Health assessment method, device and equipment for automated guided vehicle, and medium
By converting the vibration signal of the AGV driving system into symmetric point mode images and performing orderly clustering, a health status evaluation model is constructed, which solves the problem of insufficient accuracy of the AGV health assessment method in the existing technology, and achieves higher health status prediction accuracy.
Patent Information
- Application Number
- CN202411052539.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-12
AI Technical Summary
The accuracy of the prediction method based on vibration signals in the health evaluation of automatic guide vehicle (AGV) drive systems needs to be improved, and it is difficult to accurately predict the health status.
By converting the vibration signal sample data into symmetric point mode images, shape features are extracted, and orderly clustered, a health status evaluation model is constructed to improve the accuracy of health status prediction.
The accuracy of health status prediction of AGV drive system is improved, and by considering the local characteristics of the vibration signal and the similarity of samples in the cluster, it is more in line with the change law of the fault process, and the scientificity and accuracy of health status grading are enhanced.
Smart Images

Figure CN118940064A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of automated guided vehicle detection technology, and in particular to an automated guided vehicle health assessment method, device, equipment and medium. Background Art
[0002] The technical research on fault prognostics and health management (PHM) for automated guided vehicles (AGVs) is carried out to obtain the health status of the AGV system or perform fault prediction, and then take corresponding management measures to prevent the occurrence of faults based on different prediction results, which can effectively improve the work efficiency of the AGV system and reduce its operation and maintenance costs. However, in the PHM research of AGV systems, reference can be made to the PHM technology research of industrial robots or other related systems. In current research, the method based on vibration signals is widely used in the health monitoring of industrial robots and has become one of the representative methods. Since the vibration signal can be obtained from the surface of the industrial robot, it has almost no effect on the operation of the industrial robot. However, the accuracy of the method based on vibration signals to predict the health status in the health assessment of the AGV drive system needs to be improved. Summary of the invention
[0003] The purpose of this application is to provide a health assessment method, device, equipment and medium for an automated guided vehicle, which can improve the accuracy of health status prediction.
[0004] To achieve the above objectives, this application provides the following solutions:
[0005] In a first aspect, the present application provides a health assessment method for an automated guided vehicle, comprising:
[0006] Convert each vibration signal sample in the vibration signal sample data set into a symmetrical point pattern image; each vibration signal sample is a vibration signal of a driving system of an automatic guided vehicle;
[0007] extracting shape features of each of the symmetrical point pattern images as image features of each vibration signal sample;
[0008] Orderly clustering is performed on each image feature, and the health status of the automatic guided vehicle is graded according to the clustering result, each image feature corresponds to the health status level of the driving system of the automatic guided vehicle, and each image feature and the health status level corresponding to each image feature constitute a feature data set;
[0009] Using the feature data set, taking the image features as input and the health status level as output to train a classification model, a health status assessment model for an automated guided vehicle is obtained;
[0010] The vibration signal to be evaluated is output to the automatic guided vehicle health status evaluation model to obtain the health status classification of the automatic guided vehicle.
[0011] Optionally, converting each vibration signal sample in the vibration signal sample data set into a symmetrical point pattern image specifically includes:
[0012] The local binary pattern operator is used to describe the local characteristics of vibration signal samples;
[0013] Based on the local features, the vibration signal samples are converted into a symmetrical point pattern image.
[0014] Optionally, the local feature of the vibration signal sample described by the local binary pattern operator is expressed as:
[0015]
[0016] Wherein, L(t) is the local eigenvalue of the vibration signal sample between t and t+1, x(t) is the sampling value of the vibration signal sample at time t, x(t+τ) is the sampling value of the vibration signal at time t+τ, l is the time length, s(j) is the intermediate function, and j is the input of s(j);
[0017] Based on the local features, converting the vibration signal sample into a symmetrical point pattern image specifically includes:
[0018]
[0019]
[0020] Among them, r(t) is the polar value of the vibration signal sample converted to polar coordinates at time t, θ(t) and The two polar angle values of the vibration signal sample converted to polar coordinates at time t; x min is the minimum value of the signal; x max is the maximum value of the signal; θ is the rotation angle of the mirror symmetry plane; g is the angle magnification factor.
[0021] Optionally, the shape features include region area, region centroid and compactness.
[0022] Optionally, the image features are clustered in order, and the health status of the automatic guided vehicle is graded according to the clustering result, wherein each image feature corresponds to the health status level of the driving system of the automatic guided vehicle, specifically including:
[0023] The image features of n vibration signal samples are sorted according to the signal acquisition time; n is an integer greater than 1; each vibration signal sample corresponds to a signal acquisition time;
[0024] Divide the sorted n vibration signal samples into m-1 categories; m is an integer greater than 1;
[0025] The objective function of ordered clustering is constructed based on the silhouette coefficient between two adjacent clusters;
[0026] The clustering result is solved according to the objective function with the goal of maximizing the average silhouette coefficient.
[0027] Optionally, the objective function is expressed as:
[0028]
[0029] Among them, S A is the average silhouette coefficient, For the kth i Class and k i+1 The contour coefficient between the corresponding features of the vibration signal samples;
[0030]
[0031] in, For the kth i The average distance between the corresponding features of the vibration signal samples; k i The vibration signal samples correspond to the samples in the features to k i+1 The average distance between samples in the corresponding features of the vibration signal samples.
[0032] Optionally, taking maximization of the average silhouette coefficient as a goal, solving the clustering result according to the objective function specifically includes:
[0033] Using a genetic algorithm or a simulated annealing algorithm to maximize the average silhouette coefficient as a goal, and solving the clustering result according to the objective function;
[0034] The classification module is a support vector machine, a random forest, a decision tree or a neural network algorithm.
[0035] In a second aspect, the present application provides a health assessment device for an automated guided vehicle, comprising:
[0036] A vibration signal sample conversion module, used to convert each vibration signal sample in the vibration signal sample data set into a symmetrical point pattern image; each vibration signal sample is a vibration signal of a driving system of an automatic guided vehicle;
[0037] A feature extraction module, used to extract shape features of each of the symmetrical point pattern images as image features of each vibration signal sample;
[0038] A feature clustering module, used for orderly clustering each image feature, and grading the health status of the automatic guided vehicle according to the clustering result, each image feature corresponds to the health status level of the driving system of the automatic guided vehicle, and each image feature and the health status level corresponding to each image feature constitute a feature data set;
[0039] A model training module, used to use the feature data set and the image features as input and the health status level as output to train a classification model, so as to obtain an automatic guided vehicle health status assessment model;
[0040] The model application module is used to output the vibration signal to be evaluated to the automatic guided vehicle health status evaluation model to obtain the health status classification of the automatic guided vehicle.
[0041] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the above-mentioned methods for health assessment of an automated guided vehicle.
[0042] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the above-mentioned methods for health assessment of an automated guided vehicle.
[0043] According to the specific embodiments provided in this application, this application discloses the following technical effects:
[0044] The present application provides a health assessment method, device, equipment and medium for an automated guided vehicle. The method converts each vibration signal sample in a vibration signal sample data set into a symmetrical point pattern image, extracts the shape features of each symmetrical point pattern image as the image features of each vibration signal sample, and takes into account the local features of the vibration signal, thereby improving the accuracy of the health status prediction; orderly clusters each image feature, and grades the health status of the automated guided vehicle according to the clustering result, each image feature corresponds to the health status level of the drive system of the automated guided vehicle, and each image feature and the health status level corresponding to each image feature constitute a feature data set; takes into account the similarity of samples within a cluster and the degree of difference between a sample and samples between other clusters, so that the health status classification is more in line with the changing law of the fault sending process of the drive system, and further improves the accuracy of the health status prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0046] Figure 1 A schematic diagram of a flow chart of a health assessment method for an automated guided vehicle provided in one embodiment of the present application;
[0047] Figure 2 A principle block diagram of a health assessment method for an automated guided vehicle provided in one embodiment of the present application;
[0048] Figure 3 A schematic diagram of the SDP transformation principle incorporating local feature information provided in an embodiment of the present application;
[0049] Figure 4 A schematic diagram of an SDP image after a vibration signal is converted according to an embodiment of the present application;
[0050] Figure 5 A schematic diagram of vibration signal sampling provided in an embodiment of the present application;
[0051] Figure 6 A schematic diagram of an SDP image of multiple vibration signal samples provided by an embodiment of the present application;
[0052] Figure 7 A schematic diagram of a health assessment ROC curve provided in an embodiment of the present application;
[0053] Figure 8 A schematic diagram of functional modules of a health assessment device for an automated guided vehicle provided in one embodiment of the present application;
[0054] Fig. 9 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION
[0055] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0056] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.
[0057] This application provides a health assessment method for an automated guided vehicle, such as Figure 1 As shown, the health assessment method of the automatic guided vehicle includes:
[0058] Step 101: convert each vibration signal sample in the vibration signal sample data set into a symmetrical point pattern image; each vibration signal sample is a vibration signal of a driving system of an automatic guided vehicle.
[0059] Step 102: extracting shape features of each of the symmetrical point pattern images as image features of each vibration signal sample.
[0060] Step 103: clustering the image features in an orderly manner, and grading the health status of the automatic guided vehicle according to the clustering result, wherein each image feature corresponds to a health status level of a driving system of the automatic guided vehicle, and each image feature and the health status level corresponding to each image feature constitute a feature data set.
[0061] Step 104: Using the feature data set, the classification model is trained with the image features as input and the health status level as output to obtain an automatic guided vehicle health status assessment model.
[0062] Step 105: output the vibration signal to be evaluated to the automatic guided vehicle health status evaluation model to obtain the health status classification of the automatic guided vehicle.
[0063] This application analyzes the vibration signal of the AGV drive system to achieve a health assessment of the AGV drive system, such as Figure 2 As shown, by converting each vibration signal sample in the vibration signal sample data set into a symmetrical dot pattern image, extracting the shape features of each symmetrical dot pattern (Symmetrized Dot Pattern, SDP) image as the image feature of each vibration signal sample, taking into account the local features of the vibration signal, thereby improving the accuracy of health status prediction; by orderly clustering each image feature, the health status of the automatic guided vehicle is graded according to the clustering result, each image feature corresponds to the health status level of the automatic guided vehicle drive system, and each image feature and the health status level corresponding to each image feature constitute a feature data set; taking into account the similarity of samples within the cluster and the degree of difference between a sample and other cluster samples, the health status classification is more in line with the changing law of the drive system sending fault process, further improving the accuracy of health status prediction. After obtaining the health level of the AGV drive system, the feature data set marked with the health level is used to train and test the classification, and it is used as a model for AGV drive system health assessment (automatic guided vehicle health status assessment model).
[0064] The classification module is a support vector machine (SVM), a random forest, a decision tree or a neural network algorithm.
[0065] Among them, step 101 specifically includes: using a local binary pattern (LBP) operator to describe the local features of the vibration signal sample; based on the local features, converting the vibration signal sample into a symmetric point pattern image. This application incorporates the use of a local binary pattern operator to describe the local features of the vibration signal sample into the symmetric point pattern transformation, which is used to convert the vibration signal into an SDP image, and extract the features of the SDP image as the final features describing the vibration signal.
[0066] The local features of the vibration signal samples described by the local binary pattern operator are expressed as:
[0067]
[0068]
[0069] Among them, L(t) is the local feature of the vibration signal sample between t and t+1, x(t) is the sampling value of the vibration signal sample at time t, x(t+τ) is the sampling value of the vibration signal at time t+τ, the sample size is n, l is the time length, s(j) is the intermediate function, j is the input of s(j), t=1,2,3,…,nl.
[0070] Based on the local features, converting the vibration signal sample into a symmetrical point pattern image specifically includes:
[0071]
[0072] Among them, r(t) is the polar value of the vibration signal sample converted to polar coordinates at time t, θ(t) and The two polar angle values of the vibration signal sample converted to polar coordinates at time t; x min is the minimum value of the signal; x max is the maximum value of the signal, x min and x max is the maximum and minimum value of a complete vibration signal; θ is the rotation angle of the mirror symmetry plane; g is the angle magnification factor. The SDP transformation principle that incorporates local feature information is as follows: Figure 3 shown.
[0073] The shape features include region area, region centroid and compactness.
[0074] The shape features (region area, region centroid, and compactness) of the SDP image are calculated as follows:
[0075]
[0076] Where S is the area feature of the SDP image, I(i,j) is the pixel value at position (i,j), and A is the selected pixel value.
[0077]
[0078] Where x and y are the horizontal and vertical coordinates of the region centroid, respectively.
[0079]
[0080] Where C is the region compactness, L is the perimeter of the selected region, that is, the number of boundary pixels of the selected region, and S is the region area.
[0081] In step 102, feature extraction is performed by plotting the signal in a graph according to the SDP transformation result to obtain an SDP image, and extracting shape features of the SDP image (such as area, center of mass, compactness, etc.) as the final features of the signal. The converted vibration signal SDP image is shown in FIG. Figure 4 As shown, Figure 4 The SDP image after the vibration signal conversion shown in the figure is a graph of the vibration signals of the two channels drawn under different parameters. The green part in the figure is the image of the vibration signal channel 1, and the red part is the image of channel 2. During the SDP conversion, the vibration signal conversion parameters at channel 1 are θ=60°, l=9, g=50; the vibration signal conversion parameters at channel 2 are θ=120°, l=9, g=50.
[0082] In order to obtain a scientific and reasonable AGV system health level, the ordered clustering algorithm is used to analyze the vibration signal characteristics of the AGV system. The AGV system health level is obtained based on the clustering results, and the final vibration characteristics are divided into different health levels, laying the foundation for the training and testing of the AGV system health status assessment model based on vibration signal analysis. In the ordered clustering, the ordered clustering process is regarded as dividing the ordered samples into different types of data, and then the division results are evaluated, and the division method with the highest evaluation result is selected as the optimal clustering result.
[0083] Wherein, step 103 specifically includes:
[0084] The image features of n vibration signal samples are sorted according to the signal acquisition time; n is an integer greater than 1; each vibration signal sample corresponds to a signal acquisition time. Let {x1,x2,...,x n} is an ordered arrangement of SDP image features of n samples (vibration signal samples) data.
[0085] Ordered clustering principle: divide the sorted n vibration signal samples into m-1 categories; m is an integer greater than 1. There exists K = [k1 = 1, k2, k3, ..., k m =n], so that n sample data are divided into m-1 categories.
[0086] The m-1 classes are represented as follows:
[0087]
[0088] Where 1 = k1 <k2<…<k m =n,k i is the sequence number of the i-th sample.
[0089] Clustering result evaluation: The K with the highest clustering evaluation value is the optimal clustering result. The clustering result evaluation can use the silhouette coefficient, variance ratio criterion, Davies-Bouldin index (DB value), homogeneity, integrity and v measure, etc. This application uses the silhouette coefficient as a method and example for evaluating clustering results.
[0090] The silhouette coefficient for evaluating ordered clustering results is expressed as:
[0091]
[0092] in, For the kth i Class and k i+1 The contour coefficient between the samples of the vibration signal, i = 1, 2, ..., m-1, For the kth i The average distance between two samples of vibration signal samples; k i The samples of the vibration signal are k i+1 The average distance between samples in the vibration signal sample.
[0093]
[0094] Where x j and d(x m ,x p ) is x m and x p The distance measure between two samples can be Euclidean distance, Manhattan distance, etc.
[0095] The objective function of ordered clustering is constructed based on the silhouette coefficient between two adjacent clusters. When S>0, the larger S is, the more reasonable the clustering is; when S=0, it means that there is overlap in the clustering results; when S<0, the smaller S is, the more serious the situation of samples being misassigned to adjacent classes is. To find the optimal clustering result, by solving K=[k1=1,k2,k3,…,k m =n] to optimize the objective function value.
[0096] The clustering result is solved according to the objective function with the goal of maximizing the average silhouette coefficient.
[0097] The objective function is expressed as:
[0098]
[0099] Among them, S A is the average silhouette coefficient, For the kth i Class and k i+1 The contour coefficient between the samples of the vibration signal; S A ∈[-1,1], the closer S is to 1, the better the clustering effect.
[0100] The method of solving the clustering result K according to the objective function with the goal of maximizing the average silhouette coefficient specifically includes: using a genetic algorithm or a simulated annealing algorithm with the goal of maximizing the average silhouette coefficient and solving the clustering result according to the objective function.
[0101] In determining the health level number of the AGV system, the number of clusters in the ordered clustering method of this application can be set, and then the final characteristics of the vibration signal can be analyzed, and the number of clusters with the best clustering effect (clustering result) can be selected as the health level number, that is, the health status classification is specifically a health level for each cluster, and a health level is marked for the image features of each vibration signal sample; it can also be determined according to actual conditions or needs, but the health level corresponding to the final characteristics of the vibration signal needs to be determined based on the ordered clustering results.
[0102] Finally, the final feature samples of the vibration signal with the health level are used to train or test the classification model, and the classification model that passes the training and test is used as the model for AGV system health assessment.
[0103] In order to illustrate the scientificity and effectiveness of ordered clustering analysis in a health assessment method for automated guided vehicles, ordered clustering experiments and analysis were carried out. In order to illustrate the superiority of the SDP transformation incorporating local features compared with other methods, a health assessment performance comparison and analysis was carried out. In the verification experiment, the vibration signal comes from the sampled AGV drive system vibration signal during the actual operation of the AGV. During the vibration signal sampling process, the acceleration sensor is used to measure the vibration information of two normally serving AGVs at a constant speed (power) when they are empty. Vibration signal sampling is performed every 2-3 days. During sampling, 5 samples are collected from each AGV, and measurements are taken 5 times at position A and position B to avoid random error interference; each sample consists of the following: Figure 5 The vibration signals at position A and position B shown in the figure are composed of two channels of vibration information. The sampling frequency is 500 Hz / s and the sampling time is 10 seconds, so 5000 data points are collected in each channel. A total of 1240 experimental samples were collected in this AGV system health assessment study based on vibration signal analysis.
[0104] 1) Ordered clustering experiment and analysis: Before the ordered clustering experiment, Figure 2 The feature extraction step in the AGV system health assessment method shown in the figure extracts vibration signal features. First, according to formulas (1)-(5), the two channel vibration signals of the vibration signal sample are respectively subjected to SDP transformation (drawn in the same image) to obtain the SDP image of the vibration signal; secondly, the shape features (region area, region centroid, and compactness) of the vibration signal SDP image are extracted as the final features of the vibration signal. Since the vibration signal SDP image features of channel A and channel B are extracted separately, the final vibration signal feature dimension is 8. After the vibration signal feature is extracted, the genetic algorithm is used to calculate the K value to obtain the best clustering result for ordered clustering. In addition, when the ordered clustering is performed, the number of clusters needs to be determined. When the number of clusters is too large, the number of health levels of the AGV system will be too large, resulting in too small differences between adjacent health levels, which is not conducive to the development of health management. Therefore, in the ordered clustering experiment, the number of clusters m-1=2,…,10 is taken, and the optimal number of clusters from 2 to 10 is selected as the health level number of the AGV system.
[0105] As shown in Table 1, the experimental results of the feature ordered clustering algorithm (GA algorithm) based on the optimization algorithm of the present application are shown. The optimal clustering results and the corresponding silhouette coefficient values when the number of clusters is 2-10 are compared and analyzed. It can be seen that when the number of clusters is 3, the average silhouette coefficient S A =0.6287 reaches the maximum. Combined Figure 6 SDP images of 1240 vibration signals, sample SDP numbers 481-980.
[0106] There are significant differences between the image and the sample SDP images with serial numbers 1-480 and 981-1240. Therefore, it can be seen that the feature clustering analysis results are consistent with the visual effects of the vibration signal SDP image.
[0107] Table 1 Experimental results of feature ordered clustering based on optimization algorithm
[0108]
[0109]
[0110] In the ordered clustering analysis of the vibration signal SDP image features, the optimal clustering result of the ordered clustering is solved by the optimization algorithm. In order to illustrate its effectiveness and necessity, it is compared with the Fisher algorithm in the ordered clustering analysis experiment. In the Fisher ordered clustering experiment, after the vibration signal SDP image features are extracted, the Fisher ordered clustering algorithm is used for clustering analysis. The experimental results are shown in Table 2. Compared with the results in Table 1, it can be seen that the ordered clustering results based on the GA algorithm of this application are better than the results of the Fisher ordered clustering algorithm. It can be seen that the proposed ordered clustering analysis method based on the optimization algorithm is effective for clustering analysis of ordered samples (vibration signals).
[0111] Table 2 Experimental results of feature ordered clustering based on Fisher algorithm
[0112] Number of clusters Optimal clustering results <![CDATA[Average silhouette coefficient S A > 2 <![CDATA[{x1,…,x 1016 },{x 1017 ,…,x 1240 }]]> 0.5184 3 <![CDATA[{x1,…,x 816 },{x 817 ,…,x 1006 },{x 1007 ,…,x 1240 }]]> 0.2718 4 <![CDATA[{x1,…,x 812 },{x 813 ,…,x 1004 },…,{x 1226 ,…,x 1240 }]]> 0.1185 5 <![CDATA[{x1,…,x 793 },…,{x 1161 ,…,x 1240 }]]> 0.1117 6 <![CDATA[{x1,…,x 775 },…,{x 1199 ,…,x 1240 }]]> 0.1076 7 <![CDATA[{x1,…,x 771 },…,{x 1222 ,…,x 1240 }]]> 0.1051 8 <![CDATA[{x1,…,x 759 },…,{x 1201 ,…,x 1240 }]]> 0.1038 9 <![CDATA[{x1,…,x 754 },…,{x 1222 ,…,x 1240 }]]> 0.0960 10 <![CDATA[{x1,…,x 748 },…,{x 1225 ,…,x 1240 }]]> 0.1080
[0113] In the method proposed in this application, the clustering effect is evaluated to solve the method of solving the optimal clustering result, which comprehensively considers the similarity between the sample and the samples in the cluster to which it belongs and the degree of difference between the samples in other nearest clusters. Compared with the Fisher ordered clustering algorithm which only considers the principle of minimizing the distance between samples of the same type, the ordered clustering result of the proposed method is more scientific and reasonable.
[0114] 2) Comparison and analysis of health assessment performance: In the health assessment of the AGV system, the LBP operator is used to describe the local features of the vibration signal in the SDP image transformation of the vibration signal, and it is integrated into the SDP image transformation of the vibration signal to enhance the difference between different categories of vibration signals after conversion to SDP images. After the SDP image conversion, the area, center of mass, and compactness characteristics of the SDP image are extracted as characteristics to describe the vibration signal, and the vibration signal characteristics are used to train and verify the health assessment model of the AGV drive system. In order to illustrate the effectiveness of this method, it is compared with the traditional SDP image transformation method.
[0115] Table 3 Comparison of health assessment accuracy of AGV drive system (10-fold cross validation)
[0116] Number of tests The present application SDP image transformation method Original SDP image transformation method First fold verification accuracy 98.39% 90.32% Second fold verification accuracy 96.78% 89.52% The third fold verification accuracy 97.58% 88.71% Fourth fold verification accuracy 96.76% 93.55% Fifth fold verification accuracy 95.97% 91.13% Sixth fold verification accuracy 96.80% 91.13% Seventh fold verification accuracy 97.58% 90.32% The eighth fold verification accuracy 96.74% 89.52% Ninth fold verification accuracy 96.75% 91.94% 10th fold verification accuracy 96.79% 90.32% Average accuracy 97.01% 90.65%
[0117] As shown in Table 3, the average accuracy of the health assessment of the AGV system proposed in this application is 97.01%, which is 6.36% higher than the traditional SDP image conversion method. From the results in Table 3, it can be seen that in the method proposed in this application, in the extraction of vibration signal features, incorporating local vibration signal features into SDP image transformation can more accurately extract the features of the vibration signal, effectively improving the accuracy of the health assessment of the AGV drive system.
[0118] Figure 7 The Receiver Operating Characteristic Curve (ROC) of the AGV system health assessment is plotted in the paper, and the proposed method of incorporating the local features of the vibration signal into the SDP image transformation is compared with the AGV drive system health assessment results based on the traditional SDP algorithm. According to the ROC curve in the figure, the AUC value of the proposed algorithm is larger. Therefore, the proposed AGV drive system health assessment results are more effective.
[0119] The AGV system health level assessment method based on vibration signal analysis in the present application can be used not only to conduct health assessments on AGV systems, but also to conduct health assessments on other similar electromechanical systems or their key components. The method mentioned therein of extracting the signal SDP image shape features as the final features of the signal can be replaced by other image features, that is, the extracted image features include but are not limited to features such as area, region centroid, and compactness.
[0120] Based on the same inventive concept, the embodiment of the present application also provides an automatic guided vehicle health assessment device for implementing the automatic guided vehicle health assessment method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiments of one or more automatic guided vehicle health assessment devices provided below can refer to the limitations of the automatic guided vehicle health assessment method above, and will not be repeated here.
[0121] In an exemplary embodiment, Figure 8 As shown, a health assessment device for an automated guided vehicle is provided, comprising:
[0122] The vibration signal sample conversion module 201 is used to convert each vibration signal sample in the vibration signal sample data set into a symmetrical point pattern image; each vibration signal sample is a vibration signal of the driving system of the automatic guided vehicle.
[0123] The feature extraction module 202 is used to extract the shape features of each of the symmetrical point pattern images as image features of each vibration signal sample.
[0124] The feature clustering module 203 is used to perform orderly clustering on each image feature, and classify the health status of the automatic guided vehicle according to the clustering result. Each image feature corresponds to the health status level of the driving system of the automatic guided vehicle. Each image feature and the health status level corresponding to each image feature constitute a feature data set.
[0125] The model training module 204 is used to use the feature data set, take the image features as input and the health status level as output to train the classification model, and obtain the automatic guided vehicle health status assessment model.
[0126] The model application module 205 is used to output the vibration signal to be evaluated to the automatic guided vehicle health status evaluation model to obtain the health status classification of the automatic guided vehicle.
[0127] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Fig. 9 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store health assessment data of an automatic guided vehicle. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a health assessment method for an automatic guided vehicle is implemented.
[0128] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.
[0129] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0130] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0131] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0132] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0133] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a data processing logic of a programmable logic device, etc., but is not limited thereto.
[0134] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0135] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.
Claims
1. A health assessment method for an automated guided vehicle, characterized in that: The health assessment method of the automatic guided vehicle comprises: Convert each vibration signal sample in the vibration signal sample data set into a symmetrical point pattern image; each vibration signal sample is a vibration signal of a driving system of an automatic guided vehicle; extracting shape features of each of the symmetrical point pattern images as image features of each vibration signal sample; Orderly clustering is performed on each image feature, and the health status of the automatic guided vehicle is graded according to the clustering result, each image feature corresponds to the health status level of the driving system of the automatic guided vehicle, and each image feature and the health status level corresponding to each image feature constitute a feature data set; Using the feature data set, taking the image features as input and the health status level as output to train a classification model, a health status assessment model for an automated guided vehicle is obtained; The vibration signal to be evaluated is output to the automatic guided vehicle health status evaluation model to obtain the health status classification of the automatic guided vehicle.
2. The health assessment method of an automated guided vehicle according to claim 1, characterized in that: Each vibration signal sample in the vibration signal sample data set is converted into a symmetrical point pattern image, specifically including: The local binary pattern operator is used to describe the local characteristics of vibration signal samples; Based on the local features, the vibration signal samples are converted into a symmetrical point pattern image.
3. The health assessment method of an automated guided vehicle according to claim 2, characterized in that: The local features of the vibration signal samples described by the local binary pattern operator are expressed as: Wherein, L(t) is the local eigenvalue of the vibration signal sample between t and t+1, x(t) is the sampling value of the vibration signal sample at time t, x(t+τ) is the sampling value of the vibration signal at time t+T, l is the time length, s(j) is the intermediate function, and j is the input of s(j); Based on the local features, converting the vibration signal sample into a symmetrical point pattern image specifically includes: Among them, r(t) is the polar value of the vibration signal sample converted to polar coordinates at time t, θ(t) and The two polar angle values of the vibration signal sample converted to polar coordinates at time t; x min is the minimum value of the signal; x max is the maximum value of the signal; θ is the rotation angle of the mirror symmetry plane; g is the angle magnification factor.
4. The health assessment method of an automated guided vehicle according to claim 1, characterized in that: The shape features include region area, region centroid and compactness.
5. The health assessment method of an automated guided vehicle according to claim 1, characterized in that: The image features are clustered in order, and the health status of the automatic guided vehicle is graded according to the clustering results. Each image feature corresponds to the health status level of the drive system of the automatic guided vehicle, specifically including: The image features of n vibration signal samples are sorted according to the signal acquisition time; n is an integer greater than 1; each vibration signal sample corresponds to a signal acquisition time; Divide the sorted n vibration signal samples into m-1 categories; m is an integer greater than 1; The objective function of ordered clustering is constructed based on the silhouette coefficient between two adjacent clusters; The clustering result is solved according to the objective function with the goal of maximizing the average silhouette coefficient.
6. The health assessment method of an automated guided vehicle according to claim 5, characterized in that: The objective function is expressed as: Among them, S A is the average silhouette coefficient, For the kth i Class and k i+1 The contour coefficient between the corresponding features of the vibration signal samples; in, For the kth i The average distance between the corresponding features of the vibration signal samples; k i The corresponding features of the vibration signal samples are samples to k i+1 The average distance between samples in the corresponding features of the vibration signal samples.
7. The health assessment method of an automated guided vehicle according to claim 5, characterized in that: Taking the maximization of the average silhouette coefficient as the goal, the clustering result is solved according to the objective function, specifically including: Using a genetic algorithm or a simulated annealing algorithm to maximize the average silhouette coefficient as a goal, and solving the clustering result according to the objective function; The classification module is a support vector machine, a random forest, a decision tree or a neural network algorithm.
8. A health assessment device for an automated guided vehicle, characterized in that: The health assessment device of the automatic guided vehicle comprises: A vibration signal sample conversion module, used to convert each vibration signal sample in the vibration signal sample data set into a symmetrical point pattern image; each vibration signal sample is a vibration signal of a driving system of an automatic guided vehicle; A feature extraction module, used to extract shape features of each of the symmetrical point pattern images as image features of each vibration signal sample; A feature clustering module, used for orderly clustering each image feature, and grading the health status of the automatic guided vehicle according to the clustering result, each image feature corresponds to the health status level of the driving system of the automatic guided vehicle, and each image feature and the health status level corresponding to each image feature constitute a feature data set; A model training module, used to use the feature data set and the image features as input and the health status level as output to train a classification model, so as to obtain an automatic guided vehicle health status assessment model; The model application module is used to output the vibration signal to be evaluated to the automatic guided vehicle health status evaluation model to obtain the health status classification of the automatic guided vehicle.
9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the health assessment method for an automated guided vehicle according to any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the health assessment method for an automated guided vehicle according to any one of claims 1 to 7 is implemented.
