Truck scale signal abnormity identification method and system based on historical weighing signals
By combining the historical weighing signal and the current weighing signal, the time correlation and information correlation are calculated, and the vehicle scale signal abnormality is quickly identified, which solves the problem of slow processing speed in the prior art and meets the real-time requirements.
Patent Information
- Application Number
- CN202510335136.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-20
- Publication Date
- 2025-06-24
AI Technical Summary
The prior art has the problem of slow processing speed and inability to meet real-time requirements in the recognition of weighing signal abnormalities of automobile scales.
The abnormal identification method of automobile scale signal based on historical weighing signals is adopted. By combining historical weighing signals and current weighing signals, the in-class signals and extra-class signals are divided, the time correlation degree and information correlation degree are calculated, and the abnormal score is calculated to quickly identify abnormal signals.
On the premise of ensuring recognition accuracy, more efficient real-time exception recognition is achieved without the need for complex processing algorithms.
Smart Images
Figure CN120197102A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of signal processing, and in particular, to a method for identifying abnormal vehicle scale signals based on historical weighing signals, a system for identifying abnormal vehicle scale signals based on historical weighing signals, a computer device, a computer-readable storage medium, and a computer program product. Background Art
[0002] The statements in this section merely provide background art related to the present invention and do not necessarily constitute prior art.
[0003] A vehicle scale, also known as a truck scale, is a large weighing scale set on the ground, usually used to weigh the cargo tons of trucks. It is the main weighing equipment for factories, mines, merchants, etc. to measure bulk goods. Abnormalities often occur in the weighing signals of vehicle scales, and the main reasons include: (1) Sensor aging or damage, poor connection between the sensor output cable and the instrument, and displacement of the sensor after being affected by moisture or water immersion; (2) Display failure, internal failure of the instrument; (3) Unstable power supply, and the presence of equipment such as electric bells, high-voltage lines, or interference sources such as ongoing welding operations nearby; (4) The limit device between the scale body and the foundation may be abnormal, resulting in unstable weighing signals, debris blocking at the bottom of the scale body, deformation of the scale body, or settlement of the foundation.
[0004] In a real environment, identifying abnormal weighing signals of a vehicle scale is a very complex process. Field workers often cannot identify weighing abnormalities manually, and existing signal abnormality identification methods may involve complex algorithms and data processing processes, resulting in slow processing speeds and unable to meet the requirements of weighing scenarios with high real-time requirements. Summary of the Invention
[0005] To solve the deficiencies of the prior art, the present invention provides a method and system for identifying abnormal vehicle scale signals based on historical weighing signals. By combining historical weighing signals and current weighing signals, class-in signals and class-out signals are divided. According to the division results, the time correlation degree and information correlation degree are calculated respectively. Based on the time correlation degree and information correlation degree, the abnormal score is calculated, and based on the abnormal score, rapid identification of abnormal vehicle scale signals is achieved. Without complex processing algorithms, more efficient real-time abnormality identification is realized on the premise of ensuring the identification accuracy.
[0006] To achieve the above object, the present invention adopts the following technical solutions:
[0007] In the first aspect, the present invention provides a method for identifying abnormal vehicle scale signals based on historical weighing signals.
[0008] A method for identifying abnormal vehicle scale signals based on historical weighing signals includes the following processes:
[0009] Obtain the current weighing signal and multiple historical weighing signals of the weighbridge;
[0010] Separate the current weighing signal and the historical weighing signals individually for signal division to obtain the in-class signals corresponding to the current weighing signal and the historical weighing signals respectively;
[0011] Extract signal segments from the current weighing signal and the historical weighing signals, and combine the signal segments to obtain out-of-class signals;
[0012] Calculate the in-class time correlation degree based on the in-class signals to obtain the in-class time correlation degree, and calculate the out-of-class time correlation degree based on the out-of-class signals to obtain the out-of-class time correlation degree;
[0013] Calculate the in-class information correlation degree based on the in-class signals to obtain the in-class information correlation degree, and calculate the out-of-class information correlation degree based on the out-of-class signals to obtain the out-of-class information correlation degree;
[0014] Calculate the anomaly score based on the differences between the in-class time correlation degrees and the out-of-class time correlation degrees, and the differences between the in-class information correlation degrees and the out-of-class information correlation degrees. When the anomaly score is greater than the set threshold, it is determined that there is an abnormal signal; otherwise, it is determined that there is no abnormal signal.
[0015] As a further limitation of the first aspect of the present invention, extracting signal segments from the current weighing signal and the historical weighing signals, and combining the signal segments includes:
[0016] Place the current weighing signal and the historical weighing signals of the same length one above the other in sequence, intercept the current weighing signal and the historical weighing signals within a set time window to obtain multiple signal segments, and combine the obtained signal segments to obtain out-of-class signals.
[0017] As a further limitation of the first aspect of the present invention, for any in-class signal, calculate the change degree of the in-class signal at the i-th time point and the j-th time point as the in-class time correlation degree of this in-class signal, and calculate the in-class time correlation degrees of each in-class signal;
[0018] For the out-of-class signal, calculate the change degree of the out-of-class signal at the i-th time point and the j-th time point as the out-of-class time correlation degree.
[0019] As a further limitation of the first aspect of the present invention, for any in-class signal, calculate the in-class query vector, in-class key vector, and in-class value vector at the i-th time point, and obtain the in-class information correlation degree corresponding to this in-class signal according to the in-class query vector, in-class key vector, and in-class value vector, and calculate the in-class information correlation degrees of each in-class signal;
[0020] For out-of-class signals, calculate the out-of-class query vector, out-of-class key vector, and out-of-class value vector at the i-th time point, and obtain the out-of-class information correlation degree according to the out-of-class query vector, out-of-class key vector, and out-of-class value vector.
[0021] As a further limitation of the first aspect of the present invention, the calculation of the anomaly score includes:
[0022] Calculate the temporal JS divergence between each intra-class temporal correlation degree and the out-of-class temporal correlation degree respectively, and superimpose the obtained multiple temporal JS divergences to obtain the first JS divergence;
[0023] Calculate the information JS divergence between each intra-class information correlation degree and the out-of-class information correlation degree respectively, and superimpose the obtained multiple information JS divergences to obtain the second JS divergence;
[0024] Obtain the final anomaly score according to the first JS divergence and the second JS divergence.
[0025] As a further limitation of the first aspect of the present invention, where TC m represents the m-th intra-class temporal correlation degree, TC n represents the out-of-class temporal correlation degree, IC m represents the m-th intra-class information correlation degree, IC n represents the out-of-class information correlation degree, JS(TC m ||TC n ) represents the m-th temporal JS divergence, IC m ||IC n represents the m-th information JS divergence, and M represents the total number of signals.
[0026] In a second aspect, the present invention provides a weighbridge signal anomaly recognition system based on historical weighing signals.
[0027] A weighbridge signal anomaly recognition system based on historical weighing signals, comprising:
[0028] A signal acquisition unit, configured to: acquire the current weighing signal of the weighbridge and multiple historical weighing signals;
[0029] An intra-class signal division unit, configured to: separately perform signal division on the current weighing signal and the historical weighing signals, and respectively obtain the intra-class signals corresponding to the current weighing signal and the historical weighing signals;
[0030] An out-of-class signal division unit, configured to: extract signal segments from the current weighing signal and the historical weighing signals, and combine the signal segments to obtain out-of-class signals;
[0031] A time correlation calculation unit, configured to: perform intra-class time correlation calculation based on the intra-class signals to obtain an intra-class time correlation, and perform inter-class time correlation calculation based on the inter-class signals to obtain an inter-class time correlation;
[0032] An information correlation calculation unit, configured to: perform intra-class information correlation calculation based on the intra-class signals to obtain an intra-class information correlation, and perform inter-class information correlation calculation based on the inter-class signals to obtain an inter-class information correlation;
[0033] An anomaly determination unit, configured to: calculate an anomaly score based on the differences between the intra-class time correlations and the inter-class time correlations, and the differences between the intra-class information correlations and the inter-class information correlations, and determine that there is an anomaly signal when the anomaly score is greater than a set threshold, otherwise, determine that there is no anomaly signal.
[0034] In a third aspect, the present invention provides a computer device, including: a processor and a computer-readable storage medium;
[0035] A processor, adapted to execute a computer program;
[0036] The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, it implements the method for identifying abnormal signals of a weighbridge based on historical weighing signals as described in the first aspect of the present invention.
[0037] In a fourth aspect, the present invention provides a computer-readable storage medium, which stores a computer program, and the computer program is adapted to be loaded and executed by a processor to implement the method for identifying abnormal signals of a weighbridge based on historical weighing signals as described in the first aspect of the present invention.
[0038] In a fifth aspect, the present invention provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the method for identifying abnormal signals of a weighbridge based on historical weighing signals as described in the first aspect of the present invention.
[0039] Compared with the prior art, the beneficial effects of the present invention are:
[0040] 1. The present invention innovatively proposes a strategy for identifying abnormal signals of truck scales based on historical weighing signals. By combining historical weighing signals and current weighing signals, it divides intra-class signals and extra-class signals, calculates the time correlation degree and information correlation degree respectively according to the division results, calculates the abnormal score based on the time correlation degree and information correlation degree, and quickly identifies abnormal truck scale signals according to the abnormal score. Without complex processing algorithms, it realizes more efficient real-time abnormal identification while ensuring the identification accuracy.
[0041] 2. The present invention separately performs signal division on the current weighing signal and the historical weighing signal, respectively obtains the intra-class signals corresponding to the current weighing signal and the historical weighing signal, extracts signal segments from the current weighing signal and the historical weighing signal, combines the signal segments to obtain extra-class signals, effectively processes and fuses the current weighing signal and the historical weighing signal, can better draw on the signal characteristics in the historical weighing signal, and ensures the accuracy of signal processing.
[0042] 3. The present invention calculates the abnormal score according to the differences between the intra-class time correlation degree and the extra-class time correlation degree, and the differences between the intra-class information correlation degree and the extra-class information correlation degree. When the abnormal score is greater than the set threshold, it is determined that there is an abnormal signal; otherwise, it is determined that there is no abnormal signal. The abnormal score integrates the differences in intra-class and extra-class time correlation degrees and the differences in intra-class and extra-class information correlation degrees, ensuring the accuracy of abnormal score calculation.
[0043] Advantages of additional aspects of the present invention will be partly given in the following description, partly will become obvious from the following description, or will be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] The accompanying drawings forming a part of this specification are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention.
[0045] Figure 1 It is a schematic flow chart of a method for identifying abnormal truck scale signals based on historical weighing signals provided in Embodiment 1 of the present invention;
[0046] Figure 2 It is a schematic diagram of a system for identifying abnormal truck scale signals based on historical weighing signals provided in Embodiment 2 of the present invention;
[0047] Figure 3 It is a schematic diagram of a computer device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0048] The present invention will be further described below in conjunction with the drawings and embodiments.
[0049] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.
[0050] In the case of no conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.
[0051] Embodiment 1:
[0052] As described in the background art, although certain progress has been made in dealing with signal anomalies in the weighing scenario by existing signal anomaly recognition methods, many challenges still remain. These methods may involve complex algorithms and cumbersome data processing procedures, which not only increase the computational complexity but also often result in a slow processing speed, making it difficult to meet those weighing scenarios with high real-time requirements.
[0053] Existing signal anomaly recognition methods usually rely on advanced algorithms and technologies, such as machine learning, deep learning, statistical analysis, etc. These methods have significant advantages in dealing with large-scale data, but in practical applications, especially in the weighing scenario, their complexity often becomes the key factor restricting real-time performance.
[0054] For example, machine learning algorithms, especially supervised learning and unsupervised learning, have been widely used in signal anomaly recognition. Supervised learning trains a model to recognize known signal patterns and classify new signals. However, this method requires a large amount of labeled data, and the training process of the model may be very time-consuming. Unsupervised learning attempts to discover the hidden structure in the data without labels, but this also requires complex clustering or dimensionality reduction algorithms, which may be inadequate when dealing with real-time data.
[0055] For another example, deep learning, especially convolutional neural networks (CNNs) and recurrent neural networks (RNNs), have achieved remarkable results in the field of signal processing and anomaly detection in recent years. These technologies can extract complex features from the original signal and perform non-linear transformations through multiple layers of neural networks, thereby achieving accurate recognition of abnormal signals. However, the training and optimization of deep learning models usually require a large amount of computing resources and time, and in practical applications, the inference speed of the model may also be restricted by hardware conditions.
[0056] For another example, statistical analysis methods, such as hypothesis testing, variance analysis, time series analysis, etc., are also commonly used means for signal anomaly recognition. These methods detect anomalies by comparing the differences between the actual signal and the expected signal. Although statistical analysis methods have high accuracy in some cases, their calculation processes may involve a large amount of mathematical operations and iterative optimizations, thus affecting the processing speed.
[0057] In addition to the complexity of the algorithms themselves, existing signal anomaly recognition methods also involve cumbersome data processing procedures, which further increase the processing time. Specifically, they include:
[0058] (1) Data preprocessing: Data preprocessing is an important step in the process of signal anomaly recognition. It includes operations such as data cleaning, denoising, normalization, etc., to ensure the quality and consistency of the input data; however, these preprocessing steps often require a large amount of time and computing resources, especially when dealing with large-scale datasets;
[0059] (2) Feature extraction: Feature extraction is the process of converting the original signal into a feature vector that can be used in machine learning or deep learning models. This process requires careful design of feature extraction algorithms to ensure that the extracted features can effectively reflect the anomaly information in the signal; however, the design and optimization of feature extraction algorithms themselves are complex issues, and in practical applications, the speed of feature extraction may also be limited by hardware conditions;
[0060] (3) Model training and validation: In signal anomaly recognition, model training and validation are essential steps. Model training involves the process of adjusting model parameters to minimize the loss function, while model validation is used to evaluate the generalization ability of the model. Both of these steps require a large amount of computing resources and time, especially when dealing with large-scale datasets and complex models;
[0061] (4) Real-time inference: In real-time weighing scenarios, signal anomaly recognition methods need to be able to accurately classify input signals in a short time. However, existing methods often cannot meet this requirement due to the complexity of the algorithms and the cumbersomeness of the data processing procedures. The bottleneck of real-time inference speed may lie in the model inference itself, or in the preprocessing steps such as data preprocessing and feature extraction.
[0062] Weighing scenarios, especially those with high real-time requirements, such as industrial automation, logistics transportation, intelligent manufacturing and other fields, also pose higher challenges to signal anomaly recognition methods.
[0063] In an industrial automation system, sensors are used to monitor the operating status and process parameters of equipment in real time. When the sensor signals are abnormal, the system needs to be able to quickly identify and take corresponding measures to avoid failures or accidents. However, due to the slow processing speed of existing signal abnormality recognition methods, they may not be able to respond to these abnormal signals in a timely manner.
[0064] In the field of logistics transportation, load cells are widely used in cargo weight monitoring and vehicle load control. When the cargo weight exceeds the preset threshold, the system needs to be able to immediately issue an alarm and take measures to prevent overloading or transportation accidents. However, due to the lack of real-time performance of existing methods, problems such as alarm delay, false alarms, and missed alarms may occur.
[0065] In an intelligent manufacturing environment, sensors and intelligent devices are used to monitor various parameters in the production process in real time. When these parameters are abnormal, the system needs to be able to quickly identify and take corrective measures to ensure production quality and efficiency. However, existing signal abnormality recognition methods may be inadequate when dealing with this real-time data.
[0066] In view of the problems existing in the prior art, this implementation proposes a method for identifying abnormal signals of truck scales based on historical weighing signals, as Figure 1 shown, including the following processes:
[0067] S1: Obtain the current weighing signal of the truck scale and multiple historical weighing signals;
[0068] S2: Separately perform signal division on the current weighing signal and the historical weighing signals to obtain the in-class signals corresponding to the current weighing signal and the historical weighing signals respectively;
[0069] S3: Extract signal segments from the current weighing signal and the historical weighing signals, and combine the signal segments to obtain out-of-class signals;
[0070] S4: Calculate the in-class time correlation degree based on the in-class signals to obtain the in-class time correlation degree, and calculate the out-of-class time correlation degree based on the out-of-class signals to obtain the out-of-class time correlation degree;
[0071] S5: Calculate the in-class information correlation degree based on the in-class signals to obtain the in-class information correlation degree, and calculate the out-of-class information correlation degree based on the out-of-class signals to obtain the out-of-class information correlation degree;
[0072] S6: Calculate the abnormal score based on the differences between the in-class time correlation degrees and the out-of-class time correlation degrees, and the differences between the in-class information correlation degrees and the out-of-class information correlation degrees. When the abnormal score is greater than the set threshold, it is determined that there is an abnormal signal; otherwise, it is determined that there is no abnormal signal.
[0073] When there is an abnormal signal, the historical signal and the current signal are respectively compared and matched with the corresponding factory standard signals (such as by calculating distances, etc.). Based on the results of the comparison and matching, it is determined which signal is abnormal. For example, when the matching distance of a certain signal is greater than the set threshold, or when the matching distance of a certain signal is the largest, or when the difference in the matching distances between a certain signal and other signals is greater than the set threshold, it is determined that this signal is abnormal.
[0074] In step S1 of this implementation method, the current weighing signal and multiple historical weighing signals of the weighbridge are obtained. Specifically, it includes:
[0075] The current weighing signal is obtained, the length of the current weighing signal is determined, and then multiple segments of historical signals are obtained. Each segment of historical signal does not overlap, and multiple segments of historical signals with a time span greater than the set time threshold are taken. For example, historical signals with a time span of one month are taken respectively. Here, preferably, three segments of historical signals can be selected for joint processing with the current weighing signal.
[0076] It can be understood that in some other implementation methods, more segments of historical signals can also be selected. For example, 4 segments, 5 segments or 6 segments, etc. can be selected. Details are not elaborated here as long as the length can meet the requirements of data processing.
[0077] In step S2 of this implementation method, the current weighing signal and the historical weighing signal are respectively and separately divided into signal segments. Specifically, it includes: The current weighing signal and the historical weighing signal are respectively divided into multiple equal-length segments according to the preset window size (i.e., time window), and there is usually no overlap between each segment.
[0078] In step S3 of this implementation method, the current weighing signal and the historical weighing signal are used to extract signal segments, and the signal segments are combined to obtain outlier signals. Specifically, it includes:
[0079] The current weighing signal and the historical weighing signal of the same length are placed one above the other in sequence (keeping the starting points aligned), and within the set time window, the current weighing signal and the historical weighing signal are intercepted to obtain multiple signal segments. The obtained signal segments are sequentially combined to obtain outlier signals. For the combined signal, the time period with the time point of the current weighing signal as the initial period, and the signals of other signal segments are sequentially extended at the same time interval. For example, if the time period of the current weighing signal is 1s - 100s, then the time period of the next signal segment is 101s - 200s, and so on.
[0080] More specifically, for example, when there are three historical signals and one current signal, there are four signals from top to bottom. After intercepting using a time window, 4*N signal segments can be obtained, where N is a positive integer greater than or equal to 1. Arranging the obtained 4*N signal segments in sequence can obtain outlier signals.
[0081] In step S4 of this implementation manner, according to the intra-class signals, intra-class time correlation degree calculation is performed to obtain the intra-class time correlation degree. Taking two historical signals and one current signal as an example, specifically, it includes:
[0082] For the first intra-class signal, calculate the change degree of the intra-class signal at the i-th time point and other each time point as the intra-class time correlation degree of this intra-class signal, and calculate the intra-class time correlation degrees of each intra-class signal:
[0083]
[0084] Among them, TC m1 is the first intra-class time correlation degree, σ m1 =W σm1 I m1 Among them, W σ1 is a trainable weight matrix, I m1 is the first intra-class signal, and i and j respectively represent the i-th time point and the j-th time point.
[0085] For the second intra-class signal, calculate the change degree of the intra-class signal at the i-th time point and other each time point as the intra-class time correlation degree of this intra-class signal, and calculate the intra-class time correlation degrees of each intra-class signal:
[0086]
[0087] Among them, TC m2 is the second intra-class time correlation degree, σ m2 =W σm2 I m2 Among them, W σm2 is a trainable weight matrix, I m2 is the second intra-class signal, and i and j respectively represent the i-th time point and the j-th time point.
[0088] For the third intra-class signal, calculate the change degree of the intra-class signal at the i-th time point and other each time point as the intra-class time correlation degree of this intra-class signal, and calculate the intra-class time correlation degrees of each intra-class signal:
[0089]
[0090] Among them, TC m3 is the third intra-class time correlation degree, σm3 = W σm3 I m3 , where W σm3 is a trainable weight matrix, and I m3 is the third in-class signal, and i and j represent the i-th time point and the j-th time point respectively.
[0091] In step S4 of this implementation method, for the out-of-class signal, calculate the degree of change of the out-of-class signal at the i-th time point and the j-th time point as the out-of-class time correlation degree. Specifically, it includes:
[0092]
[0093] where TC n is the out-of-class time correlation degree at the i-th time point, and σ n = W n I n , where W n is a trainable weight matrix, and I n is the out-of-class signal.
[0094] In step S5 of this implementation method, calculate for each in-class signal:
[0095] For the first in-class signal, calculate the in-class query vector, in-class key vector, and in-class value vector at the i-th time point. According to the in-class query vector, in-class key vector, and in-class value vector, obtain the in-class information correlation degree corresponding to this in-class signal, and calculate the in-class information correlation degrees of each in-class signal;
[0096]
[0097] where IC m1 is the in-class information correlation degree at the i-th time point, Q m1 = W Qm1 I m1 , V m1 = W Vm1 I m1 , K m1 = W Km1 I m1 , W Qm1 , W Vm1 and W Km1 are the corresponding weight matrices respectively, I m1 is the first in-class signal, d k is the feature dimension, and Q im1 , K im1 , V im1 are the in-class query vector, in-class key vector, and in-class value vector respectively.
[0098] For the second intra-class signal, calculate the intra-class query vector, intra-class key vector, and intra-class value vector at the \(i\)-th time point. Based on the intra-class query vector, intra-class key vector, and intra-class value vector, obtain the intra-class information correlation degree corresponding to this intra-class signal, and calculate the intra-class information correlation degrees of each intra-class signal;
[0099]
[0100] where, \(IC\) m2 is the intra-class information correlation degree at the \(i\)-th time point, \(Q\) m2 = \(W\) Qm2 I m2 , \(V\) m2 = \(W\) Vm2 I m2 , \(K\) m2 = \(W\) Km2 I m2 , \(W\) Qm2 , \(W\) Vm2 and \(W\) Km2 are the corresponding weight matrices respectively, \(I\) m2 is the second intra-class signal, \(d\) k is the feature dimension, \(Q\) im2 , \(K\) im2 , \(V\) im2 are the intra-class query vector, intra-class key vector, and intra-class value vector respectively.
[0101] For the third intra-class signal, calculate the intra-class query vector, intra-class key vector, and intra-class value vector at the \(i\)-th time point. Based on the intra-class query vector, intra-class key vector, and intra-class value vector, obtain the intra-class information correlation degree corresponding to this intra-class signal, and calculate the intra-class information correlation degrees of each intra-class signal;
[0102]
[0103] where, \(IC\) m3 is the intra-class information correlation degree at the \(i\)-th time point, \(Q\) m3 = \(W\) Qm3 I m3 , \(V\) m3 = \(W\) Vm3 I m3 , \(K\) m3 = \(W\) Km3 I m3 , \(W\) Qm3 , \(W\) Vm3 and \(W\) Km3 are the corresponding weight matrices respectively, \(I\) m3 is the third intra-class signal, \(d\) k is the feature dimension, \(Q\) im3 , \(K\) im3 , \(V\) im3They are the in-class query vector, in-class key vector, and in-class value vector respectively.
[0104] In step S5 of this implementation method, for out-of-class signals, calculate the out-of-class query vector, out-of-class key vector, and out-of-class value vector at the i-th time point, and obtain the out-of-class information correlation degree according to the out-of-class query vector, out-of-class key vector, and out-of-class value vector. Specifically, it includes:
[0105] Calculate the in-class query vector, in-class key vector, and in-class value vector at the i-th time point, obtain the in-class information correlation degree corresponding to this out-of-class signal according to the out-of-class query vector, in-class key vector, and in-class value vector, and calculate the in-class information correlation degrees of each out-of-class signal;
[0106]
[0107] Among them, IC n represents the in-class information correlation degree at the i-th time point, Q n = W Qn I n , V n = W Vn I n , K n = W Kn I n , W Qn , W Vn and W Kn are the corresponding weight matrices respectively, I n is the out-of-class signal, d k is the feature dimension, Q in , K in , V in are the in-class query vector, in-class key vector, and in-class value vector respectively.
[0108] In step S6 of this implementation method, perform anomaly score calculation, including:
[0109] Calculate the temporal JS divergence between each in-class temporal correlation degree and the out-of-class temporal correlation degree respectively, and obtain the first JS divergence after superimposing the obtained multiple temporal JS divergences; calculate the information JS divergence between each in-class information correlation degree and the out-of-class information correlation degree respectively, and obtain the second JS divergence after superimposing the obtained multiple information JS divergences; obtain the final anomaly score according to the first JS divergence and the second JS divergence. Specifically, it includes:
[0110]
[0111] Among them, TC m represents the m-th in-class temporal correlation degree, TC n represents the out-of-class temporal correlation degree, IC mRepresents the within-class information correlation degree of the m-th class, IC n Represents the out-of-class information correlation degree, JS(TC m ||IC n ) represents the JS divergence of the m-th time, IC m ||IC n Represents the JS divergence of the m-th information, and M represents the total number of signals.
[0112] Embodiment 2:
[0113] As Figure 2 shown, this implementation provides a vehicle scale signal anomaly recognition system based on historical weighing signals, including:
[0114] A signal acquisition unit, configured to: acquire the current weighing signal of the vehicle scale and multiple historical weighing signals. The specific working process is as described in the process of step S1 in Embodiment 1, which will not be elaborated here;
[0115] A within-class signal division unit, configured to: separately perform signal division on the current weighing signal and the historical weighing signals to obtain the within-class signals corresponding to the current weighing signal and the historical weighing signals. The specific working process is as described in the process of step S2 in Embodiment 1, which will not be elaborated here;
[0116] An out-of-class signal division unit, configured to: extract signal segments from the current weighing signal and the historical weighing signals, and combine the signal segments to obtain out-of-class signals. The specific working process is as described in the process of step S3 in Embodiment 1, which will not be elaborated here;
[0117] A time correlation degree calculation unit, configured to: calculate the within-class time correlation degree based on the within-class signals to obtain the within-class time correlation degree, and calculate the out-of-class time correlation degree based on the out-of-class signals to obtain the out-of-class time correlation degree. The specific working process is as described in the process of step S4 in Embodiment 1, which will not be elaborated here;
[0118] An information correlation degree calculation unit, configured to: calculate the within-class information correlation degree based on the within-class signals to obtain the within-class information correlation degree, and calculate the out-of-class information correlation degree based on the out-of-class signals to obtain the out-of-class information correlation degree. The specific working process is as described in the process of step S5 in Embodiment 1, which will not be elaborated here;
[0119] An anomaly determination unit, configured to: calculate an anomaly score based on the differences between the within-class time correlation degrees and the out-of-class time correlation degrees, and the differences between the within-class information correlation degrees and the out-of-class information correlation degrees. When the anomaly score is greater than a set threshold, it is determined that there is an abnormal signal; otherwise, it is determined that there is no abnormal signal. The specific working process is as described in the process of step S6 in Embodiment 1, which will not be elaborated here.
[0120] It can be understood that each of the above units can be separately or all combined into one or several other units to form, or some of them can be further split into multiple smaller units in terms of function to form, which can achieve the same operation without affecting the realization of the technical effects of the embodiments of the present application. The above units are divided based on logical functions. In practical applications, the function of one unit can also be realized by multiple units, or the functions of multiple units can be realized by one unit. In other embodiments of the present application, the system can also include other units. In practical applications, these functions can also be assisted by other units and can be realized by the cooperation of multiple units.
[0121] According to another embodiment of the present application, the system described in this embodiment can be constructed, and the method of Embodiment 1 of the present application can be realized by running a computer program (including program code) capable of executing the respective steps involved in the corresponding method described in Embodiment 1 on a general computing device such as a computer including processing elements and storage elements such as a central processing unit (CPU), a random access memory (RAM), and a read-only memory (ROM). The computer program can be recorded on, for example, a computer-readable recording medium, loaded into the above computing device through the computer-readable recording medium, and run therein.
[0122] Embodiment 3:
[0123] As Figure 3 shown, this implementation provides an electronic device, which includes a processor 1001, a communication interface 1002, and a computer-readable storage medium 1003. Among them, the processor 1001, the communication interface 1002, and the computer-readable storage medium 1003 can be connected through a bus or other means.
[0124] Among them, the communication interface 1002 is used to receive and send data. The computer-readable storage medium 1003 can be stored in the memory of the electronic device. The computer-readable storage medium 1003 is used to store a computer program, and the computer program includes program instructions. The processor 1001 is used to execute the program instructions stored in the computer-readable storage medium 1003.
[0125] The processor 1001 (or CPU (Central Processing Unit, central processor)) is the computing core and control core of the electronic device, and is adapted to implement one or more instructions, specifically adapted to load and execute one or more instructions to thereby implement the corresponding method flow or corresponding function.
[0126] The processor 1001 is configured to execute the following process:
[0127] Obtain the current weighing signal and multiple historical weighing signals of the weighbridge. The specific working process is shown in the process of step S1 in Embodiment 1, which will not be elaborated here;
[0128] Separate signal partitioning is performed on the current weighing signal and the historical weighing signals respectively to obtain the in-class signals corresponding to the current weighing signal and the historical weighing signals. The specific working process is shown in the process of step S2 in Embodiment 1, which will not be elaborated here;
[0129] Signal segment extraction is performed on the current weighing signal and the historical weighing signals, and the signal segments are combined to obtain out-of-class signals. The specific working process is shown in the process of step S3 in Embodiment 1, which will not be elaborated here;
[0130] According to the in-class signals, in-class time correlation degree calculation is performed to obtain the in-class time correlation degree. According to the out-of-class signals, out-of-class time correlation degree calculation is performed to obtain the out-of-class time correlation degree. The specific working process is shown in the process of step S4 in Embodiment 1, which will not be elaborated here;
[0131] According to the in-class signals, in-class information correlation degree calculation is performed to obtain the in-class information correlation degree. According to the out-of-class signals, out-of-class information correlation degree calculation is performed to obtain the out-of-class information correlation degree. The specific working process is shown in the process of step S5 in Embodiment 1, which will not be elaborated here;
[0132] According to the differences between the in-class time correlation degrees and the out-of-class time correlation degrees, and the differences between the in-class information correlation degrees and the out-of-class information correlation degrees, abnormal score calculation is performed. When the abnormal score is greater than the set threshold, it is determined that there is an abnormal signal; otherwise, it is determined that there is no abnormal signal. The specific working process is shown in the process of step S6 in Embodiment 1, which will not be elaborated here.
[0133] Embodiment 4:
[0134] This implementation provides a computer-readable storage medium (Memory). A computer-readable storage medium is a memory device in an electronic device for storing programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the electronic device and, of course, the extended storage medium supported by the electronic device. The computer-readable storage medium provides a storage space, and this storage space stores the processing system of the electronic device.
[0135] Moreover, one or more instructions suitable for being loaded and executed by a processor are stored in this storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory; optionally, it can also be at least one computer-readable storage medium located far from the aforementioned processor.
[0136] In one embodiment, one or more instructions are stored in the computer-readable storage medium; the processor loads and executes one or more instructions stored in the computer-readable storage medium to implement the following process:
[0137] Obtain the current weighing signal of the weighbridge and multiple historical weighing signals. The specific working process can be seen in the process of step S1 in Embodiment 1 and will not be elaborated here;
[0138] Separate the current weighing signal and the historical weighing signals individually to obtain the within-class signals corresponding to the current weighing signal and the historical weighing signals respectively. The specific working process can be seen in the process of step S2 in Embodiment 1 and will not be elaborated here;
[0139] Extract signal segments from the current weighing signal and the historical weighing signals, and combine the signal segments to obtain out-of-class signals. The specific working process can be seen in the process of step S3 in Embodiment 1 and will not be elaborated here;
[0140] According to the within-class signals, calculate the within-class time correlation degree to obtain the within-class time correlation degree. According to the out-of-class signals, calculate the out-of-class time correlation degree to obtain the out-of-class time correlation degree. The specific working process can be seen in the process of step S4 in Embodiment 1 and will not be elaborated here;
[0141] According to the within-class signals, calculate the within-class information correlation degree to obtain the within-class information correlation degree. According to the out-of-class signals, calculate the out-of-class information correlation degree to obtain the out-of-class information correlation degree. The specific working process can be seen in the process of step S5 in Embodiment 1 and will not be elaborated here;
[0142] According to the differences between each within-class time correlation degree and the out-of-class time correlation degree, and the differences between each within-class information correlation degree and the out-of-class information correlation degree, calculate the anomaly score. When the anomaly score is greater than the set threshold, it is determined that there is an abnormal signal; otherwise, it is determined that there is no abnormal signal. The specific working process can be seen in the process of step S6 in Embodiment 1 and will not be elaborated here.
[0143] Embodiment 5:
[0144] This implementation provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, causing the electronic device to perform the following processes:
[0145] Obtain the current weighing signal of the weighbridge and multiple historical weighing signals. The specific working process can be seen in the process of step S1 in Embodiment 1, which will not be elaborated here;
[0146] Separate the current weighing signal and the historical weighing signals respectively for signal division, and obtain the in-class signals corresponding to the current weighing signal and the historical weighing signals respectively. The specific working process can be seen in the process of step S2 in Embodiment 1, which will not be elaborated here;
[0147] Extract signal segments from the current weighing signal and the historical weighing signals, and combine the signal segments to obtain out-of-class signals. The specific working process can be seen in the process of step S3 in Embodiment 1, which will not be elaborated here;
[0148] According to the in-class signals, calculate the in-class time correlation degree to obtain the in-class time correlation degree. According to the out-of-class signals, calculate the out-of-class time correlation degree to obtain the out-of-class time correlation degree. The specific working process can be seen in the process of step S4 in Embodiment 1, which will not be elaborated here;
[0149] According to the in-class signals, calculate the in-class information correlation degree to obtain the in-class information correlation degree. According to the out-of-class signals, calculate the out-of-class information correlation degree to obtain the out-of-class information correlation degree. The specific working process can be seen in the process of step S5 in Embodiment 1, which will not be elaborated here;
[0150] According to the differences between the in-class time correlation degrees and the out-of-class time correlation degrees, and the differences between the in-class information correlation degrees and the out-of-class information correlation degrees, calculate the anomaly score. When the anomaly score is greater than the set threshold, it is determined that there is an abnormal signal; otherwise, it is determined that there is no abnormal signal. The specific working process can be seen in the process of step S6 in Embodiment 1, which will not be elaborated here.
[0151] Those of ordinary skill in the art can realize that, combining the units and algorithm steps of the examples described in the embodiments disclosed in this application, can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but this implementation should not be considered to exceed the scope of this application.
[0152] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions according to the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired manner (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data processing device such as a server or a data center that includes one or more integrated available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)), etc.
[0153] The foregoing is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.
Claims
1. A method for identifying abnormal vehicle scale signals based on historical weighing signals, characterized in that: The process includes: Obtain the current weighing signal and multiple historical weighing signals of the truck scale; Performing signal division on the current weighing signal and the historical weighing signal separately to obtain intra-class signals corresponding to the current weighing signal and the historical weighing signal respectively; Extracting signal segments from the current weighing signal and the historical weighing signal, and combining the signal segments to obtain an out-of-class signal; According to the in-class signal, the in-class time correlation degree is calculated to obtain the in-class time correlation degree, and according to the out-class signal, the out-class time correlation degree is calculated to obtain the out-class time correlation degree; According to the in-class signal, calculating the in-class information correlation degree, obtaining the in-class information correlation degree; according to the out-class signal, calculating the out-class information correlation degree, obtaining the out-class information correlation degree; Anomaly scores are calculated based on the difference between the time correlation within each class and the time correlation outside the class, as well as the difference between the information correlation within each class and the information correlation outside the class. When the anomaly score is greater than the set threshold, it is determined that an abnormal signal exists; otherwise, it is determined that no abnormal signal exists.
2. The method for identifying abnormal vehicle scale signals based on historical weighing signals according to claim 1, characterized in that: Extracting signal segments from the current weighing signal and the historical weighing signal, and combining the signal segments, including: The current weighing signal and the historical weighing signal of the same length are placed one above the other, and the current weighing signal and the historical weighing signal are intercepted within a set time window to obtain multiple signal segments, and the obtained signal segments are combined to obtain an out-of-class signal.
3. The method for identifying abnormal vehicle scale signals based on historical weighing signals according to claim 1, characterized in that: For any intra-class signal, calculate the degree of change of the intra-class signal at the i-th time point and the j-th time point as the intra-class temporal correlation of the intra-class signal, and calculate the intra-class temporal correlation of each intra-class signal; For out-of-class signals, the degree of change of out-of-class signals between the i-th time point and the j-th time point is calculated as the out-of-class time correlation.
4. The method for identifying abnormal vehicle scale signals based on historical weighing signals according to claim 1, characterized in that: For any intra-class signal, calculate the intra-class query vector, the intra-class key vector and the intra-class value vector at the i-th time point, obtain the intra-class information correlation corresponding to the intra-class signal according to the intra-class query vector, the intra-class key vector and the intra-class value vector, and calculate the intra-class information correlation of each intra-class signal; For the out-of-class signal, the out-of-class query vector, the out-of-class key vector and the out-of-class value vector at the i-th time point are calculated, and the out-of-class information relevance is obtained according to the out-of-class query vector, the out-of-class key vector and the out-of-class value vector.
5. The method for identifying abnormal vehicle scale signals based on historical weighing signals according to claim 1, characterized in that: Perform anomaly score calculation, including: Calculate the time JS divergence of each intra-class time correlation degree and the out-class time correlation degree respectively, and superimpose the obtained multiple time JS divergences to obtain a first JS divergence; Calculate the information JS divergence of each intra-class information correlation degree and the out-class information correlation degree respectively, and superimpose the obtained multiple information JS divergences to obtain a second JS divergence; The final anomaly score is obtained based on the first JS divergence and the second JS divergence.
6. The method for identifying abnormal vehicle scale signals based on historical weighing signals according to claim 5, characterized in that: Among them, TC m Represents the temporal correlation within the mth class, TC n Represents the out-of-class temporal correlation, IC m Represents the information correlation within the mth class, IC n Represents the correlation of out-of-class information, JS(TC m ||TC n ) represents the JS divergence at the mth time, IC m ||IC n represents the mth information JS divergence, and M represents the total number of signals.
7. A vehicle scale signal abnormality recognition system based on historical weighing signals, characterized in that: include: The signal acquisition unit is configured to: acquire a current weighing signal and a plurality of historical weighing signals of the vehicle scale; The intra-class signal division unit is configured to: perform signal division on the current weighing signal and the historical weighing signal separately, and obtain intra-class signals corresponding to the current weighing signal and the historical weighing signal respectively; The out-of-class signal division unit is configured to: extract signal segments from the current weighing signal and the historical weighing signal, and combine the signal segments to obtain an out-of-class signal; The time correlation calculation unit is configured to: perform time correlation calculation within the class according to the in-class signal to obtain the in-class time correlation, and perform time correlation calculation outside the class according to the out-class signal to obtain the out-class time correlation; The information correlation calculation unit is configured to: calculate the information correlation within the class according to the in-class signal to obtain the in-class information correlation, and calculate the information correlation outside the class according to the out-class signal to obtain the out-class information correlation; The anomaly determination unit is configured to calculate an anomaly score based on the difference between the time correlation within each class and the time correlation outside the class, and the difference between the information correlation within each class and the information correlation outside the class. When the anomaly score is greater than a set threshold, it is determined that an abnormal signal exists; otherwise, it is determined that no abnormal signal exists.
8. A computer device, characterized in that: include: a processor and a computer readable storage medium; a processor adapted to execute a computer program; A computer-readable storage medium having a computer program stored therein, wherein the computer program, when executed by the processor, implements the method for identifying abnormal vehicle scale signals based on historical weighing signals as described in any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor and executing the method for identifying abnormal vehicle scale signals based on historical weighing signals as described in any one of claims 1 to 6.
10. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, it implements the method for identifying abnormal vehicle scale signals based on historical weighing signals as described in any one of claims 1 to 6.