Intelligent gripper grasping force control method and device
Through three-dimensional modeling and point cloud reconstruction technology, combined with sensor data optimization of the grabbing force, the damage and grabbing failure caused by the fixed grabbing force of the intelligent gripper is solved, achieving higher operational accuracy and adaptability.
Patent Information
- Application Number
- CN202411462748.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-19
- Publication Date
- 2024-12-20
- Estimated Expiration
- 2044-10-19
AI Technical Summary
The existing intelligent gripper has a fixed design and a single function, and it is impossible to sense information such as the shape, size, weight, etc. of the captured object in real time, resulting in insufficient or excessive grabbing, which can easily lead to damage to the item or failed to grab.
Using three-dimensional modeling and point cloud reconstruction technology, object data is collected through three-dimensional vision sensors, grasping force data is obtained in combination with force sensors, object 3D model is generated, structural material characteristics and quality information is analyzed, initial grasping force strategy parameters are optimized, and the Real-time monitoring and adjustment achieve flexible control of grabbing force.
It improves the operation accuracy, reliability and adaptability of intelligent grippers, reduces the risk of item damage and crawling failure, and enhances the ability to crawl complex items.
Smart Images
Figure CN118952238B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field related to robot technology, and specifically to a method and device for controlling the gripping force of an intelligent gripper. Background Art
[0002] As an important part of robot technology, the development and progress of intelligent grippers marks a significant improvement in the robot's operating capabilities. With the continuous development of science and technology, robots are gradually penetrating into all areas of human life. As the "hand" of the robot, the performance and accuracy of the intelligent gripper directly affect the effectiveness of the robot's task execution. With the rapid development of industrial automation, the requirements for robot operating accuracy and flexibility are increasing. Especially when handling complex, delicate or fragile objects, the gripping force control of the gripper is particularly important. However, traditional intelligent grippers have a fixed design and a single function, and it is difficult to adapt to complex and changing working environments. They usually rely on a preset fixed force for gripping, and often cannot perceive the shape, size, weight and other information of the grasped object in real time, which can easily lead to damage to the object or grasping failure.
[0003] Therefore, in the current technologies related to the grasping force control of intelligent grippers, there are technical problems such as reliance on a preset fixed force for grasping and lack of flexibility, which leads to damage to the grasped objects or grasping failure. Summary of the invention
[0004] The present application provides a method and device for controlling the gripping force of an intelligent gripper, and adopts technical means such as three-dimensional modeling and point cloud reconstruction to solve the technical problems of the existing intelligent gripper gripping force control that relies on a preset fixed force for gripping, lacks flexibility, and causes damage to the grasped objects or grasping failure, thereby achieving the technical effect of improving the accuracy, reliability and adaptability of the intelligent gripper operation.
[0005] The present application provides a method for controlling the grasping force of an intelligent gripper, the method comprising: S1: obtaining a target intelligent gripper, the target intelligent gripper having a built-in sensor module, wherein the sensor module comprises a three-dimensional vision sensor and a force sensor, collecting and acquiring three-dimensional appearance point cloud data of a target object through the three-dimensional vision sensor, and acquiring grasping force data of the target object through the force sensor test; S2: reconstructing the three-dimensional appearance point cloud data by using three-dimensional modeling technology to generate a three-dimensional model of the target object, and obtaining structural material feature information of the target object based on the three-dimensional model of the target object, and performing mass conversion based on the grasping force data. Obtain target object quality information; S3: Obtain design feature information of the target intelligent gripper and the preset grasping and placement state of the target object, perform grasping strategy analysis based on the design feature information and the preset grasping and placement state, the target object structural material feature information and the target object quality information, and obtain initial grasping force strategy parameters; S4: The target intelligent gripper performs grasping monitoring on the target object based on the initial grasping force strategy parameters, obtains grasping state feedback information, optimizes and adjusts the initial grasping force strategy parameters based on the grasping state feedback information, and determines the grasping force optimization strategy parameters to perform object grasping control.
[0006] In the implementation of the present application, the three-dimensional model of the target object is generated in S2, and the following processing is also performed: S21: obtaining point cloud attribute description factors, wherein the point cloud attribute description factors include spatial position, reflection intensity, and color texture; S22: using the point cloud attribute description factors to respectively describe the attributes of the three-dimensional appearance point cloud data, and obtaining a point cloud attribute description feature set of the object; S23: performing key area selection based on the object point cloud attribute description feature set, and obtaining the reference object point cloud area feature information and the remaining object point cloud area feature information set; S24: sequentially associating and mapping the remaining object point cloud area information set and the reference object point cloud area information, to obtain an object point cloud space transformation matrix set; S25: utilizing the three-dimensional modeling technology to respectively align, fuse, and reconstruct the remaining object point cloud area feature information sets with the reference object point cloud area feature information based on the object point cloud space transformation matrix set, and generate the three-dimensional model of the target object.
[0007] In the implementation method of the present application, the initial grasping force strategy parameters are obtained in S3, and the following processing is performed: S31: grasping mechanism information is obtained according to the design characteristic information, and the grasping mechanism information includes a clamping method and an adjustable threshold of the gripper; S32: a reference point is set for the structural material characteristic information of the target object, a grasping coordinate space is constructed, and the preset grasping and placement state is mapped based on the grasping coordinate space to obtain the preset grasping and placement coordinate information; S33: grasping contact point analysis is performed on the grasping mechanism information and the preset grasping and placement coordinate information based on the structural material characteristic information of the target object to obtain the object grasping contact point information; S34: grasping force analysis is performed on the object grasping contact point information and the target object mass information to obtain the grasping contact force distribution threshold information, and the grasping contact force distribution threshold information is used as a constraint condition to perform grasping strategy optimization analysis to obtain the initial grasping force strategy parameters.
[0008] In the implementation method of the present application, the object grasping contact point information is obtained in the S33, and the following processing is performed: the three-dimensional model of the target object is divided into finite element meshes to obtain a finite element mesh model of the target object, and the finite element mesh model of the target object is assigned grasping attributes based on the structural material feature information of the target object to generate a finite element simulation model of the target object; based on the finite element simulation model of the target object, the grasping contact point simulation is performed on the grasping mechanism information and the preset grasping placement coordinate information to obtain a grasping contact point simulation result set; the grasping stability is evaluated on the grasping contact point simulation result set respectively, and the grasping contact point information of the object is obtained by comparison.
[0009] In the implementation method of the present application, the initial grasping force strategy parameters are obtained in S34, and the following processing is also performed: the grasping contact force distribution threshold information is used as a constraint condition, and a grasping simulation prediction is performed on the finite element simulation model of the target object to obtain a target object grasping test data set; the object grasping target is analyzed to obtain an object grasping effect indicator set, and the object grasping effect indicator set includes grasping stability, grasping accuracy and grasping energy consumption; the target object grasping test data set is evaluated by the object grasping effect indicator set to obtain a target object grasping effect data set; according to the target object grasping test data set and the target object grasping effect data set, a gripper grasping control space is constructed, and a grasping strategy optimization analysis is performed based on the gripper grasping control space to determine the initial grasping force strategy parameters.
[0010] In the implementation method of the present application, the grasping force optimization strategy parameters are determined in S4, and the following processing is also performed: S41: If the grasping state feedback information does not achieve the preset grasping effect, the grasping state feedback information is subjected to a grasping deviation analysis to obtain the object grasping deviation parameters; S42: The object grasping deviation parameters are subjected to an associated control parameter analysis to obtain the control parameter variation direction, and the parameter variation rules are set according to the control parameter variation direction; S43: Based on the parameter variation rules, the initial grasping force strategy parameters are subjected to a variation simulation optimization to determine the grasping force optimization strategy parameters.
[0011] In the implementation of the present application, the grasping force optimization strategy parameters are determined in S43, and the following processing is also performed: the parameter variation rule is specifically: ,in, To control the parameter variation value, is the original control parameter value, is a random number between 0 and 1. is a normal distribution function; based on the parameter variation rule, the initial grasping force strategy parameters are mutated and expanded to obtain a grasping force strategy variation parameter set; based on the object grasping effect indicator set, the grasping force strategy variation parameter set is subjected to parameter simulation optimization to determine the grasping force optimization strategy parameters.
[0012] The present application also provides a smart gripper gripping force control device, comprising:
[0013] A target intelligent gripper acquisition module, the target intelligent gripper acquisition module is used to acquire a target intelligent gripper, the target intelligent gripper has a built-in sensor module, wherein the sensor module includes a three-dimensional vision sensor and a force sensor, the three-dimensional appearance point cloud data of the target object is acquired by the three-dimensional vision sensor, and the gripping force data of the target object is acquired by the force sensor test;
[0014] A target object three-dimensional model generation module, wherein the target object three-dimensional model generation module uses three-dimensional modeling technology to perform point cloud reconstruction on the three-dimensional appearance point cloud data to generate a target object three-dimensional model, and obtains structural material feature information of the target object based on the target object three-dimensional model, and performs mass conversion based on the grasping force data to obtain target object mass information;
[0015] An initial grasping force strategy parameter acquisition module, the initial grasping force strategy parameter acquisition module is used to obtain the design characteristic information of the target intelligent gripper and the preset grasping and placement state of the target object, and perform grasping strategy analysis based on the design characteristic information and the preset grasping and placement state, the structural material feature information of the target object and the quality information of the target object to obtain the initial grasping force strategy parameters;
[0016] An object grasping control module is used for the target intelligent gripper to monitor the grasping of the target object based on the initial grasping force strategy parameters, obtain grasping state feedback information, optimize and adjust the initial grasping force strategy parameters based on the grasping state feedback information, and determine the grasping force optimization strategy parameters for object grasping control.
[0017] It is intended to obtain a target intelligent gripper through a method and device for controlling the grasping force of an intelligent gripper proposed in this application. The target intelligent gripper has a built-in sensor module, obtains three-dimensional appearance point cloud data of the target object, and obtains grasping force data of the target object; performs point cloud reconstruction on the three-dimensional appearance point cloud data, generates a three-dimensional model of the target object, and obtains structural material feature information and quality information of the target object; obtains design characteristic information of the target intelligent gripper and the preset grasping and placement state of the target object, and performs grasping strategy analysis; performs grasping monitoring on the target object, obtains grasping state feedback information, optimizes and adjusts initial grasping force strategy parameters, and determines grasping force optimization strategy parameters for object grasping control, thereby solving the technical problems of existing intelligent gripper grasping force control that relies on preset fixed force for grasping and lacks flexibility, resulting in damage to the grasped object or grasping failure, and achieves the technical effect of improving the accuracy, reliability and adaptability of intelligent gripper operation. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] In order to more clearly illustrate the technical solution of the embodiment of the present disclosure, the accompanying drawings of the embodiment of the present disclosure will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the device according to the embodiment of the present application. It should be understood that the previous or following operations are not necessarily performed precisely in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.
[0019] Figure 1 A schematic diagram of a flow chart of a method for controlling the gripping force of an intelligent gripper provided in an embodiment of the present application;
[0020] Figure 2 A schematic diagram of the structure of a smart gripper gripping force control device provided in an embodiment of the present application.
[0021] Explanation of the accompanying drawings: target intelligent gripper acquisition module 10, target object three-dimensional model generation module 20, initial grasping force strategy parameter acquisition module 30, object grasping control module 40. DETAILED DESCRIPTION
[0022] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.
[0023] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.
[0024] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.
[0025] The present application embodiment provides a method for controlling the gripping force of an intelligent gripper, such as Figure 1 As shown, the method includes:
[0026] Step S1: Obtain a target intelligent gripper, which has a built-in sensor module, wherein the sensor module includes a three-dimensional vision sensor and a force sensor. The three-dimensional appearance point cloud data of the target object is collected and acquired through the three-dimensional vision sensor, and the gripping force data of the target object is acquired through the force sensor test. The target intelligent gripper is a gripping device that integrates advanced sensor modules and is used to accurately and efficiently grip and process various target objects. The built-in sensor module includes a three-dimensional vision sensor and a force sensor. By integrating the data of the three-dimensional vision sensor and the force sensor, the target intelligent gripper can accurately identify and stably grip the target object. Specifically, the three-dimensional vision sensor is the "eye" of the target intelligent gripper, which has the ability to acquire three-dimensional data with high precision and high speed. When the gripper approaches the target object, the three-dimensional vision sensor will quickly collect the three-dimensional appearance point cloud data of the target object, and record in detail the shape, size, position and slight fluctuations of the surface of the target object, providing the gripper with a comprehensive visual information of the target object. The point cloud data of the target object's three-dimensional appearance is a series of discrete point sets, representing the shape and contour of the target object's surface. By analyzing and processing these point cloud data, the gripper can identify the key features of the object; the force sensor is the "tactile nerve" in the target intelligent gripper, responsible for real-time monitoring and testing of the force data generated during the grasping process. In the process of the gripper contacting the target object and applying grasping force, the force sensor will accurately measure the size and changes of force in all directions, which is very important for judging whether the grasping is stable and whether the grasping force needs to be adjusted. Through the feedback of the force sensor, the gripper can adjust the grasping strategy in real time to ensure that the target object will not be damaged during the grasping process and the stability of the grasping can be ensured.
[0027] Step S2: Reconstruct the three-dimensional appearance point cloud data using three-dimensional modeling technology to generate a three-dimensional model of the target object, and obtain structural material feature information of the target object based on the three-dimensional model of the target object. At the same time, perform mass conversion based on the grasping force data to obtain mass information of the target object. The three-dimensional appearance point cloud data is reconstructed using three-dimensional modeling technology (for example, point cloud registration, filtering, segmentation and surface reconstruction, etc.), that is, it is processed and analyzed to reconstruct a complete three-dimensional model of the target object, that is, the three-dimensional model of the target object. For example, the target object is a cone, and the bottom surface of the cone is used as the key reference area. The remaining point cloud data generates corresponding transformation matrices, such as translation and symmetry matrices, according to their positional relationship with the reference area. The registration is calculated and the point cloud areas are fused to construct a three-dimensional model. Specifically, the three-dimensional model of the target object contains the geometric shape information of the target object, as well as the surface details and features of the target object. Based on the three-dimensional model of the target object, the structural material feature information of the target object is further extracted. For example, by analyzing the geometric shape and surface texture of the model, the material type, hardness, elasticity and other properties of the object can be inferred, which can help the gripper choose the appropriate grasping method and strength to avoid damaging the object or causing grasping failure; by analyzing the acquired grasping force data and combining it with the known gravitational acceleration, the mass of the target object can be estimated, which can help the intelligent gripper better understand the weight characteristics of the object.
[0028] In a possible implementation, step S2 further includes step S21, obtaining point cloud attribute description factors, wherein the point cloud attribute description factors include spatial position, reflection intensity, and color texture. Spatial position is the most basic and key attribute of point cloud data. Each point cloud data point contains its coordinate information (x, y, z) in three-dimensional space, accurately calibrating the specific position of each point in space. Through the spatial position information, the shape, size, and position relationship of the object can be accurately described; the reflection intensity is usually related to the principle of laser measurement, which represents the energy measurement of the laser beam reflected from the target surface. It is not only related to the surface material, roughness, and incident angle direction of the target, but also affected by the emission energy of the instrument, laser wavelength, and other factors. By analyzing the reflection intensity information, the reflection characteristics of the target object surface can be understood, and the material and surface state of the target can be further inferred; color texture is usually related to photogrammetry. Color information (such as RGB) describes the color characteristics of the point, while texture reflects the details and structure of the object surface. By obtaining the color texture information of the point cloud, the appearance and details of the object can be presented more realistically, and the sense of reality and detail of the three-dimensional model can be enhanced. It also includes step S22, using the point cloud attribute description factors to respectively describe the attributes of the three-dimensional appearance point cloud data, and obtain the object point cloud attribute description feature set. For the spatial position attribute, the three-dimensional coordinate information of each point in the point cloud data is extracted, and the overall shape, size and relative position relationship between the parts of the object are analyzed; for the reflection intensity attribute, the laser reflection intensity information of each point in the point cloud data is extracted, and the surface characteristics of the object, such as material type, surface roughness, etc., are analyzed and inferred; for the color and texture attributes, the color information and possible texture information of each point in the point cloud data are extracted, and the color distribution, texture characteristics, etc. of the object are analyzed and obtained, and they are integrated together to form the object point cloud attribute description feature set. It also includes step S23, based on the object point cloud attribute description feature set, key area selection is performed to obtain the reference object point cloud area feature information and the remaining object point cloud area feature information set. The object point cloud data is analyzed and processed to identify the key areas in the object point cloud, and the feature information of these areas and the remaining areas is extracted. Specifically, through comprehensive consideration of the attribute characteristics, it is possible to determine which areas are the key areas of the object, such as the edges of the object, raised parts, special texture areas, etc., and perform refined feature extraction on the key areas, including the shape, size, surface details, etc. of the key areas. In addition to the key areas, it is also necessary to pay attention to the feature information of the remaining object point cloud areas. Finally, the reference object point cloud area feature information and the remaining object point cloud area feature information set can be obtained. The reference object point cloud area feature information describes the key area characteristics of the object, and the remaining object point cloud area feature information set provides a supplementary description of the overall characteristics of the object.The method further includes step S24, wherein the remaining object point cloud region information set and the reference object point cloud region information are sequentially associated and mapped to obtain a set of object point cloud space transformation matrices. The object point cloud data is spatially aligned and transformed to establish a spatial correspondence between the remaining object point cloud regions and the reference object point cloud regions, thereby obtaining a series of space transformation matrices. The method further includes step S25, wherein the remaining object point cloud region feature information set is registered, fused and reconstructed with the reference object point cloud region feature information based on the object point cloud space transformation matrix set using three-dimensional modeling technology to generate a three-dimensional model of the target object. Accurately align and fuse each area in the object point cloud according to their spatial relationship to generate a complete and accurate 3D model of the target object. Specifically, select a point cloud area of the remaining objects, and transform the area to the same coordinate system as the reference object point cloud area according to the corresponding spatial transformation matrix in the object point cloud spatial transformation matrix set. Use the registration algorithm in the 3D modeling technology to accurately align the transformed area with the reference object point cloud area, and fuse the transformed and aligned area with the reference object point cloud area. Repeat the above steps until all the remaining object point cloud areas are registered and fused with the reference object point cloud area. Finally, a complete 3D model of the target object is generated by integrating all the registered and fused point cloud data.
[0029] Step S3: Obtain the design characteristic information of the target intelligent gripper and the preset grasping and placement state of the target object, perform grasping strategy analysis based on the design characteristic information and the preset grasping and placement state, the structural material feature information of the target object and the quality information of the target object, and obtain initial grasping force strategy parameters. The design characteristic information and preset grasping and placement state of the target intelligent gripper are obtained. Specifically, the design characteristic information may include the size of the gripper, grasping method (such as mechanical claws, suction cups, etc.), driving method, range of motion, etc., which directly determine the actions that the gripper can perform and the accuracy that can be achieved; the preset grasping and placement state refers to the spatial position and placement angle where the target object needs to be placed after being grasped; a comprehensive evaluation of the target intelligent gripper and the target object is performed, and the initial grasping force strategy is obtained based on the design characteristic information and preset grasping and placement state, the structural material characteristic information of the target object, and the mass information of the target object, that is, the most appropriate grasping method, grasping point, and grasping force are selected. For example, if the target object is fragile, the gripper needs to choose a gentle grasping method and accurately control the grasping force; if the target object is heavy, the gripper may need to increase the number of grasping points or use a more powerful driving method; finally, the grasping strategy is analyzed and based on the analysis results, the initial grasping force strategy parameters are obtained, including the size of the grasping force, the location of the grasping points, the grasping order, etc., which directly guide the operation of the intelligent gripper in the actual grasping process.
[0030] In a possible implementation, step S3 further includes step S31, obtaining grasping mechanism information according to the design characteristic information, wherein the grasping mechanism information includes a clamping method and an adjustable threshold value of the gripper. The grasping mechanism information is based on the design characteristic information, and further clarifies the specific parameters and methods of how the gripper performs the grasping task in actual operation, including the clamping method and the adjustable threshold value of the gripper. Specifically, the clamping method describes the specific method or technology used by the gripper when grasping an object. For example, the gripper may adopt a rigid clamping method to fix the object by mechanical locking, or adopt a flexible clamping method to adapt to objects of different shapes by using the elastic deformation of the material; the adjustable threshold value of the gripper refers to the range or limit of the gripper's ability to automatically adjust its clamping force, position and other parameters during the grasping process. For example, for fragile or sensitive objects, the gripper needs to have a smaller adjustable threshold value to ensure that the object will not be damaged during the grasping process, while for larger or heavier objects, the gripper needs to have a larger adjustable threshold value to cope with different grasping requirements. The method further includes step S32, setting reference points for the structural material characteristic information of the target object, constructing a grabbing coordinate space, mapping the preset grabbing and placing state based on the grabbing coordinate space, and obtaining the preset grabbing and placing coordinate information. The reference points are set for the structural material characteristic information of the target object. Specifically, the selection of the reference points needs to consider factors such as the geometric shape, center of gravity position, material hardness, etc. of the object to ensure that the gripper can grab the object stably and accurately. The reference points are usually located at key positions of the object, such as protruding parts, recessed places, or specific identification points. The grabbing coordinate space is constructed based on the reference points to describe the relative position and posture relationship between the gripper and the target object. The preset grabbing and placing state is mapped based on the grabbing coordinate space. For example, the fixed point of the circular quasi-object is used as the reference point to construct a coordinate system to determine the relative coordinate information of the preset grabbing and placing state. Finally, these state information are converted into specific coordinate values as the preset grabbing and placing coordinate information. The method further includes step S33, performing grabbing contact point analysis on the grabbing mechanism information and the preset grabbing and placing coordinate information based on the structural material characteristic information of the target object to obtain the object grabbing contact point information. According to the structural material feature information of the target object, a detailed grasping contact point analysis is performed to obtain specific information of the grasping contact points of the object, including the position, quantity, distribution of the contact points, and the grasping force required for each contact point, etc. The method also includes step S34, performing grasping force analysis on the grasping contact point information of the object and the mass information of the target object, obtaining grasping contact force distribution threshold information, using the grasping contact force distribution threshold information as a constraint condition to perform grasping strategy optimization analysis, and obtaining the initial grasping force strategy parameters.Based on the object grasping contact point information, the actual contact position and method between the robot and the object during the grasping process are analyzed, and the size and direction of the force that the robot needs to apply during the grasping process are calculated. The grasping contact force distribution threshold information is obtained, which describes the range of force that the robot can apply at different contact points. These grasping contact force distribution threshold information are used as constraints to perform grasping strategy optimization analysis, that is, to find the optimal grasping strategy under given constraints, that is, to determine the initial grasping force strategy parameters that the robot should adopt during the grasping process.
[0031] In a possible implementation, step S33 further includes step S331, performing finite element meshing on the three-dimensional model of the target object to obtain a finite element mesh model of the target object, assigning grabbing attributes to the finite element mesh model of the target object based on the structural material feature information of the target object, and generating a finite element simulation model of the target object. Finite element meshing is to discretize the three-dimensional model of the target object into a series of small, interconnected units (i.e., meshes) for numerical calculation. Through finite element meshing, the continuous target object can be divided into a discrete, computable unit set. Based on the structural material feature information of the target object, the finite element mesh model is assigned grabbing attributes, including the material, texture, etc. of the object, and finally, a finite element simulation model of the target object is generated. It also includes step S332, performing grabbing contact point simulation on the grabbing mechanism information and the preset grabbing placement coordinate information based on the finite element simulation model of the target object to obtain a grabbing contact point simulation result set. The model is subjected to multi-point grasping simulation test, and the grasping contact point with the best stability is selected. Specifically, the grasping mechanism information and the preset grasping placement coordinate information are combined to construct a grasping scene of the robot in a virtual environment. The grasping process of the robot is simulated in this virtual scene using finite element simulation technology. By continuously adjusting the grasping mechanism parameters and grasping coordinate information, the changes in the contact points, the distribution of the object stress, and the possible deformation during the grasping process are observed. Finally, a grasping contact point simulation result set is obtained by multiple simulations and comparisons of the grasping effects under different schemes. It also includes step S333, in which the grasping stability of the grasping contact point simulation result set is evaluated respectively, and the grasping contact point information of the object is obtained by comparison. Perform grasping stability evaluation on each simulation result in the grasping contact point simulation result set, for example, evaluate the stress distribution of the contact point, the deformation of the object, the friction coefficient between the gripper and the object, etc., compare all grasping contact point simulation results, find the most stable and reliable grasping contact point, and obtain the object grasping contact point information, including the optimal grasping contact point position, number, distribution and the grasping force required for each contact point.
[0032] In a possible implementation, step S34 further includes step S341, using the grasping contact force distribution threshold information as a constraint condition, performing grasping simulation prediction on the target object finite element simulation model, and obtaining a target object grasping test data set. The grasping contact force distribution threshold information is used as a constraint condition and applied to the finite element simulation model of the target object, applying forces corresponding to the grasping mechanism on this model, and limiting the magnitude of these forces within the threshold range, simulating the force conditions of the robot in the actual grasping process, calculating the deformation, stress distribution and possible failure modes of the object in the grasping process, recording various data in the grasping process, such as the force distribution of the contact point, the displacement and speed of the object, etc., to form a target object grasping test data set. It also includes step S342, parsing the object grasping target, and obtaining a set of object grasping effect indicators, wherein the object grasping effect indicator set includes grasping stability, grasping accuracy and grasping energy consumption. The object grasping target is analyzed in detail to obtain a set of object grasping effect indicators, including grasping stability, grasping accuracy and grasping energy consumption. Specifically, grasping stability refers to the robot's ability to keep the object stable and not fall off during the grasping process; grasping accuracy refers to the deviation between the actual grasping position of the robot and the expected grasping position; grasping energy consumption refers to the energy consumed by the robot in completing the grasping task. It also includes step S343, evaluating the effect of the target object grasping test data set through the object grasping effect indicator set to obtain the target object grasping effect data set. The grasping test data is used to evaluate the performance of the robot in the actual grasping process, and specific grasping effect data is obtained according to the preset grasping effect indicators. It also includes step S344, constructing a gripper grasping control space according to the target object grasping test data set and the target object grasping effect data set, performing grasping strategy optimization analysis based on the gripper grasping control space, and determining the initial grasping force strategy parameters. The grasping control space of the gripper is a multi-dimensional parameter space, which includes various factors that affect the grasping effect, such as grasping force, grasping position, grasping posture, etc., which constitute the variables that the robot can adjust and control during the grasping process. The optimal grasping strategy parameter combination is searched in the control space to maximize grasping stability, minimize grasping error or reduce grasping energy consumption. Finally, the initial grasping force strategy parameters are determined through grasping strategy optimization analysis.
[0033] Step S4: The target intelligent gripper monitors the target object based on the initial grasping force strategy parameters, obtains grasping state feedback information, optimizes and adjusts the initial grasping force strategy parameters based on the grasping state feedback information, and determines the grasping force optimization strategy parameters to control the object grasping. After obtaining the initial grasping force strategy parameters, the target intelligent gripper performs actual grasping operations on the target object based on the initial grasping force strategy parameters, and closely monitors and provides real-time feedback during this process. Specifically, through built-in sensor modules, such as three-dimensional vision sensors and force sensors, the intelligent gripper can obtain various information about the grasping state in real time, including the contact between the gripper and the target object, the real-time changes in grasping force, the position changes of the target object, etc. Through analysis, the intelligent gripper can evaluate whether the current grasping state is stable and whether it meets the preset grasping and placement state requirements. Based on Based on the grasping status feedback information, the intelligent gripper will optimize and adjust the initial grasping force strategy parameters. For example, if it detects that the grasping force is too large or too small, or the position of the target object is unstable, the gripper will adjust the size or distribution of the grasping force accordingly to ensure that the grasping process is more stable and reliable. Finally, by optimizing the adjusted grasping force optimization strategy parameters, the intelligent gripper can control the grasping process more accurately, ensuring that the target object can be stably and safely grasped and placed in the preset position, which not only improves the success rate and efficiency of grasping, but also reduces the risk of object damage or grasping failure due to improper operation.
[0034] In a possible implementation, step S4 further includes step S41, if the grasping state feedback information does not achieve the preset grasping effect, grasping deviation analysis is performed on the grasping state feedback information to obtain the object grasping deviation parameter. Compare the differences between these actual grasping states and the preset grasping effect. If the grasping state feedback information does not achieve the preset grasping effect, grasping deviation analysis is performed to find out the cause of the deviation, including but not limited to unreasonable gripper design, improper grasping force control, grasping position deviation, etc. Finally, through grasping deviation analysis, a series of object grasping deviation parameters are obtained, such as position deviation, force deviation, posture deviation, etc. It also includes step S42, performing associated control parameter analysis on the object grasping deviation parameter, obtaining the control parameter variation direction, and setting parameter variation rules according to the control parameter variation direction. By analyzing the relationship between grasping deviation and control parameters, the direction of optimizing grasping effect is found, and corresponding parameter adjustment rules are formulated. Specifically, the grasping deviation parameters of the object are analyzed in association with control parameters, including an in-depth analysis of the association between various control parameters (such as grasping force, speed, position, etc.) and grasping deviation parameters (such as position deviation, force deviation, etc.) in the grasping process. After analyzing the association, it is necessary to determine which control parameters should be adjusted when optimizing the grasping effect, and in which direction these parameters should be adjusted. The determination of the variation direction is usually based on the size and nature of the grasping deviation and the sensitivity of the control parameters. For example, if the position deviation is large, the control parameters of the grasping position need to be adjusted. If the force deviation is obvious, the control parameters of the grasping force need to be adjusted. Finally, the parameter variation rules are set according to the variation direction of the control parameters. It also includes step S43, based on the parameter variation rules, the initial grasping force strategy parameters are subjected to variation simulation optimization to determine the grasping force optimization strategy parameters. According to the previously determined parameter variation rules, the initial grasping force strategy parameters are mutated. Specifically, on the basis of the initial parameters, the grasping force parameters are appropriately adjusted according to factors such as the direction, range and step size set in the rules. The mutated grasping force strategy parameters are simulated and optimized using simulation software or experimental platforms. That is, by simulating the actual grasping process, the grasping effects under different parameter combinations are evaluated, and the optimal parameter configuration is found. Through multiple simulations and comparisons, the optimal grasping force strategy parameters are gradually approached. These parameters are the force values that can achieve the best grasping effect under the current task conditions and constraints.
[0035] In a possible implementation, step S43 further includes step S431, and the parameter variation rule is specifically: ,in, To control the parameter variation value, is the original control parameter value, is a random number between 0 and 1. is a normal distribution function. It also includes step S432, mutating and expanding the initial grasping force strategy parameters based on the parameter variation rule to obtain a grasping force strategy variation parameter set. By applying the parameter variation rule, a series of mutation operations are performed on the initial grasping force strategy parameters, thereby generating a variation parameter set containing multiple different parameter combinations. Specifically, the initial parameters will be subjected to multiple mutation operations in the manner set in the rule to generate multiple new parameter combinations. The new parameter combinations will deviate from the initial parameters to a certain extent, but are still within a reasonable variation range. Each parameter combination in the set is a certain variation of the initial parameters, representing a different grasping force strategy. It also includes step S433, based on the object grasping effect indicator set, the grasping force strategy variation parameter set is subjected to parameter simulation optimization to determine the grasping force optimization strategy parameters. The simulation model is updated with the current state of the target object, and the effect simulation evaluation is performed to obtain the optimized strategy parameters.
[0036] In the above, refer to Figure 1 A method for controlling the gripping force of an intelligent gripper according to an embodiment of the present invention is described in detail. Figure 2 A smart gripper grasping force control device according to an embodiment of the present invention is described.
[0037] According to an embodiment of the present invention, a smart gripper grasping force control device is used to solve the technical problems of existing smart gripper grasping force control, which relies on a preset fixed force for grasping and lacks flexibility, resulting in damage to the grasped object or grasping failure, and achieves the technical effect of improving the accuracy, reliability and adaptability of the smart gripper operation. A smart gripper grasping force control device includes: a target smart gripper acquisition module 10, a target object three-dimensional model generation module 20, an initial grasping force strategy parameter acquisition module 30, and an object grasping control module 40.
[0038] A target intelligent gripper acquisition module 10, the target intelligent gripper acquisition module 10 is used to acquire a target intelligent gripper, the target intelligent gripper has a built-in sensor module, wherein the sensor module includes a three-dimensional visual sensor and a force sensor, the three-dimensional appearance point cloud data of the target object is acquired by the three-dimensional visual sensor, and the gripping force data of the target object is acquired by the force sensor test;
[0039] A target object three-dimensional model generation module 20, wherein the target object three-dimensional model generation module 20 uses a three-dimensional modeling technology to perform point cloud reconstruction on the three-dimensional appearance point cloud data to generate a target object three-dimensional model, and obtains structural material feature information of the target object based on the target object three-dimensional model, and performs mass conversion based on the grasping force data to obtain target object mass information;
[0040] An initial grasping force strategy parameter acquisition module 30, the initial grasping force strategy parameter acquisition module 30 is used to obtain the design characteristic information of the target intelligent gripper and the preset grasping and placement state of the target object, and perform grasping strategy analysis based on the design characteristic information and the preset grasping and placement state, the structural material feature information of the target object and the quality information of the target object to obtain the initial grasping force strategy parameters;
[0041] The object grasping control module 40 is used for the target intelligent gripper to monitor the grasping of the target object based on the initial grasping force strategy parameters, obtain grasping state feedback information, optimize and adjust the initial grasping force strategy parameters based on the grasping state feedback information, and determine the grasping force optimization strategy parameters to perform object grasping control.
[0042] The specific configuration of the target object three-dimensional model generation module 20 will be described in detail below. The target object three-dimensional model generation module 20 may further include: S21: obtaining point cloud attribute description factors, the point cloud attribute description factors including spatial position, reflection intensity and color texture; S22: using the point cloud attribute description factors to respectively describe the three-dimensional appearance point cloud data, and obtaining the object point cloud attribute description feature set; S23: based on the object point cloud attribute description feature set, key area selection is performed to obtain the reference object point cloud area feature information and the remaining object point cloud area feature information set; S24: the remaining object point cloud area information set and the reference object point cloud area information are sequentially associated and mapped to obtain the object point cloud space transformation matrix set; S25: using the three-dimensional modeling technology based on the object point cloud space transformation matrix set, the remaining object point cloud area feature information set is respectively registered and fused and reconstructed with the reference object point cloud area feature information to generate the target object three-dimensional model.
[0043] The specific configuration of the initial grasping force strategy parameter acquisition module 30 will be described in detail below. The initial grasping force strategy parameter acquisition module 30 further includes: S31: obtaining grasping mechanism information according to the design characteristic information, the grasping mechanism information including the clamping method and the adjustable threshold of the gripper; S32: setting the reference point of the structural material characteristic information of the target object, constructing the grasping coordinate space, mapping the preset grasping and placement state based on the grasping coordinate space, and obtaining the preset grasping and placement coordinate information; S33: performing grasping contact point analysis on the grasping mechanism information and the preset grasping and placement coordinate information based on the structural material characteristic information of the target object, and obtaining the object grasping contact point information; S34: performing grasping force analysis on the object grasping contact point information and the target object mass information, obtaining the grasping contact force distribution threshold information, using the grasping contact force distribution threshold information as a constraint condition to perform grasping strategy optimization analysis, and obtaining the initial grasping force strategy parameters.
[0044] The specific configuration of the initial grasping force strategy parameter acquisition module 30 will be described in detail below. The initial grasping force strategy parameter acquisition module 30 may further include: performing finite element meshing on the three-dimensional model of the target object to obtain a finite element mesh model of the target object, assigning grasping attributes to the finite element mesh model of the target object based on the structural material feature information of the target object, and generating a finite element simulation model of the target object; performing grasping contact point simulation on the grasping mechanism information and the preset grasping placement coordinate information based on the finite element simulation model of the target object to obtain a grasping contact point simulation result set; performing grasping stability evaluation on the grasping contact point simulation result set respectively, and comparing to obtain the grasping contact point information of the object.
[0045] The specific configuration of the initial grasping force strategy parameter acquisition module 30 will be described in detail below. The initial grasping force strategy parameter acquisition module 30 may further include: using the grasping contact force distribution threshold information as a constraint condition, performing grasping simulation prediction on the target object finite element simulation model, and obtaining a target object grasping test data set; parsing the object grasping target, and obtaining an object grasping effect index set, wherein the object grasping effect index set includes grasping stability, grasping accuracy, and grasping energy consumption; evaluating the effect of the target object grasping test data set through the object grasping effect index set, and obtaining a target object grasping effect data set; constructing a gripper grasping control space based on the target object grasping test data set and the target object grasping effect data set, and performing a grasping strategy optimization analysis based on the gripper grasping control space to determine the initial grasping force strategy parameters.
[0046] The specific configuration of the object grasping control module 40 will be described in detail below. The object grasping control module 40 may further include: S41: if the grasping state feedback information does not achieve the preset grasping effect, the grasping state feedback information is subjected to grasping deviation analysis to obtain the object grasping deviation parameter; S42: the object grasping deviation parameter is subjected to associated control parameter analysis to obtain the control parameter variation direction, and the parameter variation rule is set according to the control parameter variation direction; S43: the initial grasping force strategy parameter is subjected to variation simulation optimization based on the parameter variation rule to determine the grasping force optimization strategy parameter.
[0047] The specific configuration of the object grasping control module 40 will be described in detail below. The object grasping control module 40 further includes: the parameter variation rule is specifically: ,in, To control the parameter variation value, is the original control parameter value, is a random number between 0 and 1. is a normal distribution function; based on the parameter variation rule, the initial grasping force strategy parameters are mutated and expanded to obtain a grasping force strategy variation parameter set; based on the object grasping effect indicator set, the grasping force strategy variation parameter set is subjected to parameter simulation optimization to determine the grasping force optimization strategy parameters.
[0048] An intelligent gripper grasping force control device provided in an embodiment of the present invention can execute an intelligent gripper grasping force control method provided in any embodiment of the present invention, and has corresponding functional modules and beneficial effects of the execution method.
[0049] Although the present application makes various references to certain modules in the device according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.
[0050] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.
Claims
1. A method for controlling the gripping force of an intelligent gripper, characterized in that: The method comprises: S1: Acquire a target intelligent gripper, wherein the target intelligent gripper has a built-in sensor module, wherein the sensor module includes a three-dimensional vision sensor and a force sensor, collects and acquires three-dimensional appearance point cloud data of the target object through the three-dimensional vision sensor, and acquires grasping force data of the target object through the force sensor test; S2: reconstructing the three-dimensional appearance point cloud data by using three-dimensional modeling technology to generate a three-dimensional model of the target object, and obtaining structural material feature information of the target object based on the three-dimensional model of the target object, and performing mass conversion based on the grasping force data to obtain mass information of the target object; S3: obtaining design characteristic information of the target intelligent gripper and a preset grasping and placing state of the target object, performing grasping strategy analysis based on the design characteristic information and the preset grasping and placing state, structural material feature information of the target object and quality information of the target object, and obtaining initial grasping force strategy parameters; S4: the target intelligent gripper performs grasping monitoring on the target object based on the initial grasping force strategy parameters, obtains grasping state feedback information, optimizes and adjusts the initial grasping force strategy parameters based on the grasping state feedback information, and determines the grasping force optimization strategy parameters to perform object grasping control; The initial grasping force strategy parameters obtained in S3 include: S31: obtaining gripping mechanism information according to the design characteristic information, wherein the gripping mechanism information includes a clamping method and an adjustable threshold of a gripper; S32: Setting reference points for the structural material feature information of the target object, constructing a grabbing coordinate space, mapping the preset grabbing and placement state based on the grabbing coordinate space, and acquiring preset grabbing and placement coordinate information; S33: performing a grasping contact point analysis on the grasping mechanism information and the preset grasping placement coordinate information based on the structural material feature information of the target object to obtain object grasping contact point information; S34: performing grasping force analysis on the grasping contact point information of the object and the mass information of the target object to obtain grasping contact force distribution threshold information, performing grasping strategy optimization analysis using the grasping contact force distribution threshold information as a constraint condition, and obtaining the initial grasping force strategy parameters; The initial grasping force strategy parameters are obtained in S34, including: Using the grasping contact force distribution threshold information as a constraint condition, performing grasping simulation prediction on the target object finite element simulation model to obtain a target object grasping test data set; Analyzing the object grasping target to obtain a set of object grasping effect indicators, wherein the set of object grasping effect indicators includes grasping stability, grasping accuracy, and grasping energy consumption; Performing an effect evaluation on the target object grasping test data set by using the object grasping effect indicator set to obtain a target object grasping effect data set; According to the target object grasping test data set and the target object grasping effect data set, a gripper grasping control space is constructed, and a grasping strategy optimization analysis is performed based on the gripper grasping control space to determine the initial grasping force strategy parameters.
2. The method for controlling the gripping force of an intelligent gripper according to claim 1, characterized in that: The step S2 generates a three-dimensional model of the target object, including: S21: Obtaining point cloud attribute description factors, where the point cloud attribute description factors include spatial position, reflection intensity, and color texture; S22: performing attribute description on the three-dimensional appearance point cloud data respectively by using the point cloud attribute description factors to obtain a point cloud attribute description feature set of the object; S23: Selecting key areas based on the object point cloud attribute description feature set to obtain reference object point cloud area feature information and other object point cloud area feature information sets; S24: sequentially performing associative mapping on the remaining object point cloud region information sets and the reference object point cloud region information to obtain an object point cloud space transformation matrix set; S25: Using three-dimensional modeling technology, based on the object point cloud space transformation matrix set, the remaining object point cloud region feature information sets are respectively registered, fused and reconstructed with the reference object point cloud region feature information to generate the target object three-dimensional model.
3. The method for controlling the gripping force of an intelligent gripper according to claim 1, characterized in that: The step S33 of obtaining the object grasping contact point information includes: Performing finite element mesh division on the three-dimensional model of the target object to obtain a finite element mesh model of the target object, assigning grabbing attributes to the finite element mesh model of the target object based on the structural material feature information of the target object, and generating a finite element simulation model of the target object; Performing grasping contact point simulation on the grasping mechanism information and the preset grasping placement coordinate information based on the target object finite element simulation model to obtain a grasping contact point simulation result set; The grasping contact point simulation result sets are respectively evaluated for grasping stability, and the grasping contact point information of the object is obtained by comparison.
4. The method for controlling the gripping force of an intelligent gripper according to claim 1, characterized in that: Determining the grabbing strength optimization strategy parameters in S4 includes: S41: If the grasping state feedback information does not achieve a preset grasping effect, performing grasping deviation analysis on the grasping state feedback information to obtain an object grasping deviation parameter; S42: analyzing the object grasping deviation parameter in association with the control parameter to obtain a control parameter variation direction, and setting a parameter variation rule according to the control parameter variation direction; S43: performing a variation simulation optimization on the initial grasping force strategy parameters based on the parameter variation rule to determine the grasping force optimization strategy parameters.
5. The method for controlling the gripping force of an intelligent gripper according to claim 4, characterized in that: Determining the grasping force optimization strategy parameters in S43 includes: The parameter variation rules are specifically as follows: ,in, To control the parameter variation value, is the original control parameter value, is a random number between 0 and 1. is the normal distribution function; Based on the parameter variation rule, the initial grasping force strategy parameters are mutated and expanded to obtain a grasping force strategy variation parameter set; The grasping force optimization strategy parameters are determined based on the parameter simulation optimization of the grasping force strategy variation parameter set through the object grasping effect indicator set.
6. An intelligent gripper gripping force control device, characterized in that: The device is used to implement the intelligent gripper grasping force control method according to any one of claims 1 to 5, and the device comprises: A target intelligent gripper acquisition module, the target intelligent gripper acquisition module is used to acquire a target intelligent gripper, the target intelligent gripper has a built-in sensor module, wherein the sensor module includes a three-dimensional vision sensor and a force sensor, the three-dimensional appearance point cloud data of the target object is acquired by the three-dimensional vision sensor, and the gripping force data of the target object is acquired by the force sensor test; A target object three-dimensional model generation module, wherein the target object three-dimensional model generation module uses three-dimensional modeling technology to perform point cloud reconstruction on the three-dimensional appearance point cloud data to generate a target object three-dimensional model, and obtains structural material feature information of the target object based on the target object three-dimensional model, and performs mass conversion based on the grasping force data to obtain target object mass information; An initial grasping force strategy parameter acquisition module, the initial grasping force strategy parameter acquisition module is used to obtain the design characteristic information of the target intelligent gripper and the preset grasping and placement state of the target object, and perform grasping strategy analysis based on the design characteristic information and the preset grasping and placement state, the structural material feature information of the target object and the quality information of the target object to obtain the initial grasping force strategy parameters; An object grasping control module is used for the target intelligent gripper to monitor the grasping of the target object based on the initial grasping force strategy parameters, obtain grasping state feedback information, optimize and adjust the initial grasping force strategy parameters based on the grasping state feedback information, and determine the grasping force optimization strategy parameters for object grasping control.