtain pattern support, wherein the clamping pattern represents the clamping type. The clamping parameter selection module is used to perform frequent pattern analysis on the clamping scheme based on the spatial coordinates of the part surface, the part model and the clamping mode, when the pattern support is greater than or equal to the first support threshold, to obtain the selected clamping position and selected clamping force when the scheme support is greater than or equal to the second support threshold. The clamping component control module is used to control the clamping component based on the selected clamping position and the selected clamping force.
Citation Information
Patent Citations
Gripping force adjustment device and gripping force adjustment system
CN111687652A
Robot control method and device, robot and readable storage medium
CN113799134A