Artificial Intelligence-Based Automotive Windshield Transfer Robot and Method

Abstract

The present invention relates to the technical field of program-controlled robots, specifically to an automotive windshield transfer robot and method based on artificial intelligence. The system includes a surface characteristic analysis module, a dynamic environment perception module, a grasping area screening module, a path and motion generation module, and an action execution and correction module. In the present invention, by accurately extracting the boundary coordinates, curvature distribution values, and light intensity offset values on the glass surface, the problem of characteristic recognition of transparent objects under complex optical interference is solved, the accuracy of surface data extraction is improved, the motion data of obstacles is collected in real time in a dynamic environment, the interference area and intensity are quickly recognized, the grasping point selection comprehensively considers the contact angle and force uniformity parameters, enhancing the stability and reliability of the grasping action. The path planning optimizes the low-interference path and adjusts the time, improving the task execution efficiency and motion coordination in a dynamic environment. The action execution is based on pose correction and force feedback adjustment, significantly improving the operation accuracy and adaptability.

Description

Technical Field

[0001] The present invention relates to the technical field of program-controlled robots, and particularly to an automotive windshield transfer robot and method based on artificial intelligence. Background Art

[0002] The technical field of program-controlled robots includes robot systems for achieving automated operations through program control. The core content of this technical field is to achieve high-precision operations of robots in industrial manufacturing, assembly, handling and other scenarios through programmed instructions, mainly involving aspects such as robot structure design, motion control, path planning and task execution. The overall scope of program-controlled robot technology covers multi-axis motion control technology, the design of end effectors and their functional extensions, as well as the development and optimization of control programs, enabling robots to complete specific tasks according to preset programs or real-time input data. In the field of automotive manufacturing, program-controlled robots are widely used in operations such as welding, painting, handling and assembly, with the characteristics of high repeatability and high precision.

[0003] Among them, an automotive windshield transfer robot based on artificial intelligence refers to using artificial intelligence technology to optimize the task execution process of the robot in the handling and assembly of automotive windshields. The technical matters covered by this patent theme include robot motion control and path optimization, precise grasping and placement of specific positions of the windshield, a real-time vision recognition system for glass positioning and feature recognition, and a real-time adjustment algorithm based on neural networks. Specifically, artificial intelligence algorithms are used to analyze sensor data to determine the spatial position of the glass, and combined with path planning algorithms to generate the robot motion trajectory. In addition, a special fixture design is used to ensure the stability and safety of the windshield during the transfer process, comprehensively realizing the combination of the robot and artificial intelligence technology to complete this specific operation task.

[0004] The prior art is greatly affected by optical interference in the characteristic recognition of transparent objects, and it is difficult to accurately extract surface features under reflection and refraction interference, resulting in deviations in characteristic data. In dynamic environment perception, relying on single-parameter analysis of the movement trend of obstacles, it is difficult to accurately predict dynamic changes in complex environments, increasing the risk of path planning failure or obstacle avoidance delay. The selection of grasping points lacks a comprehensive evaluation of the contact angle and force uniformity, and it is easy to have problems such as uneven distribution of grasping points or unstable grasping actions. In path planning, the ability to respond to real-time changing environments is insufficient, and the path rhythm cannot be flexibly adjusted, resulting in reduced efficiency of the robot in dynamic environments. In action execution, the ability to process path deviation correction and real-time force feedback is limited, and it is easy to have problems such as insufficient action accuracy or failed target grasping, restricting the operation effect in complex environments. Summary of the Invention

[0005] The object of the present invention is to solve the disadvantages existing in the prior art, and a windshield transfer robot and method based on artificial intelligence are proposed.

[0006] To achieve the above object, the present invention adopts the following technical solutions: The windshield transfer robot based on artificial intelligence includes:

[0007] The surface characteristic analysis module extracts the coordinates of the surface boundary contour points, the curvature distribution value, and the light intensity offset value of the transparent area based on the windshield sampling data, superimposes and calculates the light intensity offset value and the curvature distribution value, analyzes the curvature change trend, identifies the curvature peak value and the coordinates of the stable area, and generates the glass surface characteristic distribution data;

[0008] The dynamic environment perception module analyzes the speed value and the direction change rate based on the dynamically collected obstacle position coordinates, speed values, and motion direction change rates, calculates the obstacle movement area range, cross-analyzes the spatial position of the area and the glass surface characteristic distribution data, and generates the environmental interference distribution value;

[0009] The grasping area screening module extracts the boundary curvature peak points and the stable area coordinates based on the glass surface characteristic distribution data and the environmental interference distribution value, screens the clamping coordinate points, and generates the preferred grasping point coordinate data;

[0010] The path and action generation module extracts the spatial position sequence and time sequence of the path nodes based on the preferred grasping point coordinate data, accumulates the distances of the node paths in the spatial position sequence, and compares them point by point with the environmental interference distribution value, adjusts the time difference between the path nodes, and generates the optimized grasping path sequence data;

[0011] The action execution and correction module extracts the pose parameters of the path nodes and the contact force feedback value of the fixture based on the optimized grasping path sequence data, calculates and generates the pose correction amount, analyzes the error between the fixture contact force feedback value and the grasping point force trend value, identifies the contact force adjustment amount, and outputs the execution action calibration parameters.

[0012] As a further solution of the present invention, the glass surface characteristic distribution data includes boundary contour characteristics, curvature change characteristics, light intensity offset characteristics, and optical interference characteristics, the environmental interference distribution value includes interference area coordinates, interference intensity data, motion range distribution, and overlap range distribution, the preferred grasping point coordinate data includes grasping point coordinates, force uniformity characteristics, contact angle distribution, and interference rejection results, the optimized grasping path sequence data includes path node position sequence, time distribution sequence, low-interference path sequence, and time interval adjustment value, and the execution action calibration parameters include pose correction parameters, force adjustment parameters, path correction parameters, and grasping point force adjustment parameters.

[0013] As a further solution of the present invention, the surface characteristic analysis module includes a data extraction sub-module, an optical interference calculation sub-module, and a characteristic distribution generation sub-module;

[0014] Based on the windshield sampling data, the data extraction sub-module extracts the coordinates of the surface boundary contour points, the light intensity offset value and the curvature distribution value of the transparent area, classifies and screens according to the characteristics of the transparent area, identifies the distribution rules of the boundary data and the transparent area, and generates a surface feature data set;

[0015] Based on the surface feature data set, the optical interference calculation sub-module superimposes the light intensity offset value and the curvature distribution value of the transparent area, calculates the regional optical interference intensity, classifies the calculation results into segments to form a characteristic distribution range, analyzes the variation relationship between the optical interference and the surface geometric features, marks the differential distribution of the regional interference intensity, and generates optical interference characteristic data;

[0016] Based on the optical interference characteristic data, the characteristic distribution generation sub-module analyzes the change trend of the curvature distribution in segments, extracts the characteristic key points by summarizing the curvature change trend, maps the key points point by point to the glass surface coordinate system, marks the coordinate positions of the curvature peaks and stable regions, and generates the glass surface feature distribution data.

[0017] As a further solution of the present invention, the dynamic environment perception module includes a motion parameter extraction sub-module, a dynamic prediction range generation sub-module, and a space interference analysis sub-module;

[0018] Based on the dynamically collected obstacle position coordinates, speed values and motion direction change rates, the motion parameter extraction sub-module extracts the real-time position and motion state data of the obstacles, calculates the correlation between the speed value and the direction change rate, extracts the position coordinates and motion state data, and generates obstacle motion parameter data;

[0019] Based on the obstacle motion parameter data, the dynamic prediction range generation sub-module combines the speed value and the direction change rate to judge the motion trend of the obstacle, calculates the influence value of the trend change on the spatial range, gradually predicts the obstacle motion area, spatially marks the prediction area to generate a trend change range, and generates obstacle dynamic prediction area data;

[0020] Based on the obstacle dynamic prediction area data, the space interference analysis sub-module combines the glass surface feature distribution data, compares the spatial distribution of the prediction area and the surface features point by point, calculates the interference intensity value in the intersection range, marks the spatial distribution and intensity data of the interference area, and generates the environmental interference distribution value.

[0021] As a further solution of the present invention, the grasping area screening module includes a feature point extraction sub-module, an interference elimination sub-module, and an optimal point generation sub-module;

[0022] Based on the glass surface feature distribution data, the feature point extraction sub-module extracts the boundary curvature peak points and the coordinates of the stable regions, analyzes the curvature change values and the coordinate distribution characteristics of the boundary region point by point, screens the points with significant curvature changes as the curvature peak points, and at the same time identifies the range of the region where the curvature change tends to be stable, generating the boundary key feature point data;

[0023] Based on the boundary key feature point data and the environmental interference distribution value, the interference rejection sub-module compares the spatial coordinates of each feature point with the interference region range point by point, calculates the interference intensity value of the region where the feature point is located, rejects the feature points higher than the interference threshold, and re-marks the coordinates of the remaining feature points and the interference region range to which they belong, generating the low-interference feature point data;

[0024] Based on the low-interference feature point data, the preferred point generation sub-module calculates the clamping contact angle and the force uniformity of the feature points, sorts them according to the numerical values of the contact angle and the force uniformity parameters, screens the feature points that meet the clamping conditions, and generates the preferred grasping point coordinate data.

[0025] As a further solution of the present invention, the calculation formula of the interference intensity value is specifically:

[0026] ;

[0027] Among them, represents the interference intensity value of the feature point , represents the number of interference regions associated with the feature point , represents the interference intensity value of the feature point within the interference region , represents the area of the interference region , represents the distance between the feature point and the center point of the interference region .

[0028] As a further solution of the present invention, the path and motion generation module includes a path node extraction sub-module, a path optimization sub-module, and a time adjustment sub-module;

[0029] Based on the preferred grasping point coordinate data, the path node extraction sub-module extracts the spatial position sequence and the time sequence of the path nodes, analyzes the distribution relationship between the nodes, calculates the connection attributes between adjacent nodes, generates a path framework point by point according to the spatial connectivity of the path nodes, accumulates the distances between the nodes to form the initial information of the path nodes, and generates the initial path node sequence;

[0030] Based on the initial path node sequence and the environmental interference distribution value, the path optimization sub-module compares the spatial coordinates of path nodes point by point with the intersection of the interference regions, calculates the interference intensity value of the nodes in the interference regions, eliminates the nodes with interference intensity exceeding the threshold, reconnects the nodes, establishes an interference-optimized path, labels the adjusted path nodes, and generates low-interference path data;

[0031] Based on the low-interference path data, the time adjustment sub-module extracts the time sequence of path nodes, analyzes the time intervals between adjacent nodes in the time sequence, adjusts the intervals and optimizes the movement rhythm between nodes, corrects the time distribution, verifies the consistency between the time sequence and the spatial path, and generates optimized grasping path sequence data.

[0032] As a further solution of the present invention, the action execution and correction module includes a pose parameter correction sub-module, a contact force adjustment sub-module, and an action calibration generation sub-module;

[0033] Based on the optimized grasping path sequence data, the pose parameter correction sub-module extracts the pose parameters of path nodes, compares the pose of path nodes with the expected pose of the grasping points point by point, calculates the offset amount of path nodes in multiple directions, analyzes the directionality of the offset and the cumulative offset amount, and generates pose correction data;

[0034] Based on the pose correction data and the contact force feedback value of the fixture, the contact force adjustment sub-module compares the feedback value with the force trend value of the grasping point point by point, calculates the contact force error value, identifies its force direction, analyzes the change trend of the feedback force and the spatial relationship between the forces at the grasping point, extracts the adjustment range of the force deviation, and labels the correction amplitude to generate contact force adjustment data;

[0035] Based on the contact force adjustment data, the action calibration generation sub-module superimposes the contact force adjustment value and the pose correction value point by point, calculates the adjusted action state parameters, analyzes the state differences between the adjustment data and the path nodes item by item, converts the differences into action amplitude correction values, calibrates the action output of the path nodes, and generates execution action calibration parameters.

[0036] As a further solution of the present invention, the specific formula for calculating the interference cumulative value is:

[0037] ;

[0038] Wherein, represents the interference cumulative value of the path segment, represents the number of interference regions included in the path segment, represents the interference intensity value of the nodes in the path segment within the interference region ; represents the interference region ; Represents the distance from the node within the path segment to the center point of the interference region.

[0039] An artificial intelligence-based method for a windshield transfer robot of a vehicle, comprising the following steps:

[0040] S1: Based on the windshield sampling data, extract the coordinates of the surface boundary contour points, the curvature distribution values, and the light intensity offset values of the transparent region, analyze the curvature change trend, identify the curvature peak values and the coordinates of the stable regions, and generate the glass surface feature distribution data;

[0041] S2: Based on the dynamically collected obstacle position coordinates, speed values, and the change rate of the movement direction, calculate the range of the obstacle movement region, cross-analyze the spatial positions of the intersection region and the glass surface feature distribution data, and generate the environmental interference distribution values;

[0042] S3: Based on the glass surface feature distribution data and the environmental interference distribution values, extract the boundary curvature peak points and the coordinates of the stable regions, and generate the preferred grasping point coordinate data;

[0043] S4: Based on the preferred grasping point coordinate data, extract the spatial position sequence and the time sequence of the path nodes, accumulate the distances of the node paths in the spatial position sequence, and compare them point by point with the environmental interference distribution values to generate the optimized grasping path sequence data;

[0044] S5: Based on the optimized grasping path sequence data, extract the pose parameters of the path nodes and the contact force feedback values of the fixture, analyze the error between the fixture contact force feedback values and the force trend values at the grasping points, identify the contact force adjustment amount, and output the execution action calibration parameters.

[0045] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0046] In the present invention, by accurately extracting the boundary coordinates, curvature distribution values, and light intensity offset values of the glass surface, the problem of characteristic recognition of transparent objects under complex optical interference is solved, the accuracy of surface data extraction is improved, the obstacle movement data is collected in real time in a dynamic environment, the interference region and intensity are quickly recognized, the grasping points are selected by comprehensively considering the contact angle and force uniformity parameters, the stability and reliability of the grasping action are enhanced, the path planning is optimized through low-interference paths and time adjustment, the task execution efficiency and motion coordination in a dynamic environment are improved, and the action execution is based on pose correction and force feedback adjustment, significantly improving the operation accuracy and adaptability. Description of the Drawings

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0048] Figure 1 It is the system flow chart of the present invention;

[0049] Figure 2 It is the sub-module flow chart of the present invention;

[0050] Figure 3 It is the surface characteristic analysis module flow chart of the present invention;

[0051] Figure 4 It is the dynamic environment perception module flow chart of the present invention;

[0052] Figure 5 It is the grasping area screening module flow chart of the present invention;

[0053] Figure 6 It is the path and motion generation module flow chart of the present invention;

[0054] Figure 7 It is the motion execution and correction module flow chart of the present invention;

[0055] Figure 8 It is the method step flow chart of the present invention. Specific embodiments

[0056] The following will describe the technical solutions in the present invention in conjunction with the accompanying drawings.

[0057] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as "example" in the present invention should not be construed as more preferred or more advantageous than other embodiments or design solutions. Exactly, the use of the word "example" is intended to present concepts in a specific manner. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two.

[0058] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meanings they express are the same. "(of)", "corresponding", and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meanings they express are the same.

[0059] In the embodiments of the present invention, sometimes subscripts such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, their intended meanings are the same.

[0060] To make the technical problems, technical solutions, and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0061] Please refer to Figure 1 and Figure 2 , the windshield transfer robot based on artificial intelligence includes:

[0062] The surface characteristic analysis module extracts the coordinates of the surface boundary contour points, the curvature distribution value, and the light intensity offset value of the transparent area based on the windshield sampling data, superimposes and calculates the light intensity offset value and the curvature distribution value, analyzes the curvature change trend, identifies the curvature peak value and the coordinates of the stable area, and generates the glass surface characteristic distribution data;

[0063] The dynamic environment perception module analyzes the speed value and the direction change rate based on the dynamically collected obstacle position coordinates, speed value, and motion direction change rate, calculates the range of the obstacle movement area, cross-analyzes the spatial position of the area and the glass surface characteristic distribution data, and generates the environmental interference distribution value;

[0064] The grasping area screening module extracts the boundary curvature peak points and the coordinates of the stable area based on the glass surface characteristic distribution data and the environmental interference distribution value, screens the clamping coordinate points, and generates the preferred grasping point coordinate data;

[0065] The path and action generation module extracts the spatial position sequence and time sequence of the path nodes based on the preferred grasping point coordinate data, accumulates the distances of the node paths in the spatial position sequence, and compares them point by point with the environmental interference distribution value, adjusts the time difference between the path nodes, and generates the optimized grasping path sequence data;

[0066] The action execution and correction module extracts the pose parameters of the path nodes and the contact force feedback value of the fixture based on the optimized grasping path sequence data, calculates and generates the pose correction amount, analyzes the error between the fixture contact force feedback value and the stress trend value of the grasping point, identifies the contact force adjustment amount, and outputs the execution action calibration parameters.

[0067] The glass surface feature distribution data includes boundary contour features, curvature change characteristics, light intensity offset characteristics, and optical interference characteristics. The environmental interference distribution values include interference region coordinates, interference intensity data, motion range distribution, and overlap range distribution. The preferred grasping point coordinate data includes grasping point coordinates, force uniformity characteristics, contact angle distribution, and interference rejection results. The optimized grasping path sequence data includes path node position sequences, time distribution sequences, low-interference path sequences, and time interval adjustment values. The execution action calibration parameters include pose correction parameters, force adjustment parameters, path correction parameters, and grasping point force adjustment parameters.

[0068] Please refer to Figure 2 and Figure 3 , the surface feature analysis module includes a data extraction sub-module, an optical interference calculation sub-module, and a feature distribution generation sub-module;

[0069] Based on the windshield sampling data, the data extraction sub-module extracts the coordinates of the surface boundary contour points, the light intensity offset values and curvature distribution values of the transparent region, classifies and filters according to the characteristics of the transparent region, identifies the distribution rules of the boundary data and the transparent region, and generates a surface feature data set.

[0070] First, the coordinates of the surface boundary contour points are sampled point by point through a high-precision boundary recognition device, and the sampling error values are removed using a data denoising algorithm. At the same time, the change range of the light intensity offset values in the transparent region is recorded, and the curvature distribution value of the transparent region is calculated in real time in combination with a curvature analysis device. The classification and screening rules are constructed using the change curve of the curvature. The classification process is based on the comparison of the transparent characteristics of adjacent regions to determine whether the region belongs to the transparent boundary range, and then the distribution rules of the boundary data and the transparent region are identified. A surface feature data set is generated in combination with the recognition results of the transparent region.

[0071] Based on the surface feature data set, the optical interference calculation sub-module superimposes the light intensity offset value and the curvature distribution value of the transparent region, calculates the optical interference intensity of the region, classifies the calculation results into segments to form a characteristic distribution range, analyzes the change relationship between the optical interference and the surface geometric characteristics, marks the differential distribution of the interference intensity in the region, and generates optical interference characteristic data;

[0072] Superimpose the light intensity offset value and the curvature distribution value of the transparent region, according to the formula

[0073] ;

[0074] In the formula, represents the optical interference intensity, represents the regional curvature distribution value, represents the light intensity distribution value, represents the partial derivative of the light intensity distribution value with respect to the spatial position, and is the integration interval.

[0075] In the calculation of the optical interference intensity of the region, the curvature distribution value is measured by a high-precision curvature analysis device, and the light intensity distribution value obtains the brightness values at different spatial positions through a sampling device. The partial derivative is calculated by dividing the difference between sampling points by the difference in spatial positions. Assuming that in a region, the measured curvature value is 0.05 and the light intensity distribution satisfies a linear distribution in the interval [0, 10] , the calculation process is as follows:

[0076] , representing that the partial derivative of the light intensity with respect to the position is a constant.

[0077] Substitute into the formula for calculation:

[0078] ;

[0079] The result shows that the calculated optical interference intensity is 0.1, representing the interference effect value under the combined action of the transparent characteristics and geometric characteristics of the region, and its value reflects the optical characteristic differences of the region.

[0080] Based on the optical interference characteristic data, the characteristic distribution generation sub-module analyzes the change trend of the curvature distribution in segments, extracts the characteristic key points by summarizing the curvature change trend, maps the key points point by point to the glass surface coordinate system, marks the coordinate positions of the curvature peaks and stable regions, and generates the glass surface characteristic distribution data;

[0081] Through the segmented curvature analysis of the transparent region, the key points of the curvature change are extracted by using the curvature change rate calculation model. First, collect the curvature data in the transparent region and record the change points, analyze in segments according to the curvature change trend, extract the points with a change rate greater than a certain threshold as the characteristic key points, perform coordinate transformation on the extracted characteristic key points and map them to the glass surface coordinate system, analyze the distribution of the curvature peak points and stable points, and generate the glass surface characteristic distribution result in combination with the mapped characteristic distribution data.

[0082] Please refer to Figure 2 and Figure 4 , the dynamic environment perception module includes a motion parameter extraction sub-module, a dynamic prediction range generation sub-module, and a spatial interference analysis sub-module;

[0083] Based on the dynamically collected obstacle position coordinates, speed values, and motion direction change rates, the motion parameter extraction sub-module extracts the real-time position and motion state data of the obstacles, calculates the correlation between the speed value and the direction change rate, extracts the position coordinates and motion state data, and generates the obstacle motion parameter data;

[0084] Calculate the speed value by continuous position sampling and record the change rate of the movement direction. Combine the position coordinate data, calculate the movement trajectory by calculating the Euclidean distance between continuous position points, extract the real-time position data and movement state data through trajectory segmentation analysis, calculate the correlation degree between the speed value and the change rate of the direction by comparing the change trends between them, use the correlation coefficient method for the calculation of the correlation degree, further extract the movement parameters of the obstacle and record them in the dataset, and finally generate the obstacle movement parameter data.

[0085] Based on the obstacle movement parameter data, the dynamic prediction range generation sub-module combines the speed value and the change rate of the direction to judge the movement trend of the obstacle, calculates the influence value of the trend change on the spatial range, gradually predicts the movement area of the obstacle, performs spatial annotation on the predicted area to generate the trend change range, and generates the obstacle dynamic prediction area data;

[0086] Combine the speed value and the change rate of the direction to judge the movement trend of the obstacle, according to the formula

[0087] ;

[0088] Calculate the influence value of the trend change on the spatial range.

[0089] In the formula, represents the influence value of the trend change on the spatial range, represents the speed value of the obstacle, represents the derivative of the change rate of the direction with respect to time, and are the time integration intervals.

[0090] For the calculation of the influence value of the movement trend change on the spatial range, the speed value is obtained by collecting through the obstacle real-time monitoring device, the change rate of the direction is obtained by calculating the change of the movement direction angle of the obstacle within the time interval, and the derivative is obtained by taking the difference of the change rate of the direction over time. Assume the speed value is 3m / s, and the change rate of the direction satisfies a linear change within the interval [0,2] seconds , the calculation process is as follows:

[0091] Derivative of the change rate of the direction with respect to time: .

[0092] Substitute into the formula for calculation:

[0093] ;

[0094] The results show that the influence value of the trend change on the spatial range is 3, which represents the degree of interference of the change in the movement trend of the obstacle on the future prediction area, and its value can be used for the calculation of spatial range prediction.

[0095] Based on the data of the dynamic prediction area of the obstacle and combined with the data of the glass surface feature distribution, the spatial interference analysis sub-module compares the spatial distribution of the prediction area and the surface features point by point, calculates the interference intensity value in the intersection range, marks the spatial distribution and intensity data of the interference area, and generates the environmental interference distribution value;

[0096] By combining the dynamically generated trend change range and the glass surface feature distribution data, using the point-by-point spatial coordinate matching technology for comparison, calculate the spatial intersection range for each point in the obstacle prediction area and the glass surface feature data point. The calculation of the spatial intersection range judges the overlapping area through the distance threshold, calculates the interference intensity value in the intersection range, calculates the total interference intensity by gradually accumulating the intensity values in the interference area, and marks the coordinates of the interference area in the spatial distribution map. Combine the interference intensity value and the distribution coordinates to generate the environmental interference distribution value.

[0097] Please refer to Figure 2 and Figure 5 , the grasping area screening module includes a feature point extraction sub-module, an interference elimination sub-module, and an optimal point generation sub-module;

[0098] Based on the glass surface feature distribution data, the feature point extraction sub-module extracts the boundary curvature peak points and the coordinates of the stable area, analyzes the curvature change value and the coordinate distribution characteristics of the boundary area point by point, screens the points with significant curvature changes as the curvature peak points, and at the same time identifies the range of the area where the curvature change tends to be stable, and generates the boundary key feature point data;

[0099] Collect the glass surface curvature distribution value through a high-precision sensor, calculate the curvature change value using the differential calculation method, screen the points with significant curvature changes as the curvature peak points, use the sliding window technology to analyze the continuity characteristics of the curvature change, extract the range of the area where the curvature change tends to be stable, and calibrate the feature point coordinates through the corresponding relationship between the boundary curvature value and the position coordinates. Form the above extracted content into a boundary key feature point data set.

[0100] Based on the boundary key feature point data and the environmental interference distribution value, the interference elimination sub-module compares the spatial coordinates of each feature point with the interference area range point by point, calculates the interference intensity value of the area where the feature point is located, eliminates the feature points higher than the interference threshold, and re-marks the coordinates of the remaining feature points and the interference area range to which they belong, and generates the low-interference feature point data;

[0101] The specific calculation formula of the interference intensity value is:

[0102] ;

[0103] Among them, represents the interference intensity value of the feature point . represents the number of interference regions associated with the feature point . represents the interference intensity value of the feature point within the interference region . represents the area of the interference region . represents the distance between the feature point and the center point of the interference region .

[0104] The calculation of the feature point interference correction coefficient involves the following steps and parameters:

[0105] 1. represents the interference intensity value of the feature point within the interference region . It is calculated by measuring the optical or mechanical influence value of the interference region. The measurement method uses an interference intensity sensor to integrate the intersection range of the feature point and the region to obtain a numerical value;

[0106] 2. represents the area of the interference region . It is calculated using the coordinate data of the boundary points of the interference region through the geometric area formula, and the area unit is square meters;

[0107] 3. represents the distance from the feature point to the center point of the interference region . It is calculated using the Euclidean distance formula between two points;

[0108] 4. represents the number of interference regions related to the feature point , which is determined by the intersection of the feature point and the interference region;

[0109] 5. The summation operation is used to accumulate the influence of all interference regions on the feature point, taking the absolute value to ensure that all influences are positive values, and the square root processing is used to smooth the result.

[0110] Example:

[0111] The feature point is located at the coordinates (2, 3). There are 3 interference regions, and the center point coordinates are (4, 4), (6, 7), and (3, 5) respectively. Their areas are 5 square meters, 8 square meters, and 6 square meters. The interference intensity value Measured values are 0.8, 1.2, and 0.5. The calculations are as follows:

[0112] Calculate the distance between each interference region and the feature point:

[0113] ;

[0114] ;

[0115] ;

[0116] Calculate the influence value of each interference region on the feature point:

[0117] ;

[0118] ;

[0119] ;

[0120] Sum up and calculate the interference correction coefficient:

[0121] ;

[0122] The results show that the interference intensity value of the feature point is 1.268, representing the average interference correction value of the feature point within the range of all interference regions. The value is used to further eliminate high-interference points and optimize the distribution of interference regions.

[0123] Based on the low-interference feature point data, the preferred point generation sub-module calculates the clamping contact angle and force uniformity of the feature points, sorts them according to the numerical values of the contact angle and force uniformity parameters, screens the feature points that meet the clamping conditions, and generates the coordinate data of the preferred grasping points;

[0124] Measure the clamping contact angle and force distribution of the feature points through a high-precision force sensor. The clamping contact angle is calculated using the angle distribution mean method, and the force uniformity is calculated by the standard deviation of the force distribution of the feature points. Generate a sorted list of feature points based on the calculation results, screen the feature points with the clamping angle within the target range and good force uniformity, and output the coordinates of the screened feature points as the coordinate data of the preferred grasping points.

[0125] Please refer to Figure 2 and Figure 6 , the path and motion generation module includes a path node extraction sub-module, a path optimization sub-module, and a time adjustment sub-module;

[0126] Based on the preferred grasping point coordinate data, the path node extraction sub-module extracts the spatial position sequence and time sequence of path nodes, analyzes the distribution relationship between nodes, calculates the connection attributes between adjacent nodes, generates a path framework point by point according to the spatial connectivity of path nodes, accumulates the distances between nodes to form the initial information of path nodes, and generates an initial path node sequence;

[0127] By extracting the spatial position sequence and time sequence of path nodes point by point, using a spatial analysis tool to calculate the Euclidean distance between adjacent nodes, calculating the total path length by distance accumulation, further determining the spatial connectivity of adjacent nodes, generating a path framework point by point using geometric topology analysis, and simultaneously performing synchronous analysis on the time sequence of path nodes, recording the node time interval and sequence distribution, generating the initial information of path nodes through comprehensive analysis of the position and time between nodes, and finally outputting the initial path node sequence.

[0128] Based on the initial path node sequence and the environmental interference distribution value, the path optimization sub-module compares the spatial coordinates of path nodes point by point with the intersection of the interference region, calculates the interference intensity value of the nodes in the interference region, removes the nodes with interference intensity exceeding the threshold, reconnects the nodes, establishes an interference-optimized path, marks the adjusted path nodes, and generates low-interference path data;

[0129] The specific formula for the interference accumulation value is:

[0130] ;

[0131] Among them, represents the interference accumulation value of the path segment, represents the number of interference regions included in the path segment, represents the interference intensity value of the nodes in the interference region of the path segment, represents the interference region area, represents the distance from the nodes in the path segment to the center point of the interference region.

[0132] The interference accumulation value of the path segment represents the cumulative interference effect of all interference regions on the path nodes in the path segment. The calculation involves the following parameters:

[0133] represents the interference intensity value of the nodes in the interference region of the path segment, which is obtained by measuring with an interference intensity sensor, and the unit is Newton or the corresponding interference intensity unit.

[0134] represents the interference region The area is calculated by measuring the boundary coordinate points of the interference region and using the geometric area formula, with the unit of square meters.

[0135] Represents the distance from the node within the path segment to the center point of the interference region, calculated using the Euclidean distance formula, with the unit of meters. The distance from the node within the path segment to the center point of the interference region is calculated using the Euclidean distance formula, with the unit of meters.

[0136] Represents the total number of interference regions within the path segment, determined from the spatial intersection data of the path segment and the interference region.

[0137] The summation operation accumulates the interference contribution values for each interference region, takes the absolute value to ensure the positivity of the value, and the inverse square is used for weighting the influence of distance on interference.

[0138] Example:

[0139] The path segment contains 3 interference regions, and the corresponding parameters are as follows:

[0140] Interference region 1: , , the distance from the path node to the center point .

[0141] Interference region 2: , , the distance from the path node to the center point .

[0142] Interference region 3: , , the distance from the path node to the center point .

[0143] Calculate the contribution value of each interference region:

[0144] The contribution value of interference region 1:

[0145] ;

[0146] The contribution value of interference region 2:

[0147] ;

[0148] The contribution value of interference region 3:

[0149] ;

[0150] Accumulate the contribution values of all interference regions:

[0151] ;

[0152] This result indicates that the interference cumulative value of the path segment is 2.65, representing the total interference intensity contribution value of the nodes within the path segment to the interference area. This value is used to determine whether the path nodes exceed the interference threshold and perform path optimization processing.

[0153] The time adjustment sub-module extracts the time series of path nodes based on the low-interference path data, analyzes the time intervals between adjacent nodes in the time series, adjusts the intervals and optimizes the movement rhythm between nodes, corrects the time distribution, verifies the consistency between the time series and the spatial path, and generates optimized grasping path sequence data;

[0154] The time interval analysis model is used to optimize the node time intervals. The adjustment of the time intervals is completed by analyzing the time distribution law of adjacent nodes and optimizing the movement rhythm. After correction, the time interval distribution is consistent with the spatial path. By verifying the adjustment effect of the time series point by point, combined with the spatial path distribution, optimized grasping path sequence data is regenerated, and the corresponding relationship between the final time and spatial nodes is marked.

[0155] Please refer to Figure 2 and Figure 7 , the action execution and correction module includes a pose parameter correction sub-module, a contact force adjustment sub-module, and an action calibration generation sub-module;

[0156] The pose parameter correction sub-module extracts the pose parameters of path nodes based on the optimized grasping path sequence data, compares the pose of path nodes with the expected pose of the grasping point point by point, calculates the offset of path nodes in multiple directions, analyzes the directionality of the offset and the cumulative offset, and generates pose correction data;

[0157] The high-precision pose sensing device is used to collect the pose information of path nodes point by point, obtain the angular and position offsets of the nodes in multiple directions, calculate the magnitude and direction of the offset by comparing the pose parameters of the path nodes with the expected pose of the grasping point, analyze the offset trend in multiple directions and the cumulative offset, and generate the final pose correction data in combination with the spatial relationship between nodes.

[0158] The contact force adjustment sub-module, based on the pose correction data and the contact force feedback value of the fixture, compares the feedback value with the force trend value of the grasping point point by point, calculates the contact force error value, identifies its force direction, analyzes the change trend of the feedback force and the spatial relationship of the force at the grasping point, extracts the adjustment range of the force deviation, and marks the correction amplitude to generate contact force adjustment data;

[0159] Compare the feedback value with the force trend value of the grasping point point by point, according to the formula

[0160] ;

[0161] Calculate the contact force error value.

[0162] In the formula, represents the adjustment value of the contact force, represents the contact force distribution of feedback, Indicates the force trend value of the grasping point, is the integral range of the force-bearing area.

[0163] The contact force adjustment value is obtained by calculating the difference between the feedback contact force and the expected contact force point by point and accumulating them. The feedback contact force is recorded by the real-time contact force sensor, and the expected contact force is given by the force distribution model of the gripping point. Assume that the feedback contact force function is , the expected contact force function is , the integral range is [0,4]×[0,4]. The calculation process is as follows:

[0164] Difference function:

[0165] ;

[0166] Adjustment value calculation:

[0167] ;

[0168] Decomposition of the integral:

[0169] ;

[0170] Item-by-item integration:

[0171] ;

[0172] The results show that the contact force adjustment value is 13.6, which represents the overall error adjustment range between the feedback force and the desired force, and is used for the subsequent optimization of the contact force parameters.

[0173] Based on the contact force adjustment data, the motion calibration generation submodule superimposes the contact force adjustment value and the posture correction value point by point, calculates the adjusted motion state parameters, analyzes the state difference between the adjustment data and the path node item by item, and converts the difference into the motion amplitude correction value, calibrates the motion output of the path node, and generates the execution motion calibration parameters;

[0174] The corrected action state parameters are calculated by point-by-point superposition, and the difference between the adjustment data and the current state of the path node is further analyzed. The correction value of the action amplitude is calculated through difference analysis. The action amplitude correction value is obtained by the difference between the adjustment data and the historical state recorded in real time by the sensor. The corrected amplitude value re-marks the action output parameters in units of path nodes, and finally generates the calibrated execution action calibration parameter data.

[0175] See also Figure 8 , an artificial intelligence-based automobile windshield transfer robot method, comprising the following steps:

[0176] S1: Based on the windshield sampling data, extract the coordinates of the surface boundary contour points, the curvature distribution values, and the light intensity offset values of the transparent area, analyze the curvature change trend, identify the curvature peak values and the coordinates of the stable regions, and generate the glass surface feature distribution data;

[0177] S2: Based on the coordinates of the obstacle positions, the speed values, and the change rate of the movement directions collected dynamically, calculate the range of the obstacle movement area, cross-analyze the spatial positions of the intersection area and the glass surface feature distribution data, and generate the environmental interference distribution values;

[0178] S3: Based on the glass surface feature distribution data and the environmental interference distribution values, extract the boundary curvature peak points and the coordinates of the stable regions, and generate the preferred grasping point coordinate data;

[0179] S4: Based on the preferred grasping point coordinate data, extract the spatial position sequence and the time sequence of the path nodes, accumulate the distances of the node paths in the spatial position sequence, and compare them point by point with the environmental interference distribution values to generate the optimized grasping path sequence data;

[0180] S5: Based on the optimized grasping path sequence data, extract the pose parameters of the path nodes and the contact force feedback values of the fixture, analyze the error between the fixture contact force feedback values and the stress trend values of the grasping points, identify the contact force adjustment amount, and output the execution action calibration parameters.

[0181] As mentioned above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, and all of them should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. An artificial intelligence-based windshield transfer robot for vehicles, characterized in that: The system includes: The surface characteristic analysis module extracts the coordinates of the surface boundary contour points, the curvature distribution values, and the light intensity offset values of the transparent region based on the windshield sampling data, superimposes and calculates the light intensity offset values and the curvature distribution values, analyzes the curvature change trend, identifies the curvature peak and the coordinates of the stable region, and generates the glass surface characteristic distribution data. The dynamic environment perception module analyzes the speed value and the direction change rate based on the dynamically collected obstacle position coordinates, speed values, and motion direction change rates, calculates the range of the obstacle movement area, cross-analyzes the spatial position of the area and the glass surface characteristic distribution data, and generates the environmental interference distribution value. The grasping area screening module extracts the boundary curvature peak points and the coordinates of the stable region based on the glass surface characteristic distribution data and the environmental interference distribution value, screens the clamping coordinate points, and generates the preferred grasping point coordinate data. The path and motion generation module extracts the spatial position sequence and the time sequence of the path nodes based on the preferred grasping point coordinate data, accumulates the distances of the node paths in the spatial position sequence, and compares them point by point with the environmental interference distribution value, adjusts the time difference between the path nodes, and generates the optimized grasping path sequence data. The motion execution and correction module extracts the pose parameter of the path node and the contact force feedback value of the fixture based on the optimized grasping path sequence data, calculates and generates the pose correction amount, analyzes the error between the fixture contact force feedback value and the grasping point force trend value, identifies the contact force adjustment amount, and outputs the execution motion calibration parameter. The surface characteristic analysis module includes a data extraction sub-module, an optical interference calculation sub-module, and a characteristic distribution generation sub-module. The data extraction sub-module extracts the coordinates of the surface boundary contour points, the light intensity offset values of the transparent region, and the curvature distribution values based on the windshield sampling data, classifies and screens them according to the characteristics of the transparent region, identifies the distribution rules of the boundary data and the transparent region, and generates the surface characteristic data set. The optical interference calculation sub-module superimposes the light intensity offset value and the curvature distribution value of the transparent region based on the surface characteristic data set, calculates the regional optical interference intensity, classifies the calculation results into characteristic distribution ranges by segments, analyzes the change relationship between the optical interference and the surface geometric characteristics, marks the differential distribution of the regional interference intensity, and generates the optical interference characteristic data. The characteristic distribution generation sub-module analyzes the change trend of the curvature distribution by segments based on the optical interference characteristic data, extracts the characteristic key points by summarizing the curvature change trend, maps the key points point by point to the glass surface coordinate system, marks the coordinate positions of the curvature peak and the stable region, and generates the glass surface characteristic distribution data.

2. The windshield transfer robot for automobiles based on artificial intelligence according to claim 1, characterized in that: The glass surface feature distribution data includes boundary contour features, curvature change characteristics, light intensity offset characteristics, and optical interference characteristics. The environmental interference distribution value includes interference region coordinates, interference intensity data, motion range distribution, and overlap range distribution. The preferred grasping point coordinate data includes grasping point coordinates, force uniformity characteristics, contact angle distribution, and interference rejection results. The optimized grasping path sequence data includes path node position sequences, time distribution sequences, low-interference path sequences, and time interval adjustment values. The execution action calibration parameters include pose correction parameters, force adjustment parameters, path correction parameters, and grasping point force adjustment parameters.

3. The AI-based windshield transfer robot according to claim 1, characterized in that: The dynamic environment perception module includes a motion parameter extraction sub-module, a dynamic prediction range generation sub-module, and a spatial interference analysis sub-module; The motion parameter extraction sub-module extracts the real-time position and motion state data of the obstacle based on the dynamically collected obstacle position coordinates, speed values, and motion direction change rates, calculates the correlation between the speed value and the direction change rate, extracts the position coordinates and motion state data, and generates obstacle motion parameter data; The dynamic prediction range generation sub-module judges the motion trend of the obstacle based on the obstacle motion parameter data, combines the speed value and the direction change rate, calculates the influence value of the trend change on the spatial range, gradually predicts the obstacle motion area, spatially marks the predicted area to generate a trend change range, and generates obstacle dynamic prediction area data; The spatial interference analysis sub-module, based on the obstacle dynamic prediction area data and combined with the glass surface feature distribution data, compares the spatial distribution of the predicted area and the surface features point by point, calculates the interference intensity value in the intersection range, marks the spatial distribution and intensity data of the interference area, and generates an environmental interference distribution value.

4. The robot for transferring automobile windshields based on artificial intelligence according to claim 1, characterized in that: The grasping area screening module includes a feature point extraction sub-module, an interference rejection sub-module, and a preferred point generation sub-module; The feature point extraction sub-module extracts the boundary curvature peak points and stable area coordinates based on the glass surface feature distribution data, analyzes the curvature change value and coordinate distribution characteristics of the boundary area point by point, screens the points with significant curvature changes as curvature peak points, and at the same time identifies the area range where the curvature change tends to be stable, and generates boundary key feature point data; The interference rejection sub-module, based on the boundary key feature point data and the environmental interference distribution value, compares the spatial coordinates of each feature point with the interference region range point by point, calculates the interference intensity value of the area where the feature point is located, eliminates the feature points higher than the interference threshold, and re-marks the coordinates of the remaining feature points and the interference region range to which they belong, and generates low-interference feature point data; The preferred point generation sub-module calculates the clamping contact angle and force uniformity of the feature points based on the low-interference feature point data, sorts them according to the numerical values of the contact angle and force uniformity parameters, screens the feature points that meet the clamping conditions, and generates preferred grasping point coordinate data.

5. The robot for transferring automotive windshields based on artificial intelligence according to claim 4, wherein: The specific formula for calculating the interference intensity value is: ； Among them, represents the interference intensity value of the feature point ; represents the number of interference regions associated with the feature point ; represents the interference intensity value of the feature point within the interference region ; represents the area of the interference region ; represents the distance between the feature point and the center point of the interference region .

6. The robot for transferring automobile windshields based on artificial intelligence according to claim 1, wherein: The path and action generation module includes a path node extraction sub-module, a path optimization sub-module, and a time adjustment sub-module; Based on the preferred grasping point coordinate data, the path node extraction sub-module extracts the spatial position sequence and time sequence of path nodes, analyzes the distribution relationship between nodes, calculates the connection attributes between adjacent nodes, generates a path framework point by point according to the spatial connectivity of path nodes, accumulates the distances between nodes to form the initial information of path nodes, and generates an initial path node sequence; Based on the initial path node sequence and the environmental interference distribution value, the path optimization sub-module compares the spatial coordinates of path nodes point by point with the intersection of the interference area, calculates the interference intensity value of the nodes in the interference area, eliminates the nodes with interference intensity exceeding the threshold, reconnects the nodes, establishes an interference-optimized path, labels the adjusted path nodes, and generates low-interference path data; Based on the low-interference path data, the time adjustment sub-module extracts the time sequence of path nodes, analyzes the time interval between adjacent nodes in the time sequence, adjusts the interval and optimizes the movement rhythm between nodes, corrects the time distribution, verifies the consistency between the time sequence and the spatial path, and generates optimized grasping path sequence data.

7. The robot for transferring the automotive windshield based on artificial intelligence according to claim 6, wherein: The specific formula for calculating the interference accumulation value is: ； Among them, represents the cumulative interference value of the path segment, represents the number of interference regions included in the path segment, represents the interference intensity value of the nodes within the path segment in the interference region and represents the area of the interference region and represents the distance from the nodes within the path segment to the center point of the interference region.

8. The artificial intelligence-based windshield transfer robot for automobiles according to claim 1, wherein: The action execution and correction module includes a pose parameter correction sub-module, a contact force adjustment sub-module, and an action calibration generation sub-module; Based on the optimized grasping path sequence data, the pose parameter correction sub-module extracts the pose parameters of path nodes, compares the pose of path nodes with the expected pose of the grasping point point by point, calculates the offset amount of path nodes in multiple directions, analyzes the directionality of the offset and the cumulative offset amount, and generates pose correction data; Based on the pose correction data and the contact force feedback value of the fixture, the contact force adjustment sub-module compares the feedback value with the force trend value of the grasping point point by point, calculates the contact force error value, identifies its force direction, analyzes the change trend of the feedback force and the spatial relationship of the force on the grasping point, extracts the adjustment range of the force deviation, and marks the correction amplitude, and generates contact force adjustment data; Based on the contact force adjustment data, the action calibration generation sub-module superimposes the contact force adjustment value and the pose correction value point by point, calculates the adjusted action state parameters, analyzes the state differences between the adjustment data and the path nodes item by item, and converts the differences into action amplitude correction values, calibrates the action output of the path nodes, and generates execution action calibration parameters.

9. An artificial intelligence-based method for a windshield transfer robot of an automobile, characterized in that, The execution of the windshield transfer robot based on artificial intelligence according to any one of claims 1-8 includes the following steps: S1: Based on the windshield sampling data, extract the coordinates, curvature distribution values of the surface boundary contour points, and the light intensity offset values of the transparent area, analyze the curvature change trend, identify the curvature peak and the coordinates of the stable area, and generate the glass surface feature distribution data; S2: Based on the dynamically collected obstacle position coordinates, speed values, and the change rate of the movement direction, calculate the obstacle movement area range, cross-analyze the spatial position of the area and the glass surface feature distribution data, and generate the environmental interference distribution value; S3: Based on the glass surface feature distribution data and the environmental interference distribution value, extract the boundary curvature peak points and the coordinates of the stable area, and generate the preferred grasping point coordinate data; S4: Based on the coordinate data of the preferred grasping points, extract the spatial position sequence and time sequence of the path nodes, accumulate the distances of the node paths in the spatial position sequence, and compare them point by point with the environmental interference distribution value to generate optimized grasping path sequence data; S5: Based on the optimized grasping path sequence data, extract the pose parameters of the path nodes and the contact force feedback value of the fixture, analyze the error between the fixture contact force feedback value and the force trend value at the grasping point, identify the contact force adjustment amount, and output the execution action calibration parameters.

Images