Machine Vision AI Wheel Robot and Method Based on Artificial Intelligence Algorithm

Abstract

The present invention relates to a machine vision AI wheel robot and method based on artificial intelligence algorithms. The wheel robot includes: a working area scanning module, a force field distribution analysis module, a grasping strategy adjustment module, a tire dynamic response capture module, a target placement area scanning module, and a transplant data recording module. By analyzing the surface pattern and force field distribution of the tire, the wheel robot can comprehensively understand the state and physical properties of the tire. This accurate identification and analysis of the tire state improve its adaptability under different working conditions and the accuracy of operations, can adjust the grasping strategy in real time, and ensure stable and effective grasping and placement tasks under various conditions. This intelligent decision-making ability significantly improves the operation efficiency and reduces the operation risks caused by environmental changes.

Description

Technical Field

[0001] The present invention relates to the field of wheel transportation, and particularly to a machine vision AI wheel robot and method based on artificial intelligence algorithms. Background Art

[0002] Wheel robots for wheel transplanting belong to the application direction of the intersection of industrial robots and artificial intelligence technologies, mainly focusing on automated tire handling, grasping, moving, and precise placement tasks. This field aims to achieve efficient, precise, and safe tire transplanting operations through machine vision, mechanical perception, and intelligent control technologies. Compared with the traditional methods relying on manual labor or simple mechanical equipment, the robots in this field can real-time sense the working environment, analyze the tire state through artificial intelligence algorithms and sensor networks, and dynamically adjust the operation strategy to adapt to complex industrial scenarios, while significantly improving the operation efficiency, reducing the labor intensity and cost investment. The core technologies in this field include three-dimensional environment modeling, dynamic analysis, autonomous navigation, grasping mechanics optimization, and real-time monitoring and feedback control.

[0003] In the existing technologies for tire grasping and transplanting tasks, the lack of dynamic adaptability is an obvious shortcoming. The existing technologies usually have difficulty in real-time sensing the dynamic performance of the tire during the grasping and moving processes, lacking the immediate adjustment ability based on sensor feedback. Therefore, operation errors are likely to occur in complex environments, reducing the success rate of the task. In the target placement link, the evaluation of the placement area by traditional technologies is often simple and rough, lacking a comprehensive analysis of key factors such as the surface flatness of the area, the inclination angle, and obstacle interference, resulting in inaccurate selection of the optimal placement position, thus increasing the possibility of tire slippage or unstable placement. In addition, a perfect adaptive control mechanism is generally lacking in the existing technologies, and abnormal situations such as the weightlessness state or attitude deviation of the tire cannot be responded to in a timely manner during the operation process, and the stability and safety of the operation are thus restricted. Summary of the Invention

[0004] The present invention aims at the technical problems existing in the prior art and provides a machine vision AI wheel robot and method based on artificial intelligence algorithms.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: A machine vision AI wheel robot based on artificial intelligence algorithms, the wheel robot includes:

[0006] A working area scanning module, configured to scan the working area, construct a three-dimensional map of the area, identify the area where the tire is located and the target placement area, identify obstacles in the map, and plan a navigation path for the wheel robot;

[0007] The force field distribution analysis module is used to scan the surface pattern of the target tire, generate a pattern feature map, analyze the pattern type, detect the tire surface information, record dynamic parameters, and generate a tire force field distribution map based on the tire shape, material, and stacking method;

[0008] The grasping strategy adjustment module is used to establish a behavior prediction model for the target tire based on the tire force field distribution map, predict the dynamic performance of the tire at different grasping angles and forces, and generate a dynamic adjustment range for the grasping strategy;

[0009] The tire dynamic performance capture module is used to capture the dynamic performance of the tire in real time during the tire transplanting process, sense the weightlessness of the tire based on the clamping feedback data, and recalculate the grasping force within the dynamic adjustment range if the weightlessness signal is triggered;

[0010] The target placement area scanning module is used to scan the status of the target placement area, predict the stability of the tire after placement through the behavior prediction model, and adjust the tire angle during the tire transplanting process based on the stability prediction result;

[0011] The transplant data recording module is used to record the tire data and transplant data of this transplant after the transplant task is completed.

[0012] Another object of the present invention is to provide a working method of a machine vision AI wheel robot based on an artificial intelligence algorithm, and the working method includes:

[0013] Scan the working area, construct a three-dimensional map of the area, identify the area where the tire is located and the target placement area, identify obstacles in the map, and plan a navigation path for the wheel robot;

[0014] Scan the surface pattern of the target tire, generate a pattern feature map, analyze the pattern type, detect the tire surface information, record dynamic parameters, and generate a tire force field distribution map based on the tire shape, material, and stacking method;

[0015] For the target tire, establish a behavior prediction model based on the tire force field distribution map, predict the dynamic performance of the tire at different grasping angles and forces, and generate a dynamic adjustment range for the grasping strategy;

[0016] Capture the dynamic performance of the tire in real time during the tire transplanting process, sense the weightlessness of the tire based on the clamping feedback data, and recalculate the grasping force within the dynamic adjustment range if the weightlessness signal is triggered;

[0017] Scan the status of the target placement area, predict the stability of the tire after placement through the behavior prediction model, and adjust the tire angle during the tire transplanting process based on the stability prediction result;

[0018] After the transplanting task is completed, record the tire data and transplanting data for this transplanting.

[0019] As a further solution of the present invention, scan the working area, construct a three-dimensional map of the area, identify the area where the tire is located and the target placement area, identify obstacles in the map, and plan a navigation path for the wheel robot, specifically including:

[0020] Collect data on the working area, capture the contour information of the working area, identify the tire stacking position and the target placement position, obtain the complete spatial data of the working area, and generate a depth map;

[0021] Based on the generated depth map, convert the environmental data into point cloud data, integrate multi-view data through point cloud stitching technology, and use pose calibration and overlapping area alignment technology to establish a three-dimensional map including static objects and dynamic objects;

[0022] Mark the tire stacking position and the target placement area in the three-dimensional map, and at the same time extract the geometric features of the tire, including size, position and relative stacking relationship, and record the boundary information of the target placement area;

[0023] Analyze the obstacles in the three-dimensional map, and based on the tire stacking position and the target placement area, combine the marked obstacle range to generate the shortest obstacle avoidance path.

[0024] As a further solution of the present invention, scan the surface pattern of the target tire, generate a pattern feature map, analyze the pattern type, and at the same time detect the tire surface information and record the dynamic parameters. Based on the tire shape, material and stacking method, generate a tire force field distribution map, specifically including:

[0025] Scan the target tire to capture the surface pattern, material properties, geometric features and texture features of the tire, and obtain the image information and depth information of the tire;

[0026] Preprocess the obtained image information and depth information of the tire, extract the pattern features of the tire from the preprocessed image, generate a pattern feature map, analyze the pattern type, depth and distribution, and identify the tire wear state in combination with the tire type;

[0027] Based on the geometric features, material properties and pattern features of the tire, construct a tire force field distribution model, including the tire force area, friction force distribution and center of gravity position;

[0028] Combine the image information and depth information of the tire, the tire wear state, the geometric features, material properties and pattern features of the tire, and the force field distribution model to generate a complete tire force field distribution map.

[0029] As a further solution of the present invention, based on the tire force field distribution map, a behavior prediction model is established to predict the dynamic performance of the tire under different grasping angles and forces, and a dynamic adjustment range of the grasping strategy is generated, specifically including:

[0030] Based on the generated tire force field distribution map, a behavior prediction model is constructed, and the geometric characteristics and force field distribution data of the tire are input to predict the dynamic performance of the tire under changes in the grasping point, angle, and force, including the slip angle, tilt trend, and stability after grasping;

[0031] Based on the behavior prediction model, a grasping strategy for the target tire is generated, including the grasping point, grasping force range, and grasping angle range.

[0032] As a further solution of the present invention, the prediction of the dynamic performance of the tire under changes in the grasping point, angle, and force, including the slip angle, tilt trend, and stability after grasping, specifically is:

[0033] Calculate the slip angle:

[0034] ;

[0035] Wherein, represents the slip angle, that is, the angle at which the tire slides after grasping, is the tangential force, that is, the force parallel to the contact surface, is the normal force, that is, the force perpendicular to the contact surface;

[0036] Calculate the tilt trend:

[0037] ;

[0038] Wherein, is the change amount of the tilt angle, that is, the degree of tilt of the tire during the grasping process, is the weight of the tire, is the center of gravity offset distance, calculated from the force field distribution map, is the moment of inertia of the tire;

[0039] Calculate the stability after grasping:

[0040] ;

[0041] Wherein, is the grasping stability index. When it indicates that the grasping is in a stable state. When it indicates that the grasping fails, is the friction coefficient, calculated from the tire material properties and texture characteristics.

[0042] As a further aspect of the present invention, generating a grasping strategy for the target tire, including grasping points, grasping force range, and grasping angle range, specifically:

[0043] Calculate the comprehensive score of each candidate grasping point:

[0044] ;

[0045] Wherein, is the comprehensive score of the grasping point , is the weight factor;

[0046] Determine the grasping force range:

[0047] The frictional force generated by the clamping force is greater than the tangential force plus a safety margin , that is:

[0048] ;

[0049] The grasping force does not exceed the maximum bearing force allowed by the tire , that is:

[0050] ;

[0051] Obtain the final grasping force range:

[0052] ;

[0053] Determine the grasping angle range:

[0054] ;

[0055] Wherein:

[0056] ;

[0057] ;

[0058] and represent the minimum and maximum values of the grasping angle, represents the allowable inclination trend threshold, represents the grasping stability when the grasping angle is .

[0059] As a further aspect of the present invention, during the tire transplanting process, the dynamic performance of the tire is captured in real time, and the weight loss of the tire is sensed based on the clamping feedback data. If a weight loss signal is triggered, the grasping force is recalculated within the dynamic adjustment range;

[0060] Monitor the dynamic performance of the tire during the transplanting process in real time, including the displacement, attitude change, tilt angle and slip trend of the tire, and compare the monitoring data with the prediction results of the behavior prediction model;

[0061] The force sensors and tactile sensors installed on the gripper collect the clamping force information in real time to ensure that the grasping force is within the predetermined range, and calculate the instantaneous changes of the normal force and tangential force according to the feedback data;

[0062] Combine the real-time monitoring data and the clamping feedback data to determine whether to trigger the weightlessness signal, including:

[0063] ;

[0064] If the weightlessness signal is triggered, dynamically adjust the range according to the pre-generated grasping strategy, recalculate the grasping force, and maximize the normal force applied by the gripper based on the force field distribution map ;

[0065] ;

[0066] ;

[0067] Among them, is the current tangential force, is the additional safety margin;

[0068] Execute the dynamically adjusted grasping strategy in real time, and at the same time perform closed-loop control on the grasping action by combining the monitoring data and the feedback data.

[0069] As a further solution of the present invention, the state of the scanning target placement area is predicted for the stability of the tire after placement through a behavior prediction model, and based on the stability prediction result, the tire angle is adjusted during the tire transplanting process, specifically including:

[0070] Extract the target placement area from the three-dimensional map, partition the surface of the target placement area, divide the area into a suitable placement area and a high-risk area, and at the same time mark the area priority score by calculating the surface flatness, tilt angle and obstacle interference degree;

[0071] Based on the geometric characteristics of the target placement area and the force field distribution map of the tire, combined with the center of gravity position, calculate the stability index of the tire after placement ;

[0072] ;

[0073] Set the index threshold. If the stability index is lower than the index threshold, recalculate the placement angle and position of the tire, and adjust the direction of the tire during the transplanting process;

[0074] During the process of tire transplantation, the jaw angle and attitude of the tire are adjusted in real time to make the tire conform to the updated placement strategy, ensuring that the final placement position and angle meet the stability requirements;

[0075] After the tire is successfully placed, the force sensor and vision system are used to confirm whether the tire is in a stable state, and at the same time, the state data of the placement area and the final placement parameters of the tire are recorded.

[0076] As a further solution of the present invention, if the stability index is lower than the index threshold, the placement angle and position of the tire are recalculated, specifically:

[0077] ;

[0078] ;

[0079] Among them, is the new placement angle, is the placement angle, represents the placement angle that makes the stability index reach the maximum value , represents the new placement coordinates, represents the priority score of the target placement area, represents the placement position that makes the priority score of the target placement area reach the maximum value .

[0080] The beneficial effects of the present invention are:

[0081] By analyzing the surface pattern and force field distribution of the tire, the wheel robot can comprehensively master the state and physical characteristics of the tire. This precise identification and analysis of the tire state improve its adaptability and operation accuracy under different working conditions, and can adjust the grasping strategy in real time to ensure stable and effective grasping and placement tasks under various conditions. This intelligent decision-making ability significantly improves the operation efficiency and reduces the operation risk caused by environmental changes.

[0082] In addition, during the execution of the transplantation process, the dynamic performance of the tire is monitored in real time, and weightlessness perception is carried out based on the clamping feedback data. This function ensures that the robot can quickly respond to abnormal situations during operation, form an adaptive control mechanism, and thus optimize the force and angle adjustment during grasping and placement, making the operation safer and more reliable.

[0083] In terms of the selection of the target placement area and stability prediction, the robot can efficiently determine the optimal placement position through detailed area analysis, avoiding the time waste and resource loss caused by blind attempts. This data-driven decision-making process not only improves the success rate of tire placement but also effectively reduces the complexity of subsequent operations. Brief Description of the Drawings

[0084] Figure 1 It is a structural block diagram of a machine vision AI wheel robot based on an artificial intelligence algorithm provided by an embodiment of the present invention;

[0085] Figure 2 It is a flowchart of a working method of a machine vision AI wheel robot based on an artificial intelligence algorithm provided by an embodiment of the present invention;

[0086] Figure 3 It is a flowchart of planning a navigation path for a wheel robot provided by an embodiment of the present invention;

[0087] Figure 4 It is a flowchart of generating a tire force field distribution map provided by an embodiment of the present invention;

[0088] Figure 5 It is a flowchart of generating a dynamic adjustment range of a grasping strategy provided by an embodiment of the present invention;

[0089] Figure 6 It is a flowchart of performing weightlessness perception on a tire based on clamping feedback data provided by an embodiment of the present invention;

[0090] Figure 7 It is a flowchart of predicting the stability after tire placement provided by an embodiment of the present invention. Detailed Embodiments

[0091] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present application.

[0092] In the description of the present application, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present application, "a plurality" means two or more, unless otherwise specifically defined.

[0093] In the description of the present application, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present application is not necessarily construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without the use of these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed in the present application.

[0094] Figure 1 The structural block diagram of the machine vision AI wheel robot based on the artificial intelligence algorithm provided by the embodiment of the present invention is as Figure 1 shown, and the wheel robot includes:

[0095] A working area scanning module 100, configured to scan the working area, construct a three-dimensional map of the area, identify the area where the tire is located and the target placement area, identify obstacles in the map, and plan a navigation path for the wheel robot;

[0096] A force field distribution analysis module 200, configured to scan the surface pattern of the target tire, generate a pattern feature map, analyze the pattern type, detect the surface information of the tire, record dynamic parameters, and generate a tire force field distribution map based on the tire shape, material, and stacking method;

[0097] A grasping strategy adjustment module 300, configured to establish a behavior prediction model for the target tire based on the tire force field distribution map, predict the dynamic performance of the tire at different grasping angles and forces, and generate a dynamic adjustment range for the grasping strategy;

[0098] A tire dynamic performance capture module 400, configured to capture the dynamic performance of the tire in real time during the tire transplanting process, sense the weightlessness of the tire based on the clamping feedback data, and recalculate the grasping force within the dynamic adjustment range if a weightlessness signal is triggered;

[0099] A target placement area scanning module 500, configured to scan the state of the target placement area, predict the stability of the tire after placement through the behavior prediction model, and adjust the angle of the tire during the tire transplanting process based on the stability prediction result;

[0100] A transplant data recording module 600, configured to record the tire data and transplant data of this transplant after the transplant task is completed.

[0101] Figure 2The flowchart of the working method of the machine vision AI wheel robot based on the artificial intelligence algorithm provided by the embodiments of the present invention is as follows: Figure 2 As shown, the working method includes:

[0102] S100, scan the working area, construct a three-dimensional map of the area, identify the area where the tire is located and the target placement area, identify obstacles in the map, and plan a navigation path for the wheel robot;

[0103] During the scanning process, the robot will focus on identifying the stacking position and the target placement position of the tire, generate spatial data of the entire working area, and generate a depth map through algorithm processing. The depth map can clearly show each layer of the three-dimensional space, laying a foundation for the subsequent integration of point cloud data.

[0104] Next, the robot converts the depth map into point cloud data. Point cloud data is a high-precision three-dimensional representation method, which can accurately depict every detail in the environment through the distribution of points. Through point cloud stitching technology, the robot can use data collected from multiple angles to integrate into a seamless and complete three-dimensional map. In this process, pose calibration and overlapping area alignment technology is particularly crucial. It can calculate the relative position relationship of multi-viewpoint cloud data through algorithms, accurately match the overlapping areas, so as to ensure the accuracy of the stitched data and avoid data deviation affecting the quality of the final three-dimensional map.

[0105] In the generated three-dimensional map, the robot further marks the tire stacking position and the target placement area through a geometric feature extraction algorithm. This link not only identifies the tire stacking position, but also analyzes its size, stacking method and hierarchical relationship, as well as the boundary information of the target placement area. Through these data, the robot has a comprehensive understanding of the structure of the working area and the geometric characteristics of the tire. This step will also combine obstacle data to analyze the position, shape and range of static and dynamic obstacles in the three-dimensional map.

[0106] Finally, based on the above data, the robot combines a path planning algorithm to generate an obstacle avoidance strategy. The core goal of this strategy is to find the shortest path from the tire stacking area to the target placement area while ensuring that all obstacles are avoided on the path. The robot will comprehensively consider the length, complexity and safety of the path to generate an optimal navigation path, laying a foundation for subsequent accurate execution.

[0107] As Figure 3 shown, the scanning of the working area, constructing a three-dimensional map of the area, identifying the area where the tire is located and the target placement area, identifying obstacles in the map, and planning a navigation path for the wheel robot specifically includes:

[0108] S110, Collect data from the working area, capture the contour information of the working area, identify the tire stacking position and the target placement position, obtain the complete spatial data of the working area, and generate a depth map.

[0109] S120, Based on the generated depth map, convert the environmental data into point cloud data, integrate the multi-view data through point cloud stitching technology, and use the pose calibration and overlapping area alignment technology to establish a three-dimensional map including static and dynamic objects.

[0110] S130, Mark the tire stacking position and the target placement area in the three-dimensional map. At the same time, extract the geometric features of the tires, including size, position, and relative stacking relationship, and record the boundary information of the target placement area.

[0111] S140, Analyze the obstacles in the three-dimensional map. Based on the tire stacking position and the target placement area, combined with the marked obstacle range, generate the shortest obstacle avoidance path.

[0112] S200, Scan the surface pattern of the target tire to generate a pattern feature map, analyze the pattern type, and at the same time detect the surface information of the tire and record the dynamic parameters. Based on the tire shape, material, and stacking method, generate a tire force field distribution map.

[0113] This step will use a high-resolution camera and a laser scanning device to conduct a detailed scan of the target tire to capture multi-dimensional data such as its surface pattern, material properties, geometric features, and texture features. The robot can comprehensively perceive the surface information of the tire, including the pattern of the tire, the subtle changes in the surface texture, the characteristics of the material (such as rubber type and hardness), and the parameters of the tire geometry (such as diameter, thickness, and arc). This data collection process lays a solid foundation for subsequent analysis.

[0114] After obtaining the data, the robot will preprocess the acquired image information and depth information of the tire. This usually includes noise reduction, contrast enhancement, edge detection, and image segmentation to improve the quality and usability of the data. The preprocessed image will be used to extract the pattern features of the tire and generate a pattern feature map. This pattern feature map can clearly show the features such as the type, depth, distribution pattern, and directionality of the pattern. At the same time, the robot will further analyze the wear state of the tire by combining the pattern type and the distribution of the surface texture. For example, by comparing the changes in the pattern depth, the robot can evaluate whether there are obvious wear areas or abnormal damage on the tire.

[0115] Subsequently, based on the geometric characteristics, material properties, tread patterns, and wear conditions of the tire, the robot will start to construct a force field distribution model of the tire. The force field distribution model is a highly complex and dynamic description that synthesizes key mechanical information such as the force-bearing areas, friction force distribution, and center of gravity position of the tire. For example, the robot will calculate the magnitude of the forces acting on the tire at different contact points through algorithms, analyze which areas have higher friction forces, and even predict the possible behavior of the tire during grasping and movement through changes in the center of gravity position. This model provides the robot with an accurate description of the physical properties of the tire.

[0116] Finally, the robot will integrate all the characteristic data of the tire and its corresponding force field distribution model to generate a complete force field distribution map of the tire. This distribution map not only shows the physical and mechanical properties of the tire but also provides rich information support for subsequent grasping and movement operations. For example, through the force field distribution map, the robot can quickly determine the most suitable area for grasping, the grasping force that needs to be applied, and how to adjust the grasping angle to maintain the balance and stability of the tire.

[0117] The generation of the tread pattern map and the force field distribution map enables the robot to clearly identify the state of the tire, such as the degree of wear and material changes. This ability greatly enhances the adaptability of the robot in different task scenarios. Secondly, the construction of the force field distribution model provides a scientific basis for the robot during grasping and movement, reducing the risk of tire slipping or deformation caused by improper grasping. In addition, the comprehensive analysis of the dynamic parameters of the tire (such as friction force and center of gravity position) enables the robot to adjust the grasping strategy in real time to ensure the accuracy and safety of the operation.

[0118] As Figure 4 shown, scan the surface tread pattern of the target tire, generate a tread pattern map, analyze the tread pattern type, and at the same time detect the surface information of the tire and record the dynamic parameters. Based on the tire shape, material, and stacking method, generate a force field distribution map of the tire, specifically including:

[0119] S210, scan the target tire to capture the surface tread pattern, material properties, geometric characteristics, and texture characteristics of the tire, and obtain the image information and depth information of the tire;

[0120] S220, preprocess the obtained image information and depth information of the tire, extract the tread pattern characteristics of the tire from the preprocessed image, generate a tread pattern map, analyze the tread pattern type, depth, and distribution, and identify the tire wear state in combination with the tire type;

[0121] S230, based on the geometric characteristics, material properties, and tread pattern characteristics of the tire, construct a force field distribution model of the tire, including the force-bearing area, friction force distribution, and center of gravity position of the tire;

[0122] S240. Combine the image information and depth information of the tire, the tire wear state, the geometric features, material properties and tread patterns of the tire, and the force field distribution model to generate a complete tire force field distribution map.

[0123] S300. For the target tire, based on the tire force field distribution map, establish a behavior prediction model to predict the dynamic performance of the tire under different grasping angles and forces, and generate the dynamic adjustment range of the grasping strategy.

[0124] This step first needs to combine the geometric features of the tire (such as size, shape and structure) with the force field distribution data (such as friction force, center of gravity position and force condition) to form a comprehensive description of the tire characteristics. Through in-depth analysis of this data, the robot can identify the dynamic performance of the tire under different grasping conditions. This includes key parameters such as the slip angle, tilt trend that the tire may exhibit under different grasping points, grasping angles and grasping forces, and the stability of the tire after grasping.

[0125] In the process of constructing the behavior prediction model, the robot will use machine learning and physical simulation technologies, and through training algorithms to identify and predict the impact of different operating conditions on the tire behavior. Such a model can not only make predictions based on static data, but also be updated in real time to adapt to the rapidly changing environment. For example, when the robot collects the state data of the tire in real time during the grasping process, the model can quickly analyze and predict the changes, and even make corresponding adjustments during the grasping process, which can improve the success rate of grasping and reduce the risk of errors.

[0126] Based on the behavior prediction model, the robot will generate a grasping strategy for the target tire, which includes detailed information such as the grasping point, force range and angle range. The selection of the grasping point is based on the force field distribution map to ensure that the tire is evenly stressed during the grasping process and avoid deformation or damage caused by excessive local force application. The force range and angle range for grasping are generated through in-depth analysis of the tire characteristics to ensure that the robot can effectively prevent slipping and tilting during grasping. Especially in a complex stacking environment, the rationality of the grasping strategy directly affects the success rate of tire movement and placement.

[0127] This step significantly enhances the intelligence and autonomy of the wheeled robot during the tire transplanting task through an accurate behavior prediction model. First, the establishment of the model enables the robot to dynamically adjust its grasping strategy based on real-time data, enhancing its adaptability to environmental changes, thereby improving the success rate and efficiency of grasping. Second, through the precise prediction of the dynamic behavior of the tire, the robot can effectively avoid losses caused by improper grasping, such as tire sliding, tilting, or instability. In addition, the refinement of the grasping strategy not only improves the safety of the operation but also optimizes the overall workflow, reducing time waste and labor costs caused by incorrect operations.

[0128] As Figure 5 shown, based on the tire force field distribution map, a behavior prediction model is established to predict the dynamic behavior of the tire under different grasping angles and forces, and generate a dynamic adjustment range of the grasping strategy, specifically including:

[0129] S310, Based on the generated tire force field distribution map, construct a behavior prediction model, input the geometric features and force field distribution data of the tire, and predict the dynamic behavior of the tire under changes in the grasping position, angle, and force, including the slip angle, tilt trend, and stability after grasping;

[0130] S320, Based on the behavior prediction model, generate a grasping strategy for the target tire, including the grasping position, grasping force range, and grasping angle range.

[0131] In this step, predicting the dynamic behavior of the tire under changes in the grasping position, angle, and force, including the slip angle, tilt trend, and stability after grasping, specifically means:

[0132] Calculate the slip angle:

[0133] ;

[0134] Where represents the slip angle, that is, the angle at which the tire slides after grasping, is the tangential force, that is, the force parallel to the contact surface, is the normal force, that is, the force perpendicular to the contact surface;

[0135] Calculate the tilt trend:

[0136] ;

[0137] Where is the change in tilt angle, that is, the degree of tilt of the tire during grasping, is the tire weight, is the center of gravity offset distance, calculated from the force field distribution map, is the moment of inertia of the tire;

[0138] Calculate the stability after grasping:

[0139] ;

[0140] Among them, is the grasping stability index. When , it means that the grasping is in a stable state. When , it means that the grasping fails. is the friction coefficient, which is calculated from the tire material properties and texture characteristics.

[0141] Furthermore, the generation of the grasping strategy for the target tire includes the grasping point, the grasping force range, and the grasping angle range, specifically:

[0142] Calculate the comprehensive score of each candidate grasping point:

[0143] ;

[0144] Among them, is the comprehensive score of the grasping point , is the weight factor;

[0145] Determine the grasping force range:

[0146] The frictional force generated by the clamping force is greater than the tangential force plus a safety margin , that is:

[0147] ;

[0148] The grasping force does not exceed the maximum bearing force allowed by the tire , that is:

[0149] ;

[0150] Obtain the final grasping force range:

[0151] ;

[0152] Determine the grasping angle range:

[0153] ;

[0154] Among them:

[0155] ;

[0156] ;

[0157] and represent the minimum and maximum values of the grasping angle, represent the threshold of the allowed tilting trend, represent the grasping stability when the grasping angle is .

[0158] S400, during the tire transplanting process, it captures the dynamic performance of the tire in real time, and based on the clamping feedback data, it senses the weight loss of the tire. If the weight loss signal is triggered, it recalculates the grasping force within the dynamic adjustment range;

[0159] Monitor the dynamic performance of the tire during the transplanting process in real time, including comprehensively tracking the displacement, attitude change, tilting angle and sliding trend of the tire. These data will be carefully recorded and compared with the previously established behavior prediction model to check whether the actual behavior meets the expectations. This real-time comparison can help the robot quickly identify potential problems, such as the tilting or sliding of the tire caused by improper grasping.

[0160] At the same time, the force sensors and tactile sensors installed on the gripper play a crucial role in this process. These sensors collect the clamping force information in real time to ensure that the applied grasping force is always maintained within the predetermined range. Through these sensors, the robot can accurately calculate the instantaneous changes of the normal force (the force perpendicular to the contact surface) and the tangential force (the force along the contact surface direction), which is crucial for maintaining the stability of the tire and avoiding sliding. When the force applied by the gripper is too large or too small, the robot can immediately adjust its operation to avoid damaging the tire or losing control.

[0161] In the system monitoring data and clamping feedback data, if there are signs indicating that the weight loss signal is triggered (for example, the tire slips or tilts), the system will dynamically adjust the range according to the pre-generated grasping strategy. At this time, the robot will recalculate the grasping force and maximize the normal force applied by the gripper based on the force field distribution map to ensure that the tire can be firmly grasped. In this way, the robot can not only quickly respond to emergencies, but also effectively adjust the grasping strategy to ensure the stability and safety of the tire.

[0162] When implementing the dynamically adjusted grasping strategy, the robot performs closed-loop control by combining the monitoring data and the feedback data. This means that during the entire grasping process, the system will continuously compare the real-time data with the preset target and adjust the grasping action at any time to ensure the accuracy and safety of the operation. This closed-loop control mechanism enables the robot to have higher flexibility and adaptability in complex environments and can respond to potential unexpected situations in real time.

[0163] Real-time dynamic monitoring can ensure the robot's immediate feedback on the tire status. Such immediacy enables the system to quickly adjust operations in case of any abnormalities, thereby improving the success rate and safety of grasping. The combination of clamping feedback data and real-time monitoring forms an adaptive control mechanism, reducing errors caused by improper human operation and avoiding losses and waste. In addition, the ability to dynamically adjust the grasping strategy enables the robot to flexibly respond in complex and uncertain environments, greatly enhancing the extensiveness and practicality of its application scenarios.

[0164] As Figure 6 shown, during the tire transplant process, the dynamic performance of the tire is captured in real time, and the weight loss of the tire is sensed based on the clamping feedback data. If the weight loss signal is triggered, the grasping force is recalculated within the dynamic adjustment range;

[0165] S410, Real-time monitor the dynamic performance of the tire during the transplant process, including the displacement, attitude change, tilt angle, and slip trend of the tire, and compare the monitoring data with the prediction results of the behavior prediction model;

[0166] S420, The force sensors and tactile sensors installed on the gripper collect the clamping force information in real time to ensure that the grasping force is within the predetermined range, and calculate the immediate changes in the normal force and tangential force according to the feedback data;

[0167] S430, Combine the real-time monitoring data and the clamping feedback data to determine whether the weight loss signal is triggered, including:

[0168] ;

[0169] S440, If the weight loss signal is triggered, then according to the pre-generated dynamic adjustment range of the grasping strategy, recalculate the grasping force, and maximize the normal force applied by the gripper based on the force field distribution map ;

[0170] ;

[0171] ;

[0172] wherein, is the current tangential force, is the additional safety margin;

[0173] S450, Real-time execute the dynamically adjusted grasping strategy, and at the same time, perform closed-loop control on the grasping action by combining the monitoring data and the feedback data.

[0174] S500, Scan the status of the target placement area, predict the stability of the tire after placement through the behavior prediction model, and based on the stability prediction result, adjust the tire angle during the tire transplant process;

[0175] This step first extracts the target placement area from the 3D map. By deeply analyzing the surface characteristics of this area, it is divided into a suitable placement area and a high-risk area. This zoning is completed based on the comprehensive calculation of the surface flatness, tilt angle, and obstacle interference degree of the area. At the same time, the system will mark a priority score for each sub-area to evaluate whether it is suitable as the final placement position of the tire. The priority score reflects the placement safety and feasibility of the area, providing a key basis for the generation of subsequent placement strategies.

[0176] Next, the robot combines the geometric characteristics of the target placement area and the force field distribution map of the tire. By calculating the stability index after the tire is placed, it further optimizes the placement position and angle. The stability index is calculated based on physical models and mathematical methods and is used to quantify the balance and force conditions of the tire after placement. For example, the robot will consider the influence of the tilt angle of the target placement area on the center of gravity of the tire and predict whether the tire is likely to slip or tilt through simulation analysis. If the stability index is lower than the preset index threshold, the system will automatically recalculate the placement angle and specific position of the tire and dynamically adjust the direction and attitude of the tire during the transplant process to ensure ideal stability after placement.

[0177] During the actual tire transplant process, the robot will adjust the jaw angle and attitude in real time to make the tire gradually approach the updated placement strategy. Through monitoring and control, the robot can adjust the operation path according to the real-time state of the target placement area to ensure that the tire is accurately placed in the optimal position while keeping the angle and direction of the tire in line with the stability requirements. This real-time adjustment ability is particularly important because in a dynamic environment, the original plan may be affected by external interference and become inapplicable, and the robot's strategy correction based on real-time data can effectively solve this problem.

[0178] After the tire is successfully placed, the robot uses a force sensor and a vision system to finally confirm the state of the tire. This step ensures that the tire is indeed in a stable state, such as no tilting, slipping, or uneven force phenomenon. At the same time, the system will record the state data of the placement area and the final placement parameters of the tire. These data can be used for subsequent task optimization and model improvement and provide a reference for the next operation.

[0179] Through the surface analysis and priority scoring of the target placement area, the robot can efficiently select the most suitable position for placing the tire stack, avoiding the time waste and operation risks that may be brought by blind attempts. The calculation of stability indicators and the introduction of dynamic adjustment strategies enable the robot to be more adaptable in the face of complex and changing external environments. It can not only ensure the success rate of tire placement but also effectively avoid potential risks. In addition, the ability to adjust the angle and posture of the gripper in real time, combined with the closed-loop control mechanism, further improves the accuracy and robustness of the placement task. Finally, the confirmation of the tire state and data recording after placement not only guarantees the completion of the task but also lays a foundation for the optimization of subsequent tasks by accumulating experience data.

[0180] As Figure 7 shown, the state of the target placement area is scanned, the stability after tire placement is predicted through a behavior prediction model, and based on the stability prediction result, the tire angle is adjusted during the tire transplanting process, specifically including:

[0181] S510, Extract the target placement area from the 3D map, partition the surface of the target placement area, divide the area into a suitable placement area and a high-risk area, and at the same time mark the area priority score by calculating the surface flatness, tilt angle, and obstacle interference degree;

[0182] S520, Based on the geometric characteristics of the target placement area and the force field distribution map of the tire, combined with the center of gravity position, calculate the stability index after tire placement ;

[0183] ;

[0184] S530, Set the index threshold. If the stability index is lower than the index threshold, recalculate the placement angle and position of the tire, and adjust the direction of the tire during the transplanting process;

[0185] S540, During the tire transplanting process, adjust the gripper angle and posture of the tire in real time to make the tire conform to the updated placement strategy, ensuring that the final placement position and angle meet the stability requirements;

[0186] S550, After the tire is successfully placed, confirm whether the tire is in a stable state through the force sensor and the vision system, and at the same time record the state data of the placement area and the final placement parameters of the tire.

[0187] In this step, when the stability index is lower than the index threshold, recalculating the placement angle and position of the tire specifically means:

[0188] ;

[0189] ;

[0190] Among them, is the new placement angle, is the placement angle, represents making the stability index reach the maximum placement angle , represents the new placement coordinates, represents the target placement area priority score, represents making the target placement area priority score reach the maximum placement position .

[0191] S600. After the transplanting task is completed, record the tire data and transplanting data of this transplanting.

[0192] It should be noted that in the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0193] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program code.

[0194] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, and the combination of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded computers, or other programmable data processing devices to generate a machine, so that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0195] These computer program instructions can 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 generate a manufactured article including instruction means, and the instruction means implement the functions in Figure 1 one process or multiple processes and / or blocksFigure 1 The functions specified in one or more boxes.

[0196] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide for implementing the steps of the functions specified in one Figure 1 process or more processes and / or boxes Figure 1 or the steps of the functions specified in one or more boxes.

[0197] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concept. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0198] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A machine vision AI wheel robot based on artificial intelligence algorithm, characterized by: The wheel robot comprises: A work area scanning module is used to scan the work area, build a three-dimensional map of the area, identify the area where the tire is located and the target placement area, identify obstacles in the map, and plan a navigation path for the wheel robot; The force field distribution analysis module is used to scan the surface pattern of the target tire, generate a pattern feature map, analyze the pattern type, detect the tire surface information, and record dynamic parameters. Based on the tire shape, material and stacking method, it generates a tire force field distribution map; The grasping strategy adjustment module is used to establish a behavior prediction model for the target tire based on the tire force field distribution map, predict the dynamic performance of the tire under different grasping angles and forces, and generate the dynamic adjustment range of the grasping strategy; The tire dynamic performance capture module is used to capture the dynamic performance of the tire in real time during the tire transplanting process, and sense the weightlessness of the tire based on the clamping feedback data. If the weightlessness signal is triggered, the gripping force is recalculated within the dynamic adjustment range; The target placement area scanning module is used to scan the state of the target placement area, predict the stability of the tire after placement through the behavior prediction model, and adjust the tire angle during the tire transplantation process based on the stability prediction result; The transplanting data recording module is used to record the tire data and transplanting data of the transplanting after the transplanting task is completed.

2. The working method of the machine vision AI wheel robot based on artificial intelligence algorithm as claimed in claim 1, characterized in that: The working method comprises: Scan the working area, build a three-dimensional map of the area, identify the area where the tire is located and the target placement area, identify obstacles in the map, and plan a navigation path for the wheel robot; Scan the surface pattern of the target tire, generate a pattern feature map, analyze the pattern type, detect the tire surface information, and record dynamic parameters. Based on the tire shape, material and stacking method, generate a tire force field distribution map; For the target tire, a behavior prediction model is established based on the tire force field distribution map to predict the dynamic performance of the tire under different gripping angles and forces, and generate the dynamic adjustment range of the gripping strategy; During the tire transplanting process, the dynamic performance of the tire is captured in real time, and the tire is sensed to lose weight based on the clamping feedback data. If the weightlessness signal is triggered, the gripping force is recalculated within the dynamic adjustment range; Scan the status of the target placement area, predict the stability of the tire after placement through the behavior prediction model, and adjust the tire angle during the tire transplantation process based on the stability prediction results; After the transplanting task is completed, record the tire data and transplanting data of this transplanting.

3. The working method of the wheel robot according to claim 2, characterized in that: The scanning of the working area, building a three-dimensional map of the area, identifying the area where the tire is located and the target placement area, identifying obstacles in the map, and planning a navigation path for the wheel robot specifically includes: Collect data on the work area, capture the contour information of the work area, identify the tire stacking position and target placement position, obtain complete spatial data of the work area, and generate a depth map; Based on the generated depth map, the environmental data is converted into point cloud data, and the multi-view data is integrated through point cloud stitching technology. The pose calibration and overlapping area alignment technology are used to build a three-dimensional map containing static and dynamic objects. Mark the tire stacking position and target placement area in the 3D map, extract the geometric features of the tires, including size, position and relative stacking relationship, and record the boundary information of the target placement area; Analyze obstacles in the 3D map and generate the shortest obstacle avoidance path based on the tire stacking position and target placement area combined with the marked obstacle range.

4. The working method of the wheel robot according to claim 3, characterized in that: The method of scanning the surface pattern of the target tire, generating a pattern feature map, analyzing the pattern type, detecting the tire surface information, and recording dynamic parameters, and generating a tire force field distribution map based on the tire shape, material and stacking method, specifically includes: Scan the target tire to capture the tire's surface pattern, material properties, geometric features, and texture features, and obtain the tire's image information and depth information; Preprocess the acquired tire image information and depth information, extract tire pattern features from the preprocessed image, generate a pattern feature map, analyze the pattern type, depth and distribution, and identify the tire wear status in combination with the tire type; Based on the geometric features, material properties and pattern characteristics of the tire, a force field distribution model of the tire is constructed, including the tire force area, friction force distribution and center of gravity position; By combining the tire's image information and depth information, tire wear status, tire geometric characteristics, material properties and pattern characteristics and the force field distribution model, a complete tire force field distribution map is generated.

5. The working method of the wheel robot according to claim 4, characterized in that: Based on the tire force field distribution map, a behavior prediction model is established to predict the dynamic performance of the tire under different gripping angles and forces, and to generate a dynamic adjustment range of the gripping strategy, specifically including: Based on the generated tire force field distribution map, a behavior prediction model is constructed, and the geometric characteristics and force field distribution data of the tire are input to predict the dynamic performance of the tire under changes in the gripping point, angle and force, including the slip angle, tilt trend and stability after gripping; Based on the behavior prediction model, a grasping strategy for the target tire is generated, including grasping point position, grasping force range and grasping angle range.

6. The working method of the wheel robot according to claim 5, characterized in that: The predicted dynamic performance of the tire under changes in gripping point, angle and force, including slip angle, tilt trend and stability after gripping, is specifically: Calculate the slip angle: ； in, represents the slip angle, which is the angle at which the tire slides after grabbing. is the tangential force, i.e. the force parallel to the contact surface, is the normal force, i.e. the force perpendicular to the contact surface; Calculate the slope trend: ； in, is the change in tilt angle, that is, the degree of tilt of the tire during the gripping process. is the tire weight, is the center of gravity offset distance, calculated from the force field distribution diagram, is the moment of inertia of the tire; Calculate the stability after grabbing: ； in, To obtain the stability index, When , it means the grasping is in a stable state. , indicating that the crawling failed. is the friction coefficient, which is calculated from the tire material properties and texture characteristics.

7. The working method of the wheel robot according to claim 6, characterized in that: The generation of the grabbing strategy for the target tire includes grabbing points, grabbing force range and grabbing angle range, specifically: Calculate the comprehensive score of each candidate grab point: ； in, To grab the point The overall rating of is the weight factor; Determine the grip strength scope: The friction force generated by the clamping force is greater than the tangential force plus a safety margin ,Right now: ； The gripping force does not exceed the maximum load allowed by the tire ,Right now: ； Get the final grip strength scope: ； Determine the grip angle scope: ； in: ； ； and Indicates the minimum and maximum values ​​of the grab angle, Indicates the allowed tilt trend threshold, Indicates the grab angle is Grasping stability.

8. The working method of the wheel robot according to claim 5, characterized in that: The method captures the dynamic performance of the tire in real time during the tire transplanting process, and senses the weightlessness of the tire based on the clamping feedback data. If a weightlessness signal is triggered, the gripping force is recalculated within the dynamic adjustment range, specifically including: Real-time monitoring of the dynamic performance of the tire during the transplanting process, including tire displacement, posture change, tilt angle and slip trend, and comparing the monitoring data with the prediction results of the behavior prediction model; The force sensor and tactile sensor installed on the gripper collect gripping force information in real time to ensure that the gripping force is within the predetermined range, and calculate the instant changes in normal force and tangential force based on the feedback data; Combine real-time monitoring data and clamping feedback data to determine whether to trigger a weightlessness signal, including: ； If the weightlessness signal is triggered, the range is dynamically adjusted according to the pre-generated grasping strategy, the grasping force is recalculated, and the normal force applied by the gripper is maximized based on the force field distribution map. ; ； ； in, is the current tangential force, For additional safety margin; The dynamically adjusted grasping strategy is executed in real time, and the grasping action is closed-loop controlled by combining monitoring data and feedback data.

9. The working method of the wheel robot according to claim 8, characterized in that: The scanning target placement area state, predicting the stability of the tire after placement through a behavior prediction model, and adjusting the tire angle during the tire transplantation process based on the stability prediction result, specifically includes: Extract the target placement area from the 3D map and partition the surface of the target placement area into suitable placement areas and high-risk areas. At the same time, mark the area priority score by calculating the surface flatness, tilt angle and obstacle interference degree. Based on the geometric characteristics of the target placement area and the force field distribution map of the tire, combined with the center of gravity position, the stability index of the tire after placement is calculated ; ； in, is the friction coefficient, which is calculated from the tire material properties and texture characteristics. is the tilt angle change, is the normal force, i.e. the force perpendicular to the contact surface, is the tangential force, i.e. the force parallel to the contact surface; Set the index threshold. If the stability index is lower than the index threshold, recalculate the placement angle and position of the tire and adjust the direction of the tire during the transplanting process. During tire transplantation, the tire gripper angle and posture are adjusted in real time to make the tire conform to the updated placement strategy and ensure that the final placement position and angle meet the stability requirements; After the tire is successfully placed, the force sensor and vision system confirm whether the tire is in a stable state, and record the status data of the placement area and the final placement parameters of the tire.

10. The working method of the wheel robot according to claim 9, characterized in that: If the stability index is lower than the index threshold, the placement angle and position of the tire are recalculated, specifically: ； ； in, For the new placement angle, is the placement angle, Indicates the stability index Maximum placement angle , Represents the new placement coordinates, represents the target placement area priority score, Indicates the priority score of the target placement area The placement position that reaches the maximum value .

Images