Automobile spoiler mechanical hand polishing system and polishing method

Abstract

The application relates to the field of polishing, in particular to an automobile spoiler mechanical hand polishing system and a polishing method, which comprise a path planning module, a state monitoring module, a parameter compensation module and a waste collecting module connected in communication, wherein: the path planning module is used for dynamically adjusting a polishing path by real-time scanning of a spoiler through a visual sensor; the visual sensor is a 3D structured light sensor or a stereo visual camera; the state monitoring module is used for real-time monitoring of the vibration state of a polishing head through a vibration sensor and acquiring vibration abnormal data; the parameter compensation module is used for automatically adjusting polishing parameters according to the vibration abnormal data and adjusting the polishing head to polish; and the waste collecting module is used for centrally collecting and processing dust and waste generated in the polishing process. The application can accurately plan a polishing path, improve polishing quality, improve production efficiency, reduce cost and effectively process dust and waste.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of polishing, and particularly to a polishing system and method for a mechanical hand for polishing an automobile spoiler. BACKGROUND

[0002] With the development of the automobile industry, the production quality of automobile spoilers, which are important components for improving the performance and aesthetics of automobiles, is increasingly improved. In the manufacturing process of automobile spoilers, polishing is one of the key processes. In the prior art, the polishing path of the mechanical hand is not accurate enough, and is often preset, which cannot be adjusted adaptively according to the actual shape of the spoiler, and is prone to uneven polishing, missing or over-polishing. Moreover, in the polishing process, the vibration state of the polishing head cannot be effectively monitored, and when abnormal vibration occurs, such as uneven material of the spoiler, hard points and other problems, the polishing parameters cannot be adjusted in time, which not only affects the polishing quality, but also may damage the polishing tool and the spoiler. In addition, the collection and treatment of waste and dust are not perfect, which may cause environmental pollution and affect the working environment and the health of the operators. SUMMARY

[0003] To solve the above problems, the present application provides a polishing system and method for a mechanical hand for polishing an automobile spoiler, which can accurately plan the polishing path to improve the polishing quality, improve the production efficiency, reduce the cost and effectively treat the dust and waste.

[0004] To achieve the above-mentioned purposes, the technical solution adopted by the present application is as follows:

[0005] A polishing system for a mechanical hand for polishing an automobile spoiler, comprising a path planning module, a state monitoring module, a parameter compensation module and a waste collection module connected in communication, wherein:

[0006] The path planning module is used to dynamically adjust the polishing path by real-time scanning of the spoiler through a visual sensor; the visual sensor is a 3D structured light sensor or a stereo vision camera;

[0007] The state monitoring module is used to monitor the vibration state of the polishing head in real time through a vibration sensor and obtain vibration abnormal data;

[0008] The parameter compensation module is used to automatically adjust the polishing parameters according to the vibration abnormal data and adjust the polishing head for polishing;

[0009] The waste collection module is used to collect and treat the dust and waste generated during polishing.

[0010] Further, the real-time scanning of the spoiler through the visual sensor and the dynamic adjustment of the polishing path comprise the following steps:

[0011] The visual sensor is fixedly connected above or on the side of the mechanical hand through a support and a damping device;

[0012] The data acquisition computer and the robot control system are connected through a bidirectional communication link;

[0013] The polishing path is path planned and dynamically adjusted through the path planning model;

[0014] In the polishing process, the data acquisition computer and the monitoring program in the robot control system continuously monitor the visual sensor parameters, and an alarm is issued and the polishing is paused when the sensor parameters exceed the set threshold.

[0015] Further, the path planning model is realized through A* algorithm and feature-based model matching algorithm, wherein:

[0016] The A* algorithm is used for searching from the starting point to the target point of the path planning, generating the optimal path according to the polishing cost, and providing the reference path for subsequent dynamic adjustment;

[0017] The feature-based model matching algorithm is used for quantifying the deviation between the tail fin model and the initial model after the real-time and initial model features are aligned, and the decision basis for dynamically adjusting the polishing path.

[0018] Further, the A* algorithm comprises the following steps:

[0019] The visual sensor is started to scan the tail fin placed on the workbench and fixed in all directions, and a three-dimensional reconstruction algorithm based on stereo vision is used to reconstruct three-dimensional point cloud model of the tail fin from each frame of image;

[0020] The three-dimensional space of the tail fin is divided into uniform grids, the starting position of the polishing tool is taken as the starting point, and the target position set of the tail fin surface to be polished is taken as the target point set, and the A* algorithm searches the path by maintaining the open list and the closed list;

[0021] In each iteration, the node with the minimum cost function value is selected from the open list to determine the optimal path, and when the target point is added to the closed list, the algorithm ends, the initial polishing path is obtained by backtracking from the parent node of the target point to the starting point, and is taken as the input of the feature-based model matching algorithm.

[0022] Further, the feature-based model matching algorithm comprises the following steps:

[0023] In the polishing process, the visual sensor continuously scans the tail fin at a set frequency, converts the real-time image frame into a real-time tail fin three-dimensional point cloud model through a three-dimensional reconstruction algorithm, and extracts features from the corresponding model of the initial polishing path and the real-time model;

[0024] The corner point features are extracted by a three-dimensional Harris corner point detection algorithm, and the edge features are obtained by calculating the gradient direction change of the point cloud. The extracted feature point sets are an initial model feature point set and a real-time model feature point set;

[0025] The distance between each feature point in the real-time model feature point set and the nearest feature point in the initial model feature point set is calculated by a nearest neighbor matching algorithm. A transformation relationship including a translation vector, a rotation matrix and a scaling factor is found by iterative optimization to minimize the sum of distances between all matching feature points.

[0026] According to the transformation relationship obtained by model matching, the deviation between the real-time tail surface and the initial polishing path is analyzed, and the rotation angle deviation and the scaling factor deviation are analyzed. It is judged that the polishing path needs to be adjusted.

[0027] According to the path adjustment decision, the polishing path is modified, and the adjusted polishing path is fed back to the manipulator control system again.

[0028] Further, the vibration state of the polishing head is monitored in real time by the vibration sensor, and vibration abnormal data is obtained, including the following steps:

[0029] A vibration sensor is connected between the polishing head and the manipulator. The vibration sensor is connected to a data acquisition card by a wired method. The data acquisition card continuously acquires the output signal of the vibration sensor during polishing and performs signal preprocessing.

[0030] The direct current component of the vibration signal is obtained by calculating the signal mean value. The signal dispersion is obtained by calculating the standard deviation of the signal. The existence of impact vibration is judged by extracting the peak factor of the signal. The signal spectrum is obtained by performing fast Fourier transform on the signal, and the frequency spectrum amplitude is calculated. The main frequency component of the vibration is determined by the frequency point with large amplitude.

[0031] Threshold values are determined for the signal mean value, standard deviation, peak factor and main frequency amplitude according to the abnormal detection algorithm. When multiple features exceed the threshold value at the same time, it is judged as abnormal vibration. The mean value and standard deviation of the distribution are estimated by statistical analysis of a large amount of vibration feature data under normal polishing state. The probability of the newly collected vibration feature data under the assumed distribution is calculated. When it is lower than the probability threshold value, it is judged as abnormal vibration, and the vibration feature data is recorded.

[0032] Further, after judging as abnormal vibration, the support vector machine or multilayer perception machine is used for abnormal vibration detection to identify abnormal vibration under complex mode.

[0033] Further, the polishing parameters are automatically adjusted according to the vibration abnormal data, and the polishing head is adjusted for polishing, including the following steps:

[0034] The abnormal vibration data is fuzzed into fuzzy language variables, and the fuzzy set in each fuzzy language variable corresponds to a membership function; the fuzzed vibration features are associated with the polishing parameter adjustment amount according to the fuzzy rule base; the fuzzed polishing parameter adjustment amount is obtained by reasoning in the rule base through a fuzzy reasoning algorithm according to the current fuzzed vibration features; and the fuzzed adjustment amount is converted into an accurate parameter adjustment value;

[0035] The calculated polishing parameter adjustment value is sent to the manipulator control system to perform parameter adjustment operation, and the vibration of the polishing head is continuously monitored after parameter adjustment to determine whether it returns to normal, if it is still abnormal, the above steps are repeated to adjust the polishing parameter again, and machine learning is used for auxiliary adjustment.

[0036] Further, the dust and waste generated in the polishing process are collected and treated, including the following steps:

[0037] A dust suction port is arranged below the tail wing, and each dust suction port is connected to a central dust collecting device through a pipeline;

[0038] The dust suction system is started before polishing, the air volume and air speed of the dust suction system are controlled by adjusting the rotating speed of the fan and the opening degree of the valve, a pressure difference sensor is installed in the pipeline system, whether the pipeline is blocked is judged by monitoring the pressure difference between the two ends of the pipeline, and flow balance control is performed;

[0039] The waste is separated and classified through the cyclone separation technology, the running parameters of the treatment equipment are monitored during the waste treatment process, and the waste information is recorded to establish a waste treatment file.

[0040] A polishing method of a manipulator for an automobile tail wing, including the following steps:

[0041] The tail wing is scanned in real time by a visual sensor, and the polishing path is dynamically adjusted;

[0042] The vibration state of the polishing head is monitored in real time by a vibration sensor, and vibration abnormal data is obtained;

[0043] The polishing parameter is automatically adjusted according to the vibration abnormal data, and the polishing head is polished;

[0044] The dust and waste generated in the polishing process are collected and treated.

[0045] The beneficial effects of the present application are:

[0046] 1. By combining visual sensors with A* algorithm and feature-based model matching algorithm, the polishing path can be accurately planned and dynamically adjusted according to the actual shape of the tail wing. Whether it is a complex curved surface, edge or different geometric shape, it can ensure that the polishing tool covers every part that needs to be processed, avoiding the problem of omission and uneven polishing, making the tail wing surface quality higher, meeting the high-quality production standard. The vibration sensor cooperates with the advanced abnormal detection algorithm to timely detect the abnormal vibration of the polishing head. Through the parameter adjustment algorithm based on fuzzy logic, the polishing parameters are automatically adjusted. This effectively reduces the polishing defects caused by abnormal vibration, further improving the polishing quality.

[0047] 2. Dynamic polishing path planning does not require frequent human intervention. During the polishing process, the system can automatically adjust the path according to the shape change of the tail wing. Compared with the traditional preset fixed path polishing method, it reduces the time for re-polishing or adjustment due to poor polishing effect, improves the overall polishing efficiency. The timely vibration abnormality processing mechanism ensures the stability of the polishing process, reduces the number and time of downtime caused by polishing head damage, tail wing damage and other faults, so that the polishing work can be continuously and efficiently carried out.

[0048] 3. Precise polishing path and stable polishing process avoid over-polishing, reduce waste of tail wing materials, and effectively reduce raw material costs. By effectively controlling the polishing parameters, damage to the polishing head and manipulator caused by abnormal vibration is avoided, reducing equipment maintenance and replacement costs. The perfect waste collection system can efficiently collect dust and waste generated during the polishing process, which not only reduces environmental pollution and protects the health of operators, but also meets environmental protection regulations, avoiding potential costs and risks due to environmental problems. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 is a schematic diagram of the modules of the automobile tail wing manipulator polishing system.

[0050] Figure 2 is a flowchart of the automobile tail wing manipulator polishing method. DETAILED DESCRIPTION

[0051] Please refer to Figure 1 and Figure 2 , the present application relates to an automobile tail wing manipulator polishing system, comprising a communication connected path planning module, a state monitoring module, a parameter compensation module and a waste collection module, wherein:

[0052] The path planning module is used to scan the tail wing in real time through the visual sensor to dynamically adjust the polishing path; the visual sensor is a 3D structured light sensor or a stereo vision camera;

[0053] The state monitoring module is used for monitoring the vibration state of the polishing head in real time through the vibration sensor, and obtaining vibration abnormal data;

[0054] The parameter compensation module is used for automatically adjusting the polishing parameters according to the vibration abnormal data, and adjusting the polishing head to polish;

[0055] The waste collecting module is used for collecting and processing the dust and waste generated in the polishing process.

[0056] Further, the tail fin is scanned in real time through the visual sensor, and the polishing path is dynamically adjusted, including the following steps:

[0057] The visual sensor is fixedly connected above or on the side of the manipulator through the support and the damping device;

[0058] The data acquisition computer and the manipulator control system are connected through the bidirectional communication link;

[0059] The polishing path is planned and dynamically adjusted through the path planning model;

[0060] In the polishing process, the monitoring program in the data acquisition computer and the manipulator control system continuously monitors the visual sensor parameters, and when the sensor parameters exceed the set threshold, an alarm is issued and the polishing is paused.

[0061] Specifically, the 3D structured light sensor obtains three-dimensional information by projecting a specific light pattern onto the object surface and analyzing the reflected light, and has high precision and resolution; the stereo vision camera reconstructs a three-dimensional model by shooting an object from different angles through two or more cameras based on the triangulation principle. The installation of the visual sensor needs to ensure that its field of view is not blocked, and it needs to be firmly installed to reduce the influence of vibration and displacement on image acquisition. In the polishing process, the visual sensor collects images of the tail fin surface at a set frequency. For the 3D structured light sensor, images containing light pattern information are collected; for the stereo vision camera, multiple view images are collected at the same time. The collected images are preprocessed, including denoising, contrast enhancement and other operations, such as removing random noise in the image through Gaussian filtering, enhancing the contrast of the image through histogram equalization, and improving the image quality for subsequent processing.

[0062] The data acquisition computer is a computer device specially used for data acquisition work. It has the function of collecting various data, can obtain information from different data sources, and can perform operations such as storage, processing and analysis on these data. The mechanical hand control system refers to a system for controlling the operation of the mechanical hand. This system realizes accurate control of various actions of the mechanical hand, such as grabbing, moving, placing and other operations. Connect the data acquisition computer and the mechanical hand control system to establish a bidirectional communication link, so as to feed back the data processing results of the visual sensor to the control system, and realize real-time adjustment of the polishing path.

[0063] During the entire polishing process, by setting monitoring programs in the data acquisition computer and the control system, the working temperature of the visual sensor, the packet loss rate of data transmission, the running time of the algorithm and other parameters are continuously monitored. If the sensor temperature is too high (exceeding the set temperature threshold), an alarm is issued and the polishing is suspended, and the sensor cooling condition is checked; if the packet loss rate is higher than the set threshold, the data transmission line is checked; if the algorithm running time exceeds a certain proportion of the scanning interval time, it may cause real-time problems, and the algorithm needs to be optimized or the hardware needs to be upgraded.

[0064] Further, the path planning model is realized by A* algorithm and feature-based model matching algorithm, wherein:

[0065] The A* algorithm is used for searching from the starting point to the target point of the path planning, generating an optimal path according to the polishing cost, and providing a reference path for subsequent dynamic adjustment;

[0066] The feature-based model matching algorithm is used to align the features of the real-time and initial models, quantify the deviation between the tail wing model and the initial model, and dynamically adjust the decision basis of the polishing path.

[0067] Specifically, in the application scenario of polishing the car spoiler, the main task of the A* algorithm is to search for the path from the starting position of the polishing tool (the starting point) to all target positions (such as the edges of the spoiler, different curved surface areas, etc.) that need to be polished in the three-dimensional space of the spoiler. Due to the complex shape of the spoiler with multiple curved surfaces and irregular geometric shapes, the A* algorithm can effectively explore the path in such a three-dimensional space. The A* algorithm defines a cost function to measure the "cost" of passing through each grid point. This cost function takes into account various factors, such as the distance from the grid point to the polishing starting point, whether the grid point is on the surface of the spoiler, and the curvature change near the grid point, etc. By minimizing the total cost of the path, the initial polishing path generated by the A* algorithm is to a certain extent optimal, which can enable the polishing tool to traverse the spoiler surface in a relatively efficient and reasonable manner. The initial polishing path output by the A* algorithm serves as the input for the subsequent feature-based model matching algorithm and is the basis for the entire dynamic polishing path adjustment process. It provides a reference for the expected polishing path, and through comparison and adjustment with this benchmark path, real-time optimization of the polishing path is achieved when the shape of the spoiler surface changes during polishing.

[0068] During polishing, the real-time three-dimensional model of the spoiler is obtained through the vision sensor. The first thing the feature-based model matching algorithm needs to do is to align the features of the real-time model and the model corresponding to the initial polishing path generated by the A* algorithm. By accurately finding the correspondence of feature points in the two models, the actual position change of the spoiler surface during polishing can be determined. After completing the feature alignment, the algorithm can quantify the deviation between the real-time spoiler model and the initial model. By calculating these deviations, the regions of the spoiler surface that have undergone shape changes and the degree of change can be accurately located. This is crucial for determining the specific location and adjustment range of the polishing path that needs to be adjusted. If the displacement of the feature points in a certain region exceeds the set threshold or the rotation angle changes significantly, it indicates that the polishing of that region does not meet the expectations, which may be due to uneven spoiler material, polishing tool wear, or other factors. Based on this deviation information, the system can adjust the polishing path accordingly to ensure the uniformity and completeness of polishing, so that the quality of the polished spoiler surface meets the requirements.

[0069] Further, the A* algorithm comprises the following steps:

[0070] Starting the vision sensor to perform a full-range scan on the spoiler placed on the workbench and fixed, and obtaining a three-dimensional point cloud model of the spoiler through three-dimensional reconstruction of each frame of image based on a stereo vision three-dimensional reconstruction algorithm;

[0071] The three-dimensional space of the tail fin is divided into uniform grids, and the start position of the polishing tool is taken as the starting point, and the target position set of the tail fin surface to be polished is taken as the target point set. The A* algorithm searches the path by maintaining the open list and the closed list.

[0072] In each iteration, the node with the minimum cost function value is selected from the open list to determine the optimal path. When the target point is added to the closed list, the algorithm ends. The initial polishing path is obtained by backtracking from the target point to the parent node of the start point, and is taken as the input of the feature-based model matching algorithm.

[0073] Specifically, the visual sensor is started to perform omnidirectional scanning on the tail fin placed on the workbench and fixed. The scanning time is T, and the number of image frames collected in this time is N=fT. For each image frame, a three-dimensional reconstruction algorithm based on stereo vision is used for three-dimensional reconstruction. Specifically, the optical center coordinates of the two cameras are O1(x1, y1, z1) and O2(x2, y2, z2), respectively. For any point P(x, y) in the image, its coordinates (X, Y, Z) in three-dimensional space can be calculated by the triangulation formula:

[0074]

[0075] where C x , C y are the principal point coordinates of the camera, f is the focal length of the camera, and Z is calculated by matching the corresponding points in the images of the two cameras using the parallax principle. By processing the points in all image frames, a three-dimensional point cloud model M0 of the tail fin is obtained.

[0076] The three-dimensional space of the tail fin is divided into uniform grids, and the grid edge length is a. For each grid point G(i, j, k) (where i, j, k are the indices of the grid point in three-dimensional space), the cost function C(G) is defined as C(G)=ω1d(G)+ω2s(G)+ω3c(G); where d(G) is the Euclidean distance from the grid point G to the polishing start point, and the calculation formula is: where (x0, y0, z0) is the polishing start point coordinate, (x G , y G , z G ) is the coordinate of the grid point G; s(G) indicates whether the grid point is on the tail fin surface. If it is on the tail fin surface, s(G)=0, otherwise s(G)=1; c(G) is the curvature change near the grid point, which can be approximated by calculating the second derivative of the point cloud model near the point, and ω1, ω2, ω3 are weight coefficients determined according to experience.

[0077] The start position of the polishing tool is taken as the starting point S, and the target position set of the tail fin surface to be polished is taken as the target point set {Tn The A* algorithm searches the path by maintaining two lists: the open list and the closed list. In each iteration, the node n with the minimum cost function value is selected from the open list, and the formula is F(n) = G(n) + H(n), where G(n) is the actual cost from the starting point to node n (i.e., the sum of the cost function values of all grid points passed), and H(n) is the heuristic estimated cost from node n to the target point. The Euclidean distance is used as the heuristic function, that is:

[0078]

[0079] When the target point is added to the closed list, the algorithm ends, and the initial polishing path P0 is obtained by backtracking from the parent node of the target point to the starting point, which is represented as a sequence of grid points P0 = (G1, G2, …, G m ), which is used as the input of the feature-based model matching algorithm.

[0080] Further, the feature-based model matching algorithm comprises the following steps:

[0081] During the polishing process, the visual sensor continuously scans the tail fin at a set frequency, and converts the real-time image frames into a real-time tail fin three-dimensional point cloud model through a three-dimensional reconstruction algorithm, and extracts features from the corresponding model of the initial polishing path and the real-time model;

[0082] The corner point features are extracted through a three-dimensional Harris corner point detection algorithm, and the edge features are obtained by calculating the gradient direction change of the point cloud, and the extracted feature point sets are the initial model feature point set and the real-time model feature point set;

[0083] Through a nearest neighbor matching algorithm, the distance between each feature point in the real-time model feature point set and the nearest feature point in the initial model feature point set is calculated, and a transformation relationship that minimizes the sum of distances between all matched feature points is found through iterative optimization, and the transformation relationship includes a translation vector, a rotation matrix and a scaling factor;

[0084] According to the transformation relationship obtained through model matching, the deviation between the real-time tail fin surface and the initial polishing path is analyzed, and the rotation angle deviation and the scaling factor deviation are analyzed, and the polishing path that needs to be adjusted is determined;

[0085] According to the path adjustment decision, the polishing path is modified, and the adjusted polishing path is fed back to the robot control system again.

[0086] Specifically, during the polishing process, the visual sensor continuously scans the tail fin at a set frequency, and converts the real-time image frames into a real-time tail fin three-dimensional point cloud model M t. The features are extracted from the model M0 corresponding to the initial polishing path P0 and the real-time model M t For corner feature extraction, the three-dimensional Harris corner detection algorithm is used. Let N(p) be the neighborhood point set of point p(x, y, z) in the three-dimensional point cloud model, and H(p) be the autocorrelation matrix of p:

[0087]

[0088] where I x , I y , I z are the gradients of the point cloud in the x, y, and z directions, and ω is a weighting function (such as a Gaussian function). The eigenvalues λ1, λ2, and λ3 of H(p) are calculated, and if λ1> λ th and λ2> λ th , where λ th is a set threshold, then point p is a corner feature. For edge features, the gradient direction change of the point cloud is calculated, and when the gradient direction change exceeds a certain threshold, it is considered to be an edge point. The extracted feature point sets are F0 (initial model feature point set) and F t (real-time model feature point set).

[0089] The nearest neighbor matching algorithm is used in the feature-based model matching algorithm. For each feature point F t in the real-time model feature point set F t，i , find the nearest feature point f 0，j in the initial model feature point set F0, and calculate the distance d(f t，i , f 0，j ) between them. Through iterative optimization, find the transformation relationship that minimizes the sum of distances between all matched feature point pairs, which includes the translation vector , the rotation matrix R, and the scaling factor s. The transformation parameters can be solved by minimizing the objective function as follows:

[0090]

[0091] According to the transformation relationship obtained by model matching, analyze the deviation between the real-time tail surface and the initial polishing path. Calculate the average displacement of the feature points where n is the number of matched feature point pairs, and if where d th is a set displacement threshold, then it is considered that there is a significant deviation. At the same time, analyze the rotation angle deviation and the scaling factor deviation to determine which part of the polishing path needs to be adjusted. For example, if the feature point displacement and rotation angle deviation in a local area are large, it indicates that the polishing is uneven or the tail shape has changed, and the polishing path in that area needs to be adjusted.

[0092] According to the path adjustment decision, the polishing path is modified. Let the points on the original polishing path be P0=(G1, G2,..., G m ), and for the points G i in the region that needs to be adjusted, the adjusted position G′ i is calculated according to the deviation. If there is a translation deviation, if there is a rotation deviation, G′ i = RG i ; if there is a scaling deviation, G′ i = sG i . The adjusted polishing path P t =(G′1, G′2,..., G′ m ) is fed back to the manipulator control system again to realize real-time dynamic adjustment of the polishing process and ensure the uniformity and completeness of polishing.

[0093] Further, the vibration state of the polishing head is monitored in real time by the vibration sensor to obtain vibration abnormality data, including the following steps:

[0094] A vibration sensor is connected between the polishing head and the manipulator. The vibration sensor is connected to a data acquisition card in a wired manner. The data acquisition card continuously acquires the output signal of the vibration sensor during polishing and performs signal preprocessing.

[0095] The direct current component of the vibration signal is obtained by calculating the signal mean value. The degree of dispersion of the signal is obtained by calculating the standard deviation of the signal. The existence of impact vibration is judged by extracting the peak factor of the signal. The main frequency component of the vibration is determined by performing fast Fourier transform on the signal to obtain the signal spectrum and calculating the spectrum amplitude.

[0096] Threshold values are determined for the signal mean value, standard deviation, peak factor, and main frequency amplitude according to an abnormality detection algorithm. When multiple features simultaneously exceed the threshold values, it is judged to be abnormal vibration. The mean value and standard deviation of the distribution are estimated by statistically analyzing a large amount of vibration feature data under normal polishing conditions. The probability of the newly collected vibration feature data under the assumed distribution is calculated. When it is lower than the probability threshold value, it is judged to be abnormal vibration, and the vibration feature data is recorded.

[0097] Specifically, the vibration sensor is installed near the polishing head, such as directly on the shell of the polishing head or on the mechanical arm part connecting the polishing head and the robot. Ensure that the sensor can accurately perceive the vibration of the polishing head, and the installation method will not affect the normal work of the polishing head and the movement of the robot. According to the standard vibration source, the sensor is calibrated. A vibration generator with known frequency and amplitude is used to excite vibration at different frequencies and amplitudes within the working frequency range of the sensor, and the output signal of the sensor is recorded. According to the recorded data, a calibration curve or calibration equation is established between the sensor output signal and the actual vibration parameters (frequency, amplitude and phase). Assuming that there is a linear relationship between the sensor output voltage signal V and the vibration frequency f, the amplitude A (the actual situation may be more complex), the calibration equation V=k1f+k2A+b is obtained by least squares fitting, where k1, k2 and b are calibration coefficients.

[0098] Connect the vibration sensor and the data acquisition card in the data acquisition device, and the sampling frequency f s Should be determined according to the highest possible frequency of the polishing head vibration, according to the Nyquist-Shannon sampling theorem, usually choose f s ≥2f max , where f max is the highest frequency of the expected vibration. During the polishing process, the data acquisition system continuously acquires the output signal x(n) of the vibration sensor according to the set sampling frequency, where n represents the sampling point number x(n) is a discrete time series signal. Since the collected vibration signal may contain noise and interference components, such as electromagnetic interference, other vibrations transmitted by the mechanical structure, etc., the signal needs to be filtered. In order to facilitate subsequent data analysis and comparison, the filtered signal is normalized.

[0099] Calculate the mean value of the signal Where N is the length of the signal (number of sampling points), the mean value can reflect the direct current component of the vibration signal, and abnormal direct current offset may indicate abnormal vibration of the polishing head. Calculate the standard deviation of the signal The standard deviation can measure the dispersion of the signal, and a larger standard deviation may indicate the instability of the vibration. Extract the peak factor of the signal Where z peak is the peak value of the signal. The peak factor can be used to detect whether there is an impact vibration in the signal, such as when the polishing head encounters a hard point, a higher peak factor will be generated. Perform a fast Fourier transform on the normalized signal z(n) to obtain the frequency spectrum Z(k) of the signal, where k represents the frequency index. Calculate the amplitude |Z(k)| of the frequency spectrum. Extract the amplitude and frequency of the main frequency component by finding the frequency point with a larger amplitude to determine the main frequency component of the vibration.

[0100] In the anomaly detection algorithm, by threshold judgment method: according to experience or experimental data, set threshold for the extracted features (such as mean, standard deviation, peak factor, main frequency amplitude, etc.). When multiple features exceed their respective thresholds at the same time, the credibility of abnormal judgment is increased. For example, not only the standard deviation exceeds the threshold, but also the peak factor exceeds the normal range, so it can be more determined that the polishing head is in abnormal vibration state. By statistical model method: assume that the characteristic data of normal vibration obeys a certain probability distribution, such as normal distribution. Through statistical analysis of a large number of vibration characteristic data under normal polishing state, the parameters (such as mean and standard deviation) of the distribution are estimated. For newly collected vibration characteristic data, calculate its probability under the assumed distribution. If the probability is lower than the set probability threshold, it is judged as abnormal vibration.

[0101] When abnormal vibration data is detected, record the relevant vibration characteristic data (such as timestamp, characteristic value, frequency spectrum, etc.) for subsequent analysis of the cause of the anomaly. These data can be stored in a local database or file. Trigger the alarm system, such as reminding the worker that the polishing head is in abnormal vibration through sound, light or sending a message to the operator. The alarm information can include the type of anomaly (such as excessive vibration, abnormal frequency, etc.) and the possible location (according to the position information of the polishing head on the tail wing).

[0102] Further, after judging as abnormal vibration, support vector machine or multilayer perception machine is used for abnormal vibration detection to identify abnormal vibration under complex mode.

[0103] Specifically, after the anomaly detection algorithm, a supervised learning algorithm, support vector machine (SVM) or neural network (such as multi-layer perceptron) is used for anomaly vibration detection. First, a large number of feature data under normal and abnormal vibration states are collected as a training set. The training set is labeled (normal is 0, abnormal is 1), and then the model is trained using the training set. In the test phase, the newly extracted vibration feature data is input into the trained model, and the model outputs a probability or classification result indicating whether the vibration is abnormal. For example, SVM divides normal and abnormal data by finding an optimal hyperplane, and neural network classifies by learning the complex relationship between feature data. When the abnormal pattern of the vibration signal is complex and is affected by multiple factors (such as uneven tail wing material, worn polishing head, different polishing process parameters, etc.), machine learning algorithm can automatically mine the potential relationship between these complex factors and vibration anomaly by learning a large number of sample data. Once trained, it has certain adaptability and generalization ability. It can adapt to vibration changes under different tail wing shapes, polishing tools, polishing speeds, etc. without the need to manually adjust the threshold or statistical model for each condition. In the case of sufficient high-quality data, machine learning algorithm can make comprehensive judgments combining multiple features, avoiding the misjudgment that may occur in single feature threshold judgment. By learning the boundary between normal and abnormal vibration states, the vibration state categories can be more accurately divided, thereby improving the accuracy and reliability of vibration anomaly detection and better ensuring the stability of the polishing process and product quality.

[0104] Further, the automatic adjustment of the polishing parameters according to the vibration anomaly data, and the adjustment of the polishing head for polishing, comprises the following steps:

[0105] The abnormal vibration data is fuzzed into fuzzy language variables, and the fuzzy set in each fuzzy language variable corresponds to a membership function; the fuzzed vibration features are associated with the polishing parameter adjustment amount according to the fuzzy rule base; the current fuzzed vibration features are inferred in the rule base by a fuzzy inference algorithm to obtain the fuzzed polishing parameter adjustment amount; and the fuzzed adjustment amount is converted into an accurate parameter adjustment value.

[0106] The calculated polishing parameter adjustment value is sent to the robot control system to perform parameter adjustment operation, and the vibration of the polishing head is continuously monitored after parameter adjustment to determine whether it returns to normal. If it is still abnormal, the above steps are repeated to adjust the polishing parameters again, and machine learning is used for auxiliary adjustment.

[0107] Specifically, the detected abnormal vibration data is classified. According to the abnormal conditions of the frequency, amplitude, peak factor and other characteristics of the vibration, the abnormal vibration is divided into different types, such as high-frequency abnormal vibration, large-amplitude abnormal vibration, impact-type abnormal vibration, etc. The possible causes of each type of abnormal vibration are analyzed: high-frequency abnormal vibration may be caused by high-frequency resonance of the polishing head and the small uneven structure on the surface of the tail wing; large-amplitude abnormal vibration may be caused by the polishing head receiving a larger resistance, such as a sudden change in the local hardness of the tail wing material; impact-type abnormal vibration may be caused by the polishing head hitting a hard point or impurity on the surface of the tail wing. According to the type of abnormal vibration, the corresponding polishing parameter adjustment strategy is determined. For high-frequency abnormal vibration, the polishing head speed or the smoothness of the polishing path is adjusted; for large-amplitude abnormal vibration, the polishing pressure needs to be reduced or the feed speed of the polishing head is adjusted; for impact-type abnormal vibration, the instantaneous movement of the polishing head is paused or the polishing direction is changed to avoid the hard point. Based on empirical data, experimental results or theoretical analysis, a mapping relationship table between the adjustment strategy and the type of abnormal vibration is constructed. For example, when high-frequency abnormal vibration is detected and the frequency is within a certain range, the corresponding adjustment strategy is to reduce the polishing head speed by 10%-20% and increase the smoothness coefficient of the polishing path.

[0108] The abnormal vibration data (such as frequency deviation, amplitude deviation, etc.) is fuzzified into fuzzy language variables. For example, the amplitude deviation is divided into "positive large", "positive medium", "positive small", "zero", "negative small", "negative medium" and "negative large" fuzzy sets, each of which corresponds to a membership function. Common membership functions include triangular membership functions and trapezoidal membership functions. According to the fuzzy rule base, the fuzzified vibration characteristics are associated with the polishing parameter adjustment amount. For example, the rule is "if the amplitude deviation is positive large and the frequency deviation is positive small, then the polishing pressure is reduced a lot and the polishing head speed is reduced a little". These rules are established based on expert experience and experimental data. According to the current fuzzified vibration characteristics, the Mamdani reasoning method or Sugeno reasoning method in fuzzy reasoning algorithms is used to reason in the rule base to obtain the fuzzified polishing parameter adjustment amount. The fuzzified adjustment amount is converted into precise parameter adjustment values using the center of gravity method or the maximum membership degree method in the defuzzification method. For example, the specific adjustment values of the polishing pressure and the polishing head speed are calculated by the center of gravity method.

[0109] The calculated polishing parameter adjustment value is sent to the robot control system and the polishing head driving system to perform parameter adjustment operation. For example, the polishing head speed is changed by controlling the motor driver, and the polishing pressure is adjusted by the hydraulic or pneumatic system. After parameter adjustment, the vibration of the polishing head is continuously monitored. Through the same vibration sensor and signal processing method as the abnormal vibration detection, new vibration data is obtained. It is judged whether the vibration returns to the normal range. If it is still abnormal, repeat the above steps to adjust the polishing parameters again; if the vibration is normal, continue the polishing process and keep monitoring the vibration to deal with the possible abnormal situation again.

[0110] After the polishing parameter adjustment is executed and monitored, machine learning can be used to assist in adjustment. For example, a reinforcement learning algorithm is used. The state of the polishing system (including abnormal vibration data, current polishing parameters, etc.) is taken as the state space, the polishing parameter adjustment action is taken as the action space, and the polishing quality and vibration stability are taken as the reward signal. By continuously testing different parameter adjustment actions in actual polishing process, the optimal parameter adjustment strategy is learned according to the reward signal. For example, through the deep Q-network (DQN) algorithm, a neural network is trained to approximate the optimal Q value function, so as to select the best polishing parameter adjustment action according to the current state. This method can continuously optimize the parameter adjustment strategy in the long-term polishing process and adapt to different tail wings and polishing conditions.

[0111] Further, the dust and waste generated during the polishing process are collected and treated, including the following steps:

[0112] A dust suction port is arranged below the tail wing, and each dust suction port is connected to a central dust collection device through a pipeline;

[0113] The dust suction system is started before polishing starts, and the air volume and air speed of the dust suction system are controlled by adjusting the speed of the fan and the opening of the valve. A pressure difference sensor is installed in the pipeline system to determine whether the pipeline is blocked by monitoring the pressure difference between the two ends of the pipeline, and to control the flow balance;

[0114] Waste separation and classification are performed by cyclone separation technology, and the operating parameters of the treatment equipment are monitored during waste treatment to record waste information and establish a waste treatment file.

[0115] Specifically, a dust suction port is arranged below the tail wing, which should be as close as possible to the source of dust and waste, such as the position directly below the polishing head and below the edge of the tail wing, to improve the collection efficiency. The dust suction port is an opening part for sucking dust, debris, etc., and is connected to the dust suction pipeline to suck the surrounding dust into the equipment for collection and treatment through the generated suction force. The shape and size of the dust suction port are designed according to the size of the polishing area and the distribution of dust generation. The dust suction port is generally rectangular or circular, and its size should ensure that it can cover the main dust diffusion area. The pipelines are connected to the central dust collection equipment through the pipelines, and the pipelines should be made of smooth inner wall materials, such as galvanized steel pipes or PVC pipes, to reduce the deposition of dust in the pipelines. The diameter of the pipeline is determined according to the air volume and flow rate requirements of the dust suction system. The number of bends and changes in pipe diameter should be minimized to avoid poor air flow. The pipeline system is simulated and optimized through fluid mechanics simulation software, such as Fluent, to ensure that the air flow can smoothly transport the dust to the dust collection equipment. The appropriate dust collection equipment is selected according to the amount of dust generated by the polishing system, the size and properties of the dust particles, and other factors. Common dust collection equipment includes cyclone dust collectors, bag dust collectors, and electrostatic dust collectors. For larger and heavier dust particles, a cyclone dust collector is used for preliminary separation; for fine dust, a bag dust collector or an electrostatic dust collector is used for efficient collection. The dust collection equipment is installed in a well-ventilated location for easy maintenance and cleaning.

[0116] Before the polishing starts, the dust suction system is started to ensure that the dust and waste generated during the polishing process can be collected in time. The starting sequence of the dust suction system is to first start the fan of the dust collection equipment, and then open the valve of each dust suction port after the fan reaches the rated speed. The air volume and air speed of the dust suction system are controlled by adjusting the speed of the fan and the opening degree of the valve. Air speed sensors and pressure sensors are installed in the pipeline system to monitor the air flow parameters in real time. According to the polishing process and dust generation, the appropriate air speed range is set to ensure that the dust can be effectively sucked into the dust suction port and transported in the pipeline. A pressure difference sensor is installed in the pipeline system to determine whether the pipeline is blocked by monitoring the pressure difference between the two ends of the pipeline. When the pressure difference exceeds the set threshold, it indicates that the pipeline may be blocked. At this time, the control system issues an alarm and takes appropriate measures, such as starting the pipeline cleaning device to remove the blockage. The pipeline cleaning device includes a vibration device or a compressed air blowing device. In a system with multiple dust suction ports, flow balance control is performed to ensure that each dust suction port can effectively collect dust. Flow regulating valves are installed on the pipelines of each dust suction port, and feedback control is performed in combination with flow sensors. According to the deviation of the actual flow of each dust suction port from the set flow, the opening degree of the flow regulating valve is adjusted to keep the dust suction flow of each dust suction port balanced.

[0117] In the cyclone separation technology, there is a cyclone dust collector in the dust collection equipment, which separates particles of different sizes by using the principle of centrifugal force. When the dust-containing gas flow enters the cyclone dust collector, it rotates along the tangent direction of the wall. In the rotation process, larger particles are thrown to the wall due to the centrifugal force, and then fall along the wall to be collected, while smaller particles enter the next stage of dust removal equipment with the gas flow. For recyclable waste such as metal waste, after cleaning and drying, it can be recycled by compression or packaging. For non-recyclable waste such as plastic dust containing chemical coatings, it is treated according to environmental protection requirements. If the waste contains harmful substances, it is treated to be harmless. During the waste treatment process, the operating parameters of the treatment equipment (such as temperature, pressure, treatment time, etc.) are monitored to avoid secondary pollution. Record the treatment date, treatment amount, treatment method and other information of the waste, and establish a waste treatment file to trace and meet the environmental protection supervision requirements. These records can be stored in a database for easy query and statistical analysis.

[0118] A polishing method of a car spoiler mechanical arm, comprising the following steps:

[0119] Real-time scanning of the spoiler by a visual sensor to dynamically adjust the polishing path;

[0120] Real-time monitoring of the vibration state of the polishing head by a vibration sensor to obtain vibration abnormal data;

[0121] Automatically adjusting the polishing parameters according to the vibration abnormal data to adjust the polishing head for polishing;

[0122] Collecting and treating the dust and waste generated during the polishing process.

[0123] The above embodiments only describe the preferred embodiments of the present application, and do not limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements to the technical solutions of the present application made by ordinary engineering technicians in the art shall fall within the protection scope determined by the claims of the present application.

Claims

1. A robotic arm polishing system for automobile rear spoilers, characterized in that, It includes a path planning module for communication connections, a status monitoring module, a parameter compensation module, and a waste collection module, among which: The path planning module is used to scan the tail fin in real time using a vision sensor and dynamically adjust the polishing path; the vision sensor is a 3D structured light sensor or a stereo vision camera. The condition monitoring module is used to monitor the vibration status of the grinding head in real time through vibration sensors and obtain abnormal vibration data; The parameter compensation module is used to automatically adjust the grinding parameters based on abnormal vibration data, and adjust the grinding head for grinding. The waste collection module is used to collect and process the dust and waste generated during the grinding process. The method of monitoring the vibration state of the grinding head in real time using a vibration sensor to obtain abnormal vibration data includes the following steps: A vibration sensor is connected between the grinding head and the robotic arm. The vibration sensor is connected to a data acquisition card via a wire. The data acquisition card continuously collects the output signal of the vibration sensor during the grinding process and performs signal preprocessing. The DC component of the vibration signal is obtained by calculating the signal mean, the dispersion of the signal is obtained by calculating the signal standard deviation, the peak factor of the signal is extracted to determine the existence of impact vibration, the signal spectrum is obtained by performing a fast Fourier transform on the signal, and the spectrum amplitude is calculated. The main frequency components of the vibration are determined by the frequency points with larger amplitudes. The anomaly detection algorithm determines thresholds for signal mean, standard deviation, peak factor, and main frequency amplitude. When multiple features exceed the threshold simultaneously, it is judged as abnormal vibration. By statistically analyzing a large amount of vibration feature data under normal grinding conditions, the mean and standard deviation of the distribution are estimated, and the probability of newly collected vibration feature data under the assumed distribution is calculated. When the probability is lower than the probability threshold, it is judged as abnormal vibration, and the vibration feature data is recorded. Once the vibration is determined to be abnormal, abnormal vibration detection is performed using a support vector machine or a multilayer perceptron to identify abnormal vibrations under complex patterns. The automatic adjustment of grinding parameters based on abnormal vibration data, and the adjustment of the grinding head for grinding, includes the following steps: Abnormal vibration data is fuzzified into fuzzy linguistic variables, and each fuzzy set in the fuzzy linguistic variable corresponds to a membership function; according to the fuzzy rule base, the fuzzified vibration features are associated with the grinding parameter adjustment amount; based on the current fuzzified vibration features, the fuzzy inference algorithm is used to infer in the rule base to obtain the fuzzified grinding parameter adjustment amount; the fuzzified adjustment amount is converted into precise parameter adjustment values. The calculated grinding parameter adjustment values ​​are sent to the robot control system to perform parameter adjustment. After the parameter adjustment, the vibration of the grinding head is continuously monitored to determine whether it has returned to normal. If it is still abnormal, the above steps are repeated to adjust the grinding parameters again, and then machine learning is used for auxiliary adjustment.

2. The automotive rear wing robotic arm polishing system according to claim 1, characterized in that, The process of scanning the tail fin in real time using a visual sensor and dynamically adjusting the polishing path includes the following steps: The vision sensor is fixedly connected to the top or side of the robotic arm using a bracket and shock absorption device; The data acquisition computer and the robotic arm control system are connected via a two-way communication link; The polishing path is planned and dynamically adjusted using a path planning model. During the polishing process, the visual sensor parameters are continuously monitored by the monitoring program in the data acquisition computer and the robotic arm control system. When the sensor parameters exceed the set threshold, an alarm is issued and the polishing is paused.

3. The automotive rear wing robotic arm polishing system according to claim 2, characterized in that, The path planning model is through The algorithm and the feature-based model matching algorithm are used together to implement it, where: The algorithm is used to search from the starting point to the target point in path planning, generate the optimal path based on the polishing cost, and provide a baseline path for subsequent dynamic adjustments; Feature-based model matching algorithms are used to quantify the deviation between the tail fin model and the initial model after real-time feature alignment with the initial model, and to dynamically adjust the grinding path as a decision-making basis.

4. The automotive rear wing robotic arm polishing system according to claim 3, characterized in that, The The algorithm includes the following steps: The vision sensor is activated to perform an all-round scan of the tail fin placed on the workbench and fixed. The three-dimensional reconstruction algorithm based on stereo vision is used to reconstruct the three-dimensional image of each frame to obtain the three-dimensional point cloud model of the tail fin. The three-dimensional space of the tail fin is divided into a uniformly sized grid. The starting point is the position of the grinding tool, and the set of target positions on the tail fin surface that need to be ground is the set of target points. The algorithm searches for paths by maintaining open and closed lists; In each iteration, the node with the smallest cost function value is selected from the open list to determine the optimal path. The algorithm ends when the target point is added to the closed list. The initial polishing path is obtained by backtracking from the target point to the parent node of the starting point, and is used as the input of the feature-based model matching algorithm.

5. The automotive rear wing robotic arm polishing system according to claim 4, characterized in that, The feature-based model matching algorithm includes the following steps: During the polishing process, the vision sensor continuously scans the tail fin at a set frequency, and the real-time image frames are converted into a real-time tail fin 3D point cloud model through a 3D reconstruction algorithm, extracting features from the corresponding model of the initial polishing path and the real-time model. Corner features are extracted using the 3D Harris corner detection algorithm, and edge features are obtained by calculating the gradient direction change of the point cloud. The extracted feature point sets are the initial model feature point set and the real-time model feature point set, respectively. The nearest neighbor matching algorithm is used to calculate the distance between each feature point in the real-time model feature point set and the nearest feature point in the initial model feature point set. The transformation relationship that minimizes the sum of distances between all matched feature points is found through iterative optimization. The transformation relationship includes: translation vector, rotation matrix and scaling factor. Based on the transformation relationship obtained from model matching, the deviation between the real-time tail fin surface and the initial grinding path is analyzed. At the same time, the rotation angle deviation and scaling factor deviation are analyzed to determine the grinding path that needs to be adjusted. The grinding path is modified based on the path adjustment decision, and the adjusted grinding path is fed back to the robot control system.

6. The automotive rear wing robotic arm polishing system according to claim 1, characterized in that, The process of collecting and treating dust and waste generated during the polishing process includes the following steps: A dust suction port is installed below the tail fin, and each dust suction port is connected to a central dust collection device through pipes; Before polishing begins, start the vacuum system. Control the air volume and speed of the vacuum system by adjusting the fan speed and valve opening. Install a pressure differential sensor in the pipeline system to monitor the pressure difference between the two ends of the pipeline to determine whether the pipeline is blocked and to perform flow balance control. Waste is separated and classified using cyclone separation technology. During the waste treatment process, the operating parameters of the treatment equipment are monitored, waste information is recorded, and a waste treatment file is established.

7. A method for grinding a car rear spoiler using a robotic arm, applied to the car rear spoiler grinding system as described in any one of claims 1-6, characterized in that, Includes the following steps: The rear wing is scanned in real time using a vision sensor, and the polishing path is dynamically adjusted. The vibration status of the grinding head is monitored in real time by a vibration sensor to obtain abnormal vibration data; The grinding parameters are automatically adjusted based on abnormal vibration data, and the grinding head is adjusted accordingly for grinding. Dust and waste generated during the polishing process will be collected and disposed of in a centralized manner.

Images