A method for switching and grabbing and positioning of an electric screwdriver for different screw models

Abstract

The application discloses a kind of electric screwdriver switching and grabbing positioning method for different screw models, the intelligent switching of electric screwdriver mode and clamping mode is realized by end effector, the actual situation of target area image is dynamically adjusted according to the grabbing mode, not only improve the operation efficiency of automatic system, also ensure the grabbing accuracy when processing different targets, adapt the special needs of the mixed stacking of battery cell and screw in the process of battery pack disassembly;Through the adaptive selection operation of different screw models and electric screwdriver head, the speed and torque of electric screwdriver head can be automatically adjusted according to the identified screw model, effectively solve the problem that traditional electric screwdriver system is difficult to adapt to multiple screw models;By introducing hierarchical analysis module and attention mechanism module, the relative position of battery cell and screw is accurately calculated, on the one hand, the accuracy of subsequent mechanical arm grabbing action is guaranteed, on the other hand, the flexibility of automatic system is improved by packaging relative position data.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition and robotic arm grasping, and in particular to a method for switching electric screwdrivers and grasping and positioning screws of different screw models. Background Art

[0002] With the rapid development of electric vehicles and energy storage devices, the widespread use of battery packs has become a trend. However, after long-term use or retirement, the disassembly and recycling of battery packs has become a critical issue. Traditional battery pack disassembly relies primarily on manual operations, which is not only inefficient but also poses certain safety risks. During battery pack disassembly, it is particularly common to find battery cells mixed with other components (such as screws). How to effectively identify and handle these mixed situations has become a focus of current technical research.

[0003] Currently, the automated systems used on the market still have many shortcomings when it comes to identifying and handling mixed battery cells and screws in disordered stacking scenarios after battery pack disassembly. Existing technologies often rely on preset models to identify screws and battery cells, which often lacks flexibility when handling complex scenarios. For example, after a battery pack is disassembled, battery cells and screws may be mixed and stacked in disorder. In this case, traditional batch processing systems cannot effectively distinguish these different components, resulting in low grasping and processing accuracy. Moreover, existing automated systems often have difficulty achieving accurate grasping switching and automatic positioning when faced with a mixture of different types of screws and battery cells, which can easily lead to misoperation and inefficiency. In addition, existing robotic arm control systems often have difficulty accurately identifying and grasping mixed targets, especially when faced with randomly stacked battery cells and different types of screws, resulting in high error rates and unstable operation. These problems significantly limit the efficiency and adaptability of automated systems. They also severely restrict the level of automation in the battery pack disassembly process, making it impossible to meet the needs of large-scale battery recycling. Therefore, the research on the switching and grasping positioning methods of electric screwdrivers for different screw models and mixed stacking of battery cells has important practical application value and innovation. Summary of the Invention

[0004] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides a method for switching and grasping and positioning electric screwdrivers for different screw models to solve the problems raised in the background technology.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a method for switching and grasping and positioning electric screwdrivers for different screw models, comprising:

[0007] Acquire an image of the target area using a camera, and determine whether the current image of the target area is a disordered stacking state after the cells are separated based on a target area image recognition algorithm;

[0008] If it is determined that the current target area image is a disordered stacking state after the cells are separated, the current grasping mode is switched to the clamping mode through the end effector, and the robot arm programmable controller is used to control the robot arm to grasp the position of the cell and store the grasped cell;

[0009] Otherwise, if the current target area image shows a state where there are no battery cells and only screws, the current grasping mode is switched to the electric screw grasping mode through the end effector, the screw position in the current target area image is marked, the screw model is identified based on the screw position in the current target area image, and the robot arm programmable controller is used to control the robot arm to grasp the screw position;

[0010] The results of the current target area image recognition, the current grasping mode of the end effector, the battery cell model and the screw model are displayed in the visual operation panel.

[0011] As a preferred solution of the electric screwdriver switching and gripping positioning method for different screw models described in the present invention, it also includes:

[0012] The electric screwdriver method includes a plurality of electric screwdriver bits and the electric screwdriver bits are composed of an array, and the clamping method is a classified clamping method;

[0013] The visual operation panel is connected to the robot arm programmable controller, and the robot arm programmable controller and the end controller are regularly optimized through the display information in the visual operation panel.

[0014] As a preferred solution of the electric screwdriver switching and gripping positioning method for different screw models described in the present invention, the target area image recognition algorithm includes:

[0015] The target area image acquired by the camera is represented as unlabeled data, and a scene dataset is created to perform scene classification on the unlabeled data;

[0016] After scene classification, the scene dataset includes a state where there are no battery cells and only screws, and a disorderly stacking scene where the battery cells are separated;

[0017] By inputting the scene dataset into the DINO model, the target image in the current area is identified.

[0018] As a preferred solution of the electric screwdriver switching and gripping positioning method for different screw models described in the present invention, the DINO model includes:

[0019] Hierarchical analysis module, attention mechanism module and target detection module.

[0020] As a preferred solution of the electric screwdriver switching and gripping positioning method for different screw models described in the present invention, the hierarchical analysis module includes:

[0021] Based on the received scene dataset, all screw types are classified into several categories, each category represents a different screw thread shape, size or material, and the number of electric screw bits in each layer is determined;

[0022] Initialize the electric screwdriver array and set the initial number of array elements to zero;

[0023] According to the current screw type, the electric screwdriver bit type is gradually added to the array.

[0024] As a preferred solution of the electric screwdriver switching and grasping positioning method for different screw models described in the present invention, the attention mechanism module and the target detection module include:

[0025] The attention mechanism module performs weight assignment based on the scene dataset, extracts all the battery cells and screws in the scene dataset, assigns weights to all the battery cells and screws, sorts them from high to low according to the weights, and calculates the relative positions of the battery cells and screws in the scene in sequence according to the sorting order, thereby obtaining the relative positions of all the battery cells and screws in the scene;

[0026] The target detection module receives the relative positions of the battery cell and the screw in the scene, and encapsulates them as operation parameters and transmits them to the programmable controller of the robot arm to generate operation instructions.

[0027] As a preferred solution of the electric screwdriver switching and grasping positioning method for different screw models described in the present invention, the robot arm is controlled by a programmable controller of the robot arm, including:

[0028] The robotic arm receives the operation command issued by the DINO model and initializes the position and posture of the robotic arm, which is composed of the first rotation joint A, the second mobile joint B and the third ball joint C;

[0029] The first rotation joint A calculates the current rotation angle A1 of the robotic arm according to the position of the target object that the robotic arm needs to grasp;

[0030] Based on A1 and the placement angle L of the target object, a deviation value M=A1-L is calculated, and at the same time, a distance B1 required for the second movable joint B to move to the current object is calculated, and B1 is compared with the set target movement distance B2;

[0031] If the deviation value M is not zero and the distance between B1 and B2 gradually increases, the third ball joint C is always adjusted to make the deviation value M between A1 and L approach zero, and make B1 approach B2 or B2 approach B1. Otherwise, when the deviation value M is equal to zero and B1 and B2 overlap, the third ball joint C is not adjusted.

[0032] As a preferred solution of the electric screwdriver switching and gripping positioning method for different screw models described in the present invention, wherein: the electric screwdriver method includes a plurality of electric screwdriver heads and the electric screwdriver heads are composed of an array, and the clamping method is a classified clamping method, including:

[0033] The array contains a number of elements, each of which stores a type of electric screwdriver bit. The number of elements in the array is affected by the results of the hierarchical analysis module in the DINO model.

[0034] Before the clamping method is performed, the battery cells are classified by scanning the QR code on the battery cells, the classification results are saved in the battery cell category database, and the clamping force is set according to the battery cell category.

[0035] As a preferred solution of the electric screwdriver switching and grasping positioning method for different screw models described in the present invention, the grasping process of the disorderly stacked discharge core also includes:

[0036] By considering the clamping tension, each time the robot arm clamps a battery cell, the clamping force changes with the type of battery cell, and the clamping tension changes with the size of the battery cell;

[0037] Taking the shortest distance from the robot arm to the battery cell position as the objective function, the shortest path for the robot arm to grasp the battery cell is established.

[0038] As a preferred solution of the electric screwdriver switching and gripping positioning method for different screw models described in the present invention, the screw gripping process also includes:

[0039] The robot arm selects the corresponding electric screwdriver bit based on the currently identified screw model and sets the speed and torque of the electric screwdriver bit;

[0040] The speed and torque of the electric screwdriver bit are adjusted by considering the start-stop ratio of the electric screwdriver bit and the motion coefficient of the robot arm.

[0041] Compared with the prior art, the invention has the following beneficial effects:

[0042] 1. By introducing the DINO model, it can effectively distinguish between the mixed and disordered stacking of battery cells and screws, thereby achieving accurate scene recognition and classification. This greatly improves the adaptability of the automation system in complex scenarios and effectively reduces operational errors caused by recognition errors in traditional systems.

[0043] 2. The end effector enables intelligent switching between electric batching mode and gripping mode, dynamically adjusting the gripping method based on the actual image of the target area. This switching method not only improves the operating efficiency of the automation system, but also ensures gripping accuracy when handling different targets, adapting to the special needs of mixed stacking of battery cells and screws during battery pack disassembly.

[0044] 3. Through the adaptive selection operation of different screw models and electric screwdriver bits, the speed and torque of the electric screwdriver bit can be automatically adjusted according to the identified screw model, effectively solving the problem that traditional electric screwdriver systems are difficult to adapt to multiple screw models, and greatly improving the accuracy and efficiency of the assembly and disassembly process;

[0045] 4. By introducing a hierarchical analysis module and an attention mechanism module into the DINO model, the relative positions of the battery cells and screws can be accurately calculated. This not only ensures the accuracy of the subsequent robotic arm grasping action, but also improves the flexibility of the automation system by encapsulating relative position data. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0047] Figure 1 This is a general flow chart of a method for switching electric screwdrivers and grasping and positioning screws of different types according to an embodiment of the present invention;

[0048] Figure 2 This is a comparison chart of the grasping accuracy of the robotic arm in different scenarios for the electric screwdriver switching and grasping positioning method for different screw models according to one embodiment of the present invention;

[0049] Figure 3 This is a robustness test diagram of the DINO model in different scenarios for the electric screwdriver switching and gripping positioning method for different screw models according to an embodiment of the present invention;

[0050] Figure 4 This is a system diagram of a method for switching electric screwdrivers and grasping and positioning screws of different types according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0053] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0054] The present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0055] In the description of the present invention, it should be noted that the terms "upper, lower, inner, and outer" and other references to orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0056] In this disclosure, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they may refer to fixed, removable, or integral connections. They may also refer to mechanical, electrical, or direct connections, indirect connections through an intermediary, or internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in this disclosure.

[0057] Example 1

[0058] Reference Figure 1, which is the first embodiment of the present invention, provides a method for switching and grasping and positioning electric screwdrivers for different screw models, including:

[0059] S1. Acquire a target area image using a camera device, and determine whether the current target area image is a disordered stacking state after the battery cells are separated based on a target area image recognition algorithm;

[0060] Furthermore, the target area image captured by the camera device is represented as unlabeled data, and a scene dataset is created to perform scene classification on the unlabeled data;

[0061] It should be noted that the purpose of setting it as unlabeled data is to simulate the randomness of the target area image, so that the scene dataset is a state of mixed and disordered stacking of battery cells and screws before scene classification;

[0062] Specifically, after scene classification, the scene dataset includes a state where there are no battery cells and only screws, and a disorderly stacking scene where the battery cells are separated;

[0063] Furthermore, by inputting the scene dataset into the DINO model, the target image in the current area is identified;

[0064] It should be explained that the DINO (Distillation with No Labels) model is an unsupervised learning model that can learn image representations without labeled data through self-supervised learning.

[0065] Furthermore, the DINO model includes a hierarchical analysis module, an attention mechanism module, and an object detection module;

[0066] Furthermore, the hierarchical analysis module, based on the received scene dataset, first divides all screw types into several categories, each representing a different screw thread shape, size, or material, and determines the number of electric screwdriver bits in each layer. Secondly, it initializes the electric screwdriver bit array, setting the initial number of array elements to zero. Finally, based on the current screw type, it gradually adds electric screwdriver bit types to the electric screwdriver bit array.

[0067] For example, the steps in the hierarchical analysis module are described as follows:

[0068] First, the screw types are divided into three categories: Category 1: Small screws, thread shape: fine thread, size: diameter less than 2mm, material: aluminum alloy; Category 2: Medium screws, thread shape: coarse thread, size: diameter 2-5mm, material: stainless steel; Category 3: Large screws, thread shape: coarse thread, size: diameter greater than 5mm, material: steel. Based on the number of screws, the number of electric screwdriver bits on the current layer is 3;

[0069] Next, create an electric screwdriver bit array to store the electric screwdriver bits for different screw types. The initial number of array elements is zero, that is, the array is empty, represented as: electric_screwdriver_array = [];

[0070] According to the current screw type, gradually add the electric screwdriver bit type to the electric screwdriver bit array:

[0071] Task requirement 1: Compatible with aluminum alloy and fine threads;

[0072] Get processed category 1: small screws;

[0073] Perform action 1: add a small electric screwdriver bit to the array;

[0074] Execution result: electric_screwdriver_array.append("Small-Screwdriver-Head");

[0075] Current array status: electric_screwdriver_array = ["Small-Screwdriver-Head"];

[0076] Task requirement 2: Compatible with thick screws with a diameter of 2 to 5 mm;

[0077] Processing category 2: medium screws;

[0078] Perform action 2: add a medium-sized electric screwdriver bit to the array;

[0079] Execution result: electric_screwdriver_array.append("Medium-Screwdriver-Head");

[0080] Current array status: electric_screwdriver_array = ["Small-Screwdriver-Head","Medium-Screwdriver-Head"];

[0081] Task requirement 3: diameter greater than 5mm, steel;

[0082] Processing category 3: large screws;

[0083] Result: electric_screwdriver_array.append("Large-Screwdriver-Head");

[0084] Current array status: electric_screwdriver_array=["Small-Screwdriver-Head", "Medium-Screwdriver-Head", "Large-Screwdriver-Head"];

[0085] In summary, the final array of electric screwdriver bits contains the three types of electric screwdriver bits in the current layer, which are suitable for small, medium, and large screws respectively;

[0086] Furthermore, the attention mechanism module assigns weights based on the scene dataset, extracts all the battery cells and screws in the scene dataset, assigns weights to all the battery cells and screws, sorts them from high to low according to the weights, and calculates the relative positions of the battery cells and screws in the scene based on the sorting order, thus obtaining the relative positions of all battery cells and screws in the scene.

[0087] Preferably, the relative positions of the battery cell and the screw in the scene are calculated using a clustering-based relative position method;

[0088] It should be noted that the clustering algorithm can achieve dynamic updates of the target object's surrounding environment to adapt to changes in the target object due to human movement or application scenarios, ensuring the accuracy of the robot arm's acquisition of the target object's position;

[0089] Specifically, first, weights are assigned based on the scenarios in the scenario dataset. Then, weights are assigned to the cells and screws based on the weights assigned by the scenario dataset. The weight assignment rules for cells and screws are determined by location complexity, physical size, material properties, and task time.

[0090] Specifically, position complexity is represented by the state of the battery cell or screw in three-dimensional space:

[0091]

[0092] Among them, d x d y dz and dz are the distance coordinates of the target object (battery cell or screw) in three-dimensional space, expressed as the distance between the target object and the operation reference point (such as the initial position of the robot arm) (in meters or millimeters); for Euclidean and distance;

[0093] Specifically, the physical size is expressed as:

[0094] F m =log(S+1)(2)

[0095] Where S represents the physical size of the target; for battery cells, it is expressed as volume (unit: cubic millimeter); for screws, it is expressed as the product of the outer diameter and length of the thread (unit: square millimeter);

[0096] Specifically, the material properties are expressed as:

[0097]

[0098] Where H represents the hardness of the target material, usually expressed in Vickers hardness or Brinell hardness (unit: HV or HB); max Indicates the maximum hardness of the material in the current task; D indicates the density of the target material, usually expressed in kilograms per cubic meter (unit: kg / m 3 ), D max Expressed as the maximum density of the target material;

[0099] Specifically, the task time is expressed as:

[0100]

[0101] Among them, T left The remaining time of the current task, usually expressed in seconds or minutes. The smaller the task time, the higher the weight;

[0102] It should be noted that the weight value P can be obtained by adding the above formulas (1), (2), (3), and (4). By assigning weights, the larger the weight value, the higher the priority;

[0103] Furthermore, the target detection module receives the relative positions of the battery cell and the screw in the scene, encapsulates them as operation parameters and transmits them to the programmable controller of the robot arm to generate operation instructions;

[0104] It should be noted that the relative position data is encapsulated so that the programmable controller of the robot arm no longer needs to perform complex calculations. Instead, it can directly generate specific action instructions based on the transmitted parameters to guide the first rotation joint, the second mobile joint, and the third ball joint to adjust. In addition, the operation instructions are passed through a standardized operation parameter interface so that they can not only act on the current robot arm, but also on different robot arms. This allows different robot arms to receive relative position data in the same format, thereby improving the portability of the automation system. Even when the robot arm is replaced or the system is upgraded, the operation process can still be seamlessly connected.

[0105] S2. If it is determined that the current target area image is a disordered stacking state after the cells are separated, the current grasping mode is switched to a clamping mode through the end effector, and the robot arm programmable controller is used to control the robot arm to grasp the position of the cell and store the grasped cell;

[0106] Specifically, the clamping method is classified clamping;

[0107] Furthermore, before the clamping method is performed, the battery cells are classified by scanning the QR code on the battery cells, the classification results are saved in the battery cell category database, and the clamping force is set according to the battery cell category;

[0108] It should be noted that the clamping force needs to be set not only according to the chip type, but also according to the contact force distribution, friction coefficient and material of the end effector;

[0109] Furthermore, the robot arm programmable controller is used to control the robot arm process as follows:

[0110] The robotic arm receives the operation command issued by the DINO model and initializes the position and posture of the robotic arm. The posture is composed of the first rotation joint A, the second mobile joint B, and the third ball joint C.

[0111] It should be explained that the first revolute joint allows the robot arm to rotate around a fixed axis, the second mobile joint allows the robot arm to move in a straight line based on the first revolute joint, and the third ball joint allows the end of the robot arm to rotate freely in multiple planes based on the first revolute joint and the second mobile joint, that is, to control the rotation of the end effector;

[0112] The first rotational joint A calculates the current rotation angle A1 of the robotic arm based on the position of the target object that the robotic arm needs to grasp;

[0113] Specifically, the target object grasped by the robotic arm here is a battery cell or a screw;

[0114] Based on A1 and the placement angle L of the target object, the deviation value M = A1 - L is calculated. At the same time, the distance B1 required for the second movable joint B to move to the current object is calculated and compared with the set target movement distance B2;

[0115] If the deviation M is not zero and the distance between B1 and B2 gradually increases, the third ball joint C is always adjusted to make the deviation M between A1 and L approach zero and make B1 approach B2 or B2 approach B1. Otherwise, when the deviation M is equal to zero and B1 and B2 overlap, the third ball joint C is not adjusted.

[0116] It should be noted that by adjusting the third ball joint, the end effector does not need to pay too much attention to its own posture and angle, which improves the working efficiency of the robotic arm;

[0117] Furthermore, by considering the clamping tension, each time the robot arm clamps a cell, the clamping force changes with the cell type, and the clamping tension changes with the cell size;

[0118] It should be noted that scanning the QR code on the battery cell can only identify the type and model of the battery cell, but cannot determine the current size of the battery cell. The physical size obtained previously is the size of the battery cell relative to the image, not the actual size of the battery cell. Therefore, the actual size of the battery cell needs to be considered and the clamping tension needs to be set.

[0119] It should be explained that the clamping tension is the maximum opening distance between the two jaws (or other forms of clamps) when the robot arm is clamping an object;

[0120] Furthermore, the shortest distance from the robot arm to the battery cell position is used as the objective function to establish the shortest path for the robot arm to grasp the battery cell;

[0121] Specifically, the objective function Min(J) is:

[0122]

[0123] Where n is the number of cells, d(h i―1 ,h i ) indicates that the robot arm moves from position h i―1 Move to position h i The Euclidean sum distance d, P i Indicates the priority of cell i. The smaller the value, the higher the priority. i Expressed as the size of cell i, I i It is represented by the type of battery cell i;

[0124] It should be noted that the shortest path for grabbing the battery cell can be obtained by calculating the minimum objective function value. Since the relative position is previously obtained through the clustering algorithm, the accuracy of the robot arm's distance to the object does not need to be considered in the formula here, which greatly optimizes the shortest path calculation time.

[0125] S3. Otherwise, if the current target area image shows that there are no battery cells and only screws, the current grasping mode is switched to the electric screw grasping mode through the end effector, the screw position in the current target area image is marked, the screw model is identified based on the screw position in the current target area image, and the robot arm programmable controller is used to control the robot arm to grasp the screw position;

[0126] Specifically, the electric screwdriver method includes a plurality of electric screwdriver heads and the electric screwdriver heads are composed of an array;

[0127] Furthermore, there are several elements in the electric screwdriver bit array, each of which stores a type of electric screwdriver bit. The number of elements in the array is affected by the results of the hierarchical analysis module in the DINO model.

[0128] Furthermore, the robotic arm selects the corresponding electric screwdriver bit according to the currently identified screw model and sets the speed and torque of the electric screwdriver bit;

[0129] Specifically, at this time, the end effector will find the corresponding electric screwdriver bit through the electric screwdriver bit array to match the current screw model;

[0130] It should be noted that the speed and torque are taken into account because the traditional electric screwdriver operation can be used not only to collect screws but also to assemble and disassemble;

[0131] Furthermore, by considering the start-stop ratio of the electric screwdriver and the motion coefficient of the robot arm, the speed and torque of the electric screwdriver are adjusted;

[0132] It should be explained that the motion-stop ratio here refers to the ratio of the indexing time and the rest time of each time the end effector controls the electric screwdriver array; the robot arm motion coefficient here refers to the period ratio of each time the robot arm rotates to the electric screwdriver array;

[0133] Specifically, the formula for the start-stop ratio k of the electric screwdriver is:

[0134]

[0135] Where z is the number of elements in the electric screwdriver array;

[0136] Specifically, the formula for the motion coefficient v of the robotic arm is:

[0137]

[0138] It should be noted that the motion-stop ratio directly reflects the working efficiency of the end effector. A too high motion-stop ratio means that the end effector is performing electrical batching tasks most of the time, with less rest time and high efficiency. On the contrary, a too low motion-stop ratio indicates that the end effector spends more time waiting or adjusting, which is inefficient. Similarly, a too high motion coefficient means that the robot arm frequently performs high-energy-consuming operations.

[0139] S4, displaying the result of the current target area image recognition, the current grasping mode of the end effector, the battery cell model, and the screw model on the visual operation panel;

[0140] Furthermore, the visual operation panel is connected to the programmable controller of the robot arm, and the programmable controller of the robot arm and the end controller are optimized regularly through the display information in the visual operation panel;

[0141] It should be noted that the programmable controller of the robotic arm can be updated regularly by displaying and recording the start-stop ratio of the electric screwdriver bit and the motion coefficient of the robotic arm in the visual operation panel, and submitting the recorded values ​​to the DINO model for processing. The DINO model controls the number of electric screwdriver bits in the array of the end effector according to the screw types of all image scenes containing screws but no battery cells, so as to improve the start-stop ratio of the end effector and reduce the energy consumption of the robotic arm; the programmable controller of the robotic arm can be optimized by adjusting the weight distribution rules of the battery cells and screws to fine-tune the relative positions of the battery cells and screws to achieve the purpose of controlling the robotic arm.

[0142] Example 2

[0143] Reference Figure 2 and Figure 3 , which is the second embodiment of the present invention, and provides a method for switching and grasping and positioning electric screwdrivers for different screw models, including: In this embodiment, the purpose is to verify the effectiveness and superiority of a method for switching and grasping and positioning electric screwdrivers for different screw models;

[0144] Experimental Preparation: To ensure the simulation is realistic and effective, different types of battery cells and screws were selected as test objects. Specifically, the screw types included small aluminum alloy screws, medium stainless steel screws, and large steel screws, as well as three different battery cell models: LG INR18650 M36, Samsung SDI 94Ah, and BYD Blade LFP. The test environment was set up on a standard industrial robot arm, ABB IRB 2600, equipped with a Schunk PGN-plus 64 end effector and an ABB IRC5 programmable controller. The camera used was a Basler a2A1920-160umPRO industrial camera equipped with a DINO model for image recognition and object detection.

[0145] Experimental process: Screws and batteries of different types are randomly stacked in a 2m 2 The robot simulates the disordered stacking of disassembled battery packs in the target area. A camera is installed above the robotic arm to capture images of the area and input the image data into the DINO model.

[0146] The DINO model uses hierarchical analysis and attention mechanism modules to classify and locate different types of battery cells and screws. Based on the recognition results, the end effector's gripping method is adjusted: when a battery cell is identified, it switches to gripping mode to accurately grasp and store the cell; when a screw is identified, it switches to electric screwdriver mode, selects the corresponding electric screwdriver bit, and adjusts the speed and torque based on the screw's characteristics.

[0147] In this simulation experiment, we set up multiple control groups, including the Swin Transformer, ViT (Vision Transformer), and ResNet-50 models, as well as traditional preset models (rule-based image recognition and robotic arm control methods) and manual operation groups, to conduct a comprehensive comparative analysis. The results are shown in Table 1.

[0148] Table 1

[0149]

[0150]

[0151] Table 1 shows that the DINO model performs best, with a recognition accuracy of 98.5%, 1.5% higher than SwinTransformer, 2.2% higher than ViT, and 4.5% higher than ResNet-50. This shows that the DINO model has stronger recognition ability and accuracy when handling complex and diverse mixed stacking scenarios of screws and battery cells.

[0152] In terms of grasping efficiency, the DINO model can accurately grasp 20 battery cells per minute, surpassing the 18 of the Swin Transformer and the 17 of ViT. Furthermore, the ResNet-50 model has a grasping efficiency of 16, while the traditional preset model only has a grasping efficiency of 12. This shows that the DINO model can complete tasks more efficiently in actual operations, especially in scenarios that require efficient and continuous operation.

[0153] The DINO model achieved a 95% success rate in screw installation. This means that when handling different screw types, the DINO model can provide more precise parameter adjustments and operation guidance for the robot arm, effectively reducing installation errors and improving the overall stability of the production line.

[0154] In terms of average recognition time, the DINO model completes a recognition cycle in just 15 milliseconds, which is 3 milliseconds faster than the SwinTransformer model and 5 milliseconds faster than the ViT model. This is significantly faster than the 22 milliseconds of the ResNet-50 model. This enables the DINO model to respond more quickly even in a high-paced production environment, reducing the cumulative effect of screw and battery cell recognition time.

[0155] The DINO model's electric screwdriver bit rotation time is 2 seconds, which is faster than the 3 seconds of the Swin Transformer model. The ViT and ResNet-50 models require 3.5 seconds and 4 seconds, respectively. As we all know, the rotation time of electric screwdriver bits directly affects the continuity and efficiency of the production line. By reducing the rotation time, production downtime can be significantly reduced in actual operation, improving overall production efficiency.

[0156] The DINO model's robotic arm energy consumption was 0.45 kWh / hour, lower than the other models. This indicates that under the same operating conditions, the robotic arm controlled by the DINO model consumes less electricity and is more energy-efficient. In comparison, the Swin Transformer model and the ViT model consumed energy of 0.50 kWh / hour and 0.55 kWh / hour, respectively. The traditional preset model and manual operation consumed even higher energy consumption, at 0.75 kWh / hour and 0.70 kWh / hour, respectively. This demonstrates that the DINO model not only excels in performance but also reduces operating costs through energy management.

[0157] Finally, combine Figure 2 It can be seen that even when the scene is changed, the accuracy of the push robot arm in grasping the target object can always be maintained at 99%; and reference Figure 3 After testing the robustness of the DINO model after changing the scene, it was found that the screw installation success rate and mixed stacking recognition accuracy were always high, while the error rate and electric screwdriver head rotation time were always low, indicating that the DINO model is robust.

[0158] In summary, this simulation experiment, by introducing a variety of different types of battery cells and screws, further demonstrated the significant advantages of the DINO model in controlling the robotic arm in terms of recognition accuracy, grasping efficiency, operational stability, and energy consumption management, demonstrating its potential and innovation in actual industrial applications.

[0159] Example 3

[0160] Reference Figure 4 , which is the third embodiment of the present invention, provides an electric screwdriver switching and gripping positioning system for different screw models, including:

[0161] The target area image acquisition and recognition module first uses a Basler a2A1920-160umPRO industrial camera mounted on the robotic arm to capture an image of the target area, which includes randomly stacked battery cells and screws of different types. The captured image data is then fed into the DINO model for analysis. Finally, the DINO model uses its hierarchical analysis module to classify all screws and battery cells and identify the specific location and type of each target object.

[0162] The gripping mode switching and operation module first controls the end effector to switch modes based on the recognition results of the DINO model. When a battery cell is identified, the end effector switches to gripping mode, accurately grasping the battery cell and storing it in the specified location. When a screw is identified, the end effector switches to electric screwdriver mode and automatically selects the appropriate electric screwdriver bit based on the screw model. Secondly, before grasping the screw, the robot arm calculates the optimal gripping path through the controller to ensure efficient and accurate gripping operations. Finally, the speed and torque of the electric screwdriver bit are adjusted according to the specific screw model to ensure the success rate and stability of the installation.

[0163] The operation feedback and optimization module: First, during all operations, the robot's operating status, current gripping method, battery cell model, screw model, and other information are displayed in real time on the Siemens Simatic HMI TP1200 visual operation panel. Second, the operator can monitor the system's operation through this panel and manually intervene or adjust it as needed. Finally, the DINO model's weight distribution and hierarchical analysis module will provide data feedback through the robot's programmable controller and based on the robot's operation results, regularly optimizing the end effector to ensure the stability and efficiency of the robot's programmable controller and end controller in long-term operation.

[0164] Those skilled in the art will appreciate that the embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Furthermore, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present application may be implemented in various computer languages, for example, object-oriented programming language Java and interpreted scripting language JavaScript, etc.

[0165] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0166] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0167] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0168] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0169] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A method for switching and grasping and positioning electric screwdrivers for different screw types, characterized in that: include: Acquire an image of the target area using a camera, and determine whether the current image of the target area is a disordered stacking state after the cells are separated based on a target area image recognition algorithm; The target area image recognition algorithm includes: The target area image acquired by the camera is represented as unlabeled data, and a scene dataset is created to perform scene classification on the unlabeled data; After scene classification, the scene dataset includes a state where there are no battery cells and only screws, and a disorderly stacking scene where the battery cells are separated; By inputting the scene dataset into the DINO model, the target image in the current area is identified; The DINO model includes: Hierarchical analysis module, attention mechanism module and target detection module; If it is determined that the current target area image is a disordered stacking state after the cells are separated, the current grasping mode is switched to the clamping mode through the end effector, and the robot arm programmable controller is used to control the robot arm to grasp the position of the cell and store the grasped cell; Otherwise, if the current target area image shows a state where there are no battery cells and only screws, the current grasping mode is switched to the electric screw grasping mode through the end effector, the screw position in the current target area image is marked, the screw model is identified based on the screw position in the current target area image, and the robot arm programmable controller is used to control the robot arm to grasp the screw position; The method of controlling the robotic arm using a robotic arm programmable controller includes: The robotic arm receives the operation command issued by the DINO model and initializes the position and posture of the robotic arm, which is composed of the first rotation joint A, the second mobile joint B and the third ball joint C; The first rotation joint A calculates the current rotation angle A1 of the robotic arm according to the position of the target object that the robotic arm needs to grasp; Based on A1 and the placement angle L of the target object, a deviation value M=A1-L is calculated, and at the same time, the distance B1 required for the second movable joint B to move to the current object is calculated, and B1 is compared with the set target movement distance B2; If the deviation M is not zero and the distance between B1 and B2 gradually increases, the third ball joint C is always adjusted to make the deviation M between A1 and L approach zero and make B1 approach B2 or B2 approach B1. Otherwise, when the deviation M is zero and B1 and B2 overlap, the third ball joint C is not adjusted. The results of the current target area image recognition, the current grasping mode of the end effector, the battery cell model and the screw model are displayed in the visual operation panel.

2. The method for switching and grasping and positioning electric screwdrivers for different screw types according to claim 1, characterized in that: Also includes: The electric screwdriver method includes a plurality of electric screwdriver bits and the electric screwdriver bits are composed of an array, and the clamping method is a classified clamping method; The visual operation panel is connected to the robot arm programmable controller, and the robot arm programmable controller and the end controller are regularly optimized through the display information in the visual operation panel.

3. The method for switching and grasping and positioning electric screwdrivers for different screw types according to claim 1, characterized in that: Hierarchical analysis module, including: Based on the received scene dataset, all screw types are classified into several categories, each category represents a different screw thread shape, size or material, and the number of electric screw bits in each layer is determined; Initialize the electric screwdriver array and set the initial number of array elements to zero; According to the current screw type, the electric screwdriver bit type is gradually added to the array.

4. The method for switching and grasping and positioning electric screwdrivers for different screw types according to claim 1, characterized in that: The attention mechanism module and the target detection module include: The attention mechanism module performs weight assignment based on the scene dataset, extracts all the battery cells and screws in the scene dataset, assigns weights to all the battery cells and screws, sorts them from high to low according to the weights, and calculates the relative positions of the battery cells and screws in the scene in sequence according to the sorting order, thereby obtaining the relative positions of all the battery cells and screws in the scene; The target detection module receives the relative positions of the battery cell and the screw in the scene, and encapsulates them as operation parameters and transmits them to the programmable controller of the robot arm to generate operation instructions.

5. The method for switching and grasping and positioning electric screwdrivers for different screw types according to claim 2, characterized in that: The electric screwdriver method includes a plurality of electric screwdriver bits and the electric screwdriver bits are formed in an array. The clamping method is a classified clamping method, including: There are several elements in the array, each of which stores a type of electric screwdriver bit. The number of elements in the array is affected by the results of the hierarchical analysis module in the DINO model. Before the clamping method is performed, the battery cells are classified by scanning the QR code on the battery cells, the classification results are saved in the battery cell category database, and the clamping force is set according to the battery cell category.

6. The method for switching and grasping and positioning electric screwdrivers for different screw types according to claim 1 or 5, characterized in that: The process of grabbing the disorderly stacked cores also includes: By considering the clamping tension, each time the robot arm clamps a battery cell, the clamping force changes with the type of battery cell, and the clamping tension changes with the size of the battery cell; Taking the shortest distance from the robot arm to the battery cell position as the objective function, the shortest path for the robot arm to grasp the battery cell is established.

7. The method for switching and grasping and positioning electric screwdrivers for different screw types according to claim 1, characterized in that: The screw grabbing process also includes: The robotic arm selects the corresponding electric screwdriver bit based on the currently identified screw model and sets the speed and torque of the electric screwdriver bit. The speed and torque of the electric screwdriver bit are adjusted by considering the start-stop ratio of the electric screwdriver bit and the motion coefficient of the robotic arm.

Images