CN117591986A - Real-time automobile data processing method based on artificial intelligence - Google Patents
Real-time automobile data processing method based on artificial intelligence Download PDFInfo
- Publication number
- CN117591986A CN117591986A CN202410071350.0A CN202410071350A CN117591986A CN 117591986 A CN117591986 A CN 117591986A CN 202410071350 A CN202410071350 A CN 202410071350A CN 117591986 A CN117591986 A CN 117591986A
- Authority
- CN
- China
- Prior art keywords
- pressure data
- upper limit
- value
- data
- limit value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 20
- 238000003672 processing method Methods 0.000 title claims abstract description 18
- 230000005856 abnormality Effects 0.000 claims abstract description 84
- 238000001514 detection method Methods 0.000 claims abstract description 73
- 238000011156 evaluation Methods 0.000 claims abstract description 55
- 230000002159 abnormal effect Effects 0.000 claims abstract description 52
- 238000000034 method Methods 0.000 claims abstract description 13
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000000694 effects Effects 0.000 description 9
- 238000004458 analytical method Methods 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M17/00—Testing of vehicles
- G01M17/007—Wheeled or endless-tracked vehicles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/15—Correlation function computation including computation of convolution operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Algebra (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Measuring Fluid Pressure (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to an artificial intelligence-based real-time automobile data processing method. The method comprises the following steps: acquiring a pressure data sequence of an automobile tire and a temperature data sequence of an automobile engine; the method comprises the steps of presetting an upper limit value, calculating initial abnormality degree of each pressure data in each pressure data sequence under the upper limit value, and obtaining final abnormality degree of each pressure data under the upper limit value according to the initial abnormality degree and the temperature data sequence; obtaining an abnormality detection evaluation value of each pressure data under an upper limit value according to the final abnormality degree; obtaining an optimal K value of each pressure data according to the abnormal detection evaluation value; and carrying out anomaly detection on the pressure data according to the optimal K value of each piece of pressure data to obtain anomaly data. By setting an appropriate K value for each pressure data, the accuracy of pressure data abnormality detection is improved.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to an artificial intelligence-based real-time automobile data processing method.
Background
The vehicle can have potential safety hazards due to faults such as abrasion and aging in the use process, and the operation data of key parts of the vehicle are required to be detected in real time by utilizing the sensor, so that the operation state condition of the vehicle can be mastered in time, driving adjustment can be carried out according to the operation state condition of the vehicle, accidents on roads are effectively avoided, and the safety of drivers and passengers is ensured. Abnormal operation data is generated when an abnormality exists in the operation state of a key part of the automobile. Therefore, in order to analyze an abnormal condition of the running state of the automobile, it is necessary to extract abnormal data from the running data.
The LOF algorithm is a commonly used anomaly analysis method that generally sets a K value based on human experience, and the method sets the same K value for each data. The K value in the LOF algorithm affects the sensitivity of anomaly detection, i.e., when the K value is set too small, more anomaly data can be identified, but normal data is also easily misjudged as anomaly data. When the K value is set too large, however, the abnormal data is recognized less, but it is not easy to misjudge the normal data as the abnormal data. The suitability of the K value setting may affect the accuracy of anomaly detection. Since the variation of the operation data at each timing of the critical components of the vehicle is inconsistent, such variation difference may cause a difference in the distance between the operation data. Wherein, the inconsistent distance between some operation data and other operation data is caused by the abnormality of key parts of the automobile, and the inconsistent distance between some operation data and other operation data is caused by the interference of other factors. And for those operating data due to interference of other factors, if the K value is set too small, it is easy to determine it as abnormal data. And for those motion data caused by abnormality of critical components of the vehicle, if an excessively large K value is set, it is easy to determine it as normal data. Thus, a fixed K value is set for each data, which is not suitable for detecting abnormal data in the running data of the automobile. Therefore, how to accurately detect abnormal data in the operation data by setting an appropriate K value for each operation data becomes a problem to be solved.
Disclosure of Invention
In order to solve the technical problems, the invention provides an artificial intelligence-based real-time automobile data processing method, which adopts the following technical scheme:
acquiring a pressure data sequence of an automobile tire and a temperature data sequence of an automobile engine;
presetting an upper limit value, taking the upper limit value as neighborhood data of each pressure data in the pressure data sequence under the upper limit value, and obtaining initial abnormality degree of each pressure data under the upper limit value according to a difference value between the neighborhood data of each pressure data under the upper limit value; according to the initial abnormality degree of each pressure data under the upper limit value, the correlation of the pressure data sequence and the temperature data sequence, and the neighborhood data of each pressure data under the upper limit value, the final abnormality degree of each pressure data under the upper limit value is obtained;
obtaining an abnormality detection evaluation value of each pressure data under the upper limit value according to the difference value of the final abnormality degree between the neighborhood data of each pressure data under the upper limit value; adjusting the K value according to the abnormal detection evaluation value of each pressure data under the upper limit value to obtain an optimal K value of each pressure data;
and carrying out anomaly detection on the pressure data according to the optimal K value of each piece of pressure data to obtain anomaly data.
Preferably, the step of taking the upper limit value as the neighborhood data of each pressure data in the pressure data sequence under the upper limit value includes the following specific steps:
and based on an LOF algorithm, taking the upper limit value as the K value of each piece of pressure data, acquiring the pressure data of each piece of pressure data in the K distance neighborhood, and recording the pressure data as neighborhood data of each piece of pressure data under the upper limit value.
Preferably, the obtaining the initial abnormality degree of each pressure data under the upper limit value according to the difference value between the neighborhood data of each pressure data under the upper limit value includes the following specific steps:
obtaining local reachable density of each pressure data according to neighborhood data of each pressure data under the upper limit value, and recording the local reachable density of each pressure data under the upper limit value;
and obtaining the initial abnormality degree of each pressure data under the upper limit value according to the local reachable density of each pressure data.
Preferably, the initial abnormality degree of each pressure data under the upper limit value is obtained according to the local reachable density of each pressure data, and the specific calculation formula is:
wherein,local reachable density of the jth neighborhood data representing the ith pressure data below an upper limit,/->Representing the locally attainable density of the ith pressure data below the upper limit, < >>Euclidean distance of the ith pressure data and the jth neighborhood data below the upper limit value,/>Representing the number of neighborhood data of the ith pressure data below the upper limit, +.>The initial abnormality degree of the ith pressure data at the upper limit value is indicated.
Preferably, the step of obtaining the final abnormality degree of each pressure data under the upper limit value according to the initial abnormality degree of each pressure data under the upper limit value, the correlation between the pressure data sequence and the temperature data sequence, and the neighborhood data of each pressure data under the upper limit value comprises the following specific steps:
obtaining a first correlation coefficient and a second correlation coefficient according to the pressure data sequence and the temperature data sequence; obtaining neighborhood correlation of each pressure data under the upper limit value according to the neighborhood data of each pressure data under the upper limit value;
the calculation method for obtaining the final abnormality degree of each pressure data according to the first correlation coefficient, the second correlation coefficient and the neighborhood correlation of each pressure data comprises the following steps:
wherein,a second correlation coefficient representing the ith pressure data, +.>A first correlation coefficient representing the ith pressure data,/->Representing the neighborhood correlation of the ith pressure data below an upper limit,/for the pressure data>The initial degree of abnormality of the ith pressure data under the upper limit value, +.>Represents an exponential function based on natural constants, < ->The final degree of abnormality of the ith pressure data at the upper limit value is indicated.
Preferably, the first correlation coefficient and the second correlation coefficient are obtained according to the pressure data sequence and the temperature data sequence; obtaining the neighborhood correlation of each pressure data under the upper limit value according to the neighborhood data of each pressure data under the upper limit value, comprising the following specific steps:
recording any one pressure data as target pressure data, recording rank correlation coefficients of a pressure data sequence and a temperature data sequence as first correlation coefficients of target pressure data, recording temperature data at the same time as the target pressure data as target temperature data, recording a data sequence obtained by removing the target pressure data from the pressure data sequence as a residual sequence of the target pressure data, recording a data sequence obtained by removing the target temperature data from the temperature data sequence as a residual sequence of the target temperature data, calculating rank correlation coefficients of the residual sequence of the target pressure data and the residual sequence of the target temperature data, recording the rank correlation coefficients of the residual sequence of the target pressure data and the second correlation coefficients of the target pressure data, and obtaining the first correlation coefficients and the second correlation coefficients of each pressure data;
obtaining neighborhood data of target pressure data under an upper limit value, taking temperature data which is the same as the neighborhood data of the target pressure data under the upper limit value as the neighborhood data of the target temperature data under the upper limit value, calculating rank correlation coefficients of the target pressure data and each neighborhood data of the target temperature data under the upper limit value, recording the accumulated sum of the rank correlation coefficients of all the neighborhood data of the target pressure data and the target temperature data under the upper limit value as the neighborhood correlation of the target pressure data under the upper limit value, and obtaining the neighborhood correlation of each pressure data under the upper limit value.
Preferably, the method for obtaining the abnormality detection evaluation value of each pressure data under the upper limit value according to the difference value of the final abnormality degree between the neighborhood data of each pressure data under the upper limit value comprises the following specific steps:
obtaining a plurality of combinations of each pressure under the upper limit value according to the neighborhood data of each pressure under the upper limit value; obtaining a neighborhood radius, an increment neighborhood radius and a decrement neighborhood radius of each pressure data under an upper limit value;
according to a plurality of combinations of each pressure data under the upper limit value, the neighborhood radius, the increment neighborhood radius and the decrement neighborhood radius of each pressure data under the upper limit value are used for obtaining the abnormal detection evaluation value of each pressure data under the upper limit value, and the calculation method comprises the following steps:
wherein,incremental neighborhood radius representing the ith pressure data below the upper limit, +.>Neighborhood radius representing the ith pressure data below the upper limit, +.>A decrement neighborhood radius representing the ith pressure data below the upper limit, +.>The ith pressure data is represented by the +.>One of the neighborhood data in the combinations, +.>The ith pressure data is represented by the +.>Another neighborhood data in the combination, +.>Represents the number of combinations of the ith pressure data below the upper limit,/->An abnormality detection evaluation value indicating that the ith pressure data is below the upper limit value.
Preferably, the neighborhood data under the upper limit value according to each pressure data obtains a plurality of combinations under the upper limit value; the method for acquiring the neighborhood radius, the increment neighborhood radius and the decrement neighborhood radius of each pressure data under the upper limit value comprises the following specific steps:
combining any two neighborhood data of each pressure data under the upper limit value to obtain a plurality of combinations of each pressure data under the upper limit value; acquiring each pressure dataThe radius of the M1 distance neighborhood of the pressure data is recorded as the neighborhood radius of each pressure data under the upper limit value; acquiring each pressure dataThe radius of the distance neighborhood is recorded as the increment neighborhood radius of each pressure data under the upper limit value; acquiring +.>The radius of the distance neighborhood is recorded as a decrement neighborhood radius of each pressure data under an upper limit value, and M1 represents a preset upper limit value.
Preferably, the adjusting the K value according to the abnormality detection evaluation value of each pressure data under the upper limit value to obtain the optimal K value of each pressure data includes the specific steps of:
comparing the abnormality detection evaluation value of each pressure data under the upper limit value with a preset index threshold Y1, and taking the upper limit value as the optimal K value of the pressure data when the abnormality detection evaluation value of each pressure data under the upper limit value is larger than the preset index threshold;
when the abnormal detection evaluation value of each pressure data under the upper limit value is smaller than or equal to a preset index threshold value, marking the difference value between the upper limit value and one as a first adjustment value, acquiring the abnormal detection evaluation value of each pressure data under the first adjustment value, comparing the abnormal detection evaluation value of each pressure data under the first adjustment value with a preset index threshold value Y1, and when the abnormal detection evaluation value of each pressure data under the first adjustment value is larger than the preset index threshold value, taking the first adjustment value as the optimal K value of the pressure data;
when the abnormal detection evaluation value of each pressure data under the first adjustment value is smaller than or equal to the preset index threshold value, the difference value between the first adjustment value and one is recorded as a second adjustment value, the abnormal detection evaluation value of each pressure data under the second adjustment value is obtained, the abnormal detection evaluation value of each pressure data under the second adjustment value is compared with the preset index threshold value Y1, when the abnormal detection evaluation value of each pressure data under the second adjustment value is larger than the preset index threshold value, the second adjustment value is used as the optimal K value of the pressure data, and the like until the abnormal detection evaluation value of each pressure data under the adjustment value is larger than the preset index threshold value, and the optimal K value of each pressure data is obtained.
Preferably, the anomaly detection is performed on the pressure data according to the optimal K value of each pressure data to obtain anomaly data, and the specific steps include:
and setting the K value of each pressure data to be equal to the optimal K value, and analyzing all the pressure data by using an LOF algorithm based on the K value of each pressure data to obtain abnormal data.
The invention has the following beneficial effects:
acquiring a temperature data sequence and a pressure data sequence, presetting an upper limit value, obtaining the final abnormality degree of each pressure data under the upper limit value by eliminating interference of pressure data fluctuation caused by temperature data fluctuation on abnormality analysis of the pressure data, wherein the final abnormality degree of each pressure data under the upper limit value can reflect the possibility that each pressure data is abnormal when the upper limit value is taken as a K value, constructing a judging index capable of measuring the abnormality detection effect of each pressure data when the upper limit value is taken as the K value according to the final abnormality degree, obtaining an abnormality detection evaluation value of each pressure data under the upper limit value, reflecting the abnormality detection effect of each pressure data when the upper limit value is taken as the K value according to the abnormality detection evaluation value of each pressure data under the upper limit value, and adjusting the K value according to the abnormality detection evaluation value of each pressure data to obtain the optimal K value of each pressure data. Thereby improving the accuracy of pressure data anomaly detection by setting an appropriate K value for each pressure data.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of an artificial intelligence based real-time processing method for automobile data according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description is given below of the specific implementation, structure, characteristics and effects of the artificial intelligence-based real-time processing method for automobile data according to the invention in combination with the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
An embodiment of an automobile data real-time processing method based on artificial intelligence comprises the following steps:
the following specifically describes a specific scheme of the artificial intelligence-based real-time automobile data processing method provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of an artificial intelligence based real-time processing method for vehicle data according to an embodiment of the present invention is shown, where the method includes:
s001: a pressure data sequence of an automobile tire and a temperature data sequence of an engine are acquired.
Specifically, every 1 second, the pressure data of the automobile tire and the temperature data of the engine are collected by using the pressure sensor and the temperature sensor for N times. And arranging the acquired N pieces of pressure data according to time sequence to obtain a pressure data sequence, and arranging the acquired N pieces of temperature data according to time sequence to obtain a temperature data sequence.
In this embodiment, N is 1000 as an example, and other values may be adopted in other embodiments, and the present embodiment is not particularly limited.
S002: the method comprises the steps of presetting an upper limit value, obtaining neighborhood data of each pressure data under the upper limit value, obtaining initial abnormality degree of each pressure data under the upper limit value according to the neighborhood data of each pressure data under the upper limit value, and obtaining final abnormality degree of each pressure data under the upper limit value according to the initial abnormality degree of each pressure data under the upper limit value.
In order to determine the K value of each pressure data, it is necessary to evaluate the abnormality detection effect of each pressure data at a different K value.
Specifically, an upper limit value M1 is preset. In this embodiment, M1 is taken as 100 as an example, and other values may be taken in other embodiments, and the present embodiment is not particularly limited. And based on an LOF algorithm, taking the upper limit value as the K value of each piece of pressure data, acquiring the pressure data of each piece of pressure data in the K distance neighborhood, and recording the pressure data as neighborhood data of each piece of pressure data under the upper limit value. Based on the LOF algorithm, local reachable density of each pressure data is obtained according to neighborhood data of each pressure data under the upper limit value and recorded as the local reachable density of each pressure data under the upper limit value. The method for obtaining the K-distance neighborhood and the local reachable density in the LOF algorithm is the prior art, and will not be described here.
The abnormality detection effect of each pressure data at the upper limit value is first evaluated. To evaluate the abnormality detection effect of each pressure data at the upper limit value, it is necessary to analyze the abnormality of each pressure data at the upper limit value.
Further, since the similarity between the pressure data and the pressure data closer to each other is high, it is necessary to refer to more information of the pressure data closer to each other when performing the anomaly analysis.
Further, the calculation method for obtaining the initial abnormality degree of each pressure data under the upper limit value according to the local reachable density of each pressure data under the upper limit value comprises the following steps:
wherein,local reachable density of the jth neighborhood data representing the ith pressure data below an upper limit,/->Representing the locally attainable density of the ith pressure data below the upper limit, < >>Euclidean distance of the ith pressure data and the jth neighborhood data below the upper limit value,/>The difference between the local reachable density of the jth neighborhood data and the local reachable density of the ith pressure data below the upper limit value is reflected, and the larger the value is, the less the density of data distributed around the ith pressure data is than the density of data distributed around the neighborhood data, so that the more abnormal the ith pressure data is.Representing the number of neighborhood data for which the ith pressure data is below the upper limit. />The initial abnormality degree of the ith pressure data at the upper limit value is indicated.
The tire pressure may vary due to acceleration or deceleration of the vehicle. Whereas acceleration and deceleration of the vehicle are related to the engine operating state. The operating state of the engine may be manifested by the engine temperature. The tire pressure of the vehicle will thus be related to the engine temperature. At the same time, the pressure of the tire varies due to the variation of the engine state, which results in a large difference between the pressure data of the tire and other pressure data. When the pressure data of the tire is greatly different from other pressure data, the tire is easily misjudged as abnormal data. In order to prevent such abnormal erroneous judgment, it is necessary to eliminate the disturbance of the tire pressure fluctuation caused by the engine state fluctuation.
Further, any one pressure data is recorded as target pressure data, a rank correlation coefficient of a pressure data sequence and a temperature data sequence is recorded as a first correlation coefficient of the target pressure data, the temperature data at the same time as the target pressure data is recorded as target temperature data, a data sequence obtained by removing the target pressure data from the pressure data sequence is recorded as a residual sequence of the target pressure data, and a data sequence obtained by removing the target temperature data from the temperature data sequence is recorded as a residual sequence of the target temperature data. And calculating rank correlation coefficients of the residual sequence of the target pressure data and the residual sequence of the target temperature data, and recording the rank correlation coefficients as second correlation coefficients of the target pressure data. And similarly, obtaining a first correlation coefficient and a second correlation coefficient of each pressure data.
And acquiring neighborhood data of the target pressure data under the upper limit value, and taking the temperature data which is the same as the neighborhood data of the target pressure data under the upper limit value as the neighborhood data of the target temperature data under the upper limit value. And calculating rank correlation coefficients of each neighborhood data of the target pressure data and the target temperature data under the upper limit value, and recording an accumulated sum of the rank correlation coefficients of all the neighborhood data of the target pressure data and the target temperature data under the upper limit value as neighborhood correlation of the target pressure data under the upper limit value. And similarly, obtaining the neighborhood correlation of each pressure data under the upper limit value.
The calculation method for obtaining the final abnormality degree of each pressure data according to the first correlation coefficient, the second correlation coefficient and the neighborhood correlation of each pressure data comprises the following steps:
wherein,a second correlation coefficient representing the ith pressure data, +.>A first correlation coefficient representing the ith pressure data,/->Reflecting the ith pressureThe correlation of the pressure data and the temperature data after the removal of the force data, and the variation of the correlation of the pressure data and the temperature data before the removal, the larger the value is, the more the existence of the ith pressure data will cause the correlation of the pressure data and the temperature data to be reduced, and the existence of the ith pressure data will cause the correlation law of the pressure data and the temperature data to be poor, and the higher the possibility that the ith pressure data is abnormal data is>Representing the neighborhood correlation of the ith pressure data below an upper limit,/for the pressure data>Reflecting the correlation of the ith pressure data with the temperature data in the neighborhood, the larger the value is, the stronger the correlation of the ith pressure data with the temperature data in the neighborhood is, and thus the less likely the ith pressure data is abnormal data. />The initial degree of abnormality of the ith pressure data at the upper limit value, the larger the value is, the greater the degree of abnormality of the ith pressure data is. />An exponential function based on a natural constant is represented. />The final degree of abnormality of the ith pressure data at the upper limit value is indicated.
Thus, the final abnormality degree of each pressure data is obtained under the upper limit value, and the final abnormality degree is obtained based on the neighborhood data analysis of each pressure data under the upper limit value, and can reflect the abnormal condition of each pressure data. The abnormal condition of each pressure data under the upper limit value can be reflected well through the data.
S003: and obtaining an abnormality detection evaluation value of each pressure data under the upper limit value according to the final abnormality degree of each pressure data under the upper limit value, and obtaining an optimal K value of each pressure data according to the abnormality detection evaluation value of each pressure data under the upper limit value.
In order to obtain a proper K value for each piece of pressure data, it is necessary to evaluate the abnormality detection effect when the upper limit value is used as the K value for each piece of pressure data.
Specifically, any two neighborhood data of each pressure data under the upper limit value are combined to obtain a plurality of combinations of each pressure data under the upper limit value. Acquiring the radius of an M1 distance neighborhood of each pressure data, and recording the radius as the neighborhood radius of each pressure data under an upper limit value; acquiring each pressure dataThe radius of the distance neighborhood is recorded as the increment neighborhood radius of each pressure data under the upper limit value; acquiring +.>The radius from the neighborhood is noted as the decreasing neighborhood radius for each pressure data at the upper value. M1 represents a preset upper limit value.
The calculation method of the abnormality detection evaluation value for each pressure data under the upper limit value is:
wherein,incremental neighborhood radius representing the ith pressure data below the upper limit, +.>Neighborhood radius representing the ith pressure data below the upper limit, +.>A decrement neighborhood radius representing the ith pressure data below the upper limit, +.>Representing two adjacent pressure data below an upper limitThe larger the ratio of the difference of the neighborhood radii, the larger the value is, which indicates that the difference between the ith pressure data M1+1 and M1 is larger, and the difference between the ith pressure data M1 and M1-1 is smaller. The data in the M1 distance neighborhood of the ith pressure data is smaller in difference, and the data outside the M1 distance neighborhood is larger in difference from the data in the M1 neighborhood, so that the data in the M1 distance neighborhood of the ith pressure data possibly belong to the same type of data. The similarity between the same class of data is large. Therefore, when abnormality detection is performed, only the same type of data is generally referred to. When referring to multiple classes of data, it is easily disturbed by class differences. Thus(s)The larger the upper limit value is, the larger the degree of suitability as a K value is. />The ith pressure data is represented by the +.>One of the neighborhood data in the combinations, +.>The ith pressure data is represented by the +.>Another neighborhood data in the set.
Indicating the number of combinations of the ith pressure data below the upper limit. />The difference in degree of abnormality between neighborhood data obtained when the ith pressure data has an upper limit value as a K value is reflected. The greater the value, the better the difference between the normal data and the abnormal data can be distinguished when the ith pressure data has the upper limit value as the K value, thus the ith pressure numberWhen the upper limit value is set as the K value, the abnormality detection effect is good. />An abnormality detection evaluation value indicating that the ith pressure data is below the upper limit value.
In order to obtain an appropriate K value for each piece of pressure data, the K value for each piece of pressure data is adjusted according to the abnormality detection evaluation value for each piece of pressure data at the upper limit value, and thus an optimal K value for each piece of pressure data is obtained.
Further, the abnormality detection evaluation value of each pressure data under the upper limit value is compared with a preset index threshold value Y1, and when the abnormality detection evaluation value of each pressure data under the upper limit value is greater than the preset index threshold value, the upper limit value is taken as the optimal K value of the pressure data.
When the abnormal detection evaluation value of each pressure data under the upper limit value is smaller than or equal to a preset index threshold value, marking the difference value between the upper limit value and one as a first adjustment value, acquiring the abnormal detection evaluation value of each pressure data under the first adjustment value, comparing the abnormal detection evaluation value of each pressure data under the first adjustment value with a preset index threshold value Y1, and when the abnormal detection evaluation value of each pressure data under the first adjustment value is larger than the preset index threshold value, taking the first adjustment value as the optimal K value of the pressure data;
when the abnormal detection evaluation value of each pressure data under the first adjustment value is smaller than or equal to the preset index threshold value, the difference value between the first adjustment value and one is recorded as a second adjustment value, the abnormal detection evaluation value of each pressure data under the second adjustment value is obtained, the abnormal detection evaluation value of each pressure data under the second adjustment value is compared with the preset index threshold value Y1, when the abnormal detection evaluation value of each pressure data under the second adjustment value is larger than the preset index threshold value, the second adjustment value is used as the optimal K value of the pressure data, and the like until the abnormal detection evaluation value of each pressure data under the adjustment value is larger than the preset index threshold value, and the optimal K value of each pressure data is obtained.
In this embodiment, Y1 is taken as an example of 0.9, and other values may be taken in other embodiments, and the embodiment is not particularly limited.
S004: and performing anomaly detection based on the optimal K value of each pressure data to obtain anomaly data.
Specifically, the K value of each pressure data is set to be equal to the optimal K value, and all the pressure data are analyzed by using the LOF algorithm based on the K value of each pressure data to obtain abnormal data.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (10)
1. The real-time automobile data processing method based on artificial intelligence is characterized by comprising the following steps of:
acquiring a pressure data sequence of an automobile tire and a temperature data sequence of an automobile engine;
presetting an upper limit value, taking the upper limit value as neighborhood data of each pressure data in the pressure data sequence under the upper limit value, and obtaining initial abnormality degree of each pressure data under the upper limit value according to a difference value between the neighborhood data of each pressure data under the upper limit value; according to the initial abnormality degree of each pressure data under the upper limit value, the correlation of the pressure data sequence and the temperature data sequence, and the neighborhood data of each pressure data under the upper limit value, the final abnormality degree of each pressure data under the upper limit value is obtained;
obtaining an abnormality detection evaluation value of each pressure data under the upper limit value according to the difference value of the final abnormality degree between the neighborhood data of each pressure data under the upper limit value; adjusting the K value according to the abnormal detection evaluation value of each pressure data under the upper limit value to obtain an optimal K value of each pressure data;
and carrying out anomaly detection on the pressure data according to the optimal K value of each piece of pressure data to obtain anomaly data.
2. The real-time processing method of automobile data based on artificial intelligence according to claim 1, wherein the step of regarding the upper limit value as the neighborhood data of each pressure data in the pressure data sequence under the upper limit value comprises the following specific steps:
and based on an LOF algorithm, taking the upper limit value as the K value of each piece of pressure data, acquiring the pressure data of each piece of pressure data in the K distance neighborhood, and recording the pressure data as neighborhood data of each piece of pressure data under the upper limit value.
3. The real-time processing method of automobile data based on artificial intelligence according to claim 1, wherein the obtaining the initial abnormality degree of each pressure data under the upper limit value according to the difference value between the neighborhood data of each pressure data under the upper limit value comprises the following specific steps:
obtaining local reachable density of each pressure data according to neighborhood data of each pressure data under the upper limit value, and recording the local reachable density of each pressure data under the upper limit value;
and obtaining the initial abnormality degree of each pressure data under the upper limit value according to the local reachable density of each pressure data.
4. The method for real-time processing automobile data based on artificial intelligence according to claim 3, wherein the initial abnormality degree of each pressure data under the upper limit value is obtained according to the local reachable density of each pressure data, and the specific calculation formula is as follows:
wherein,local reachable density of the jth neighborhood data representing the ith pressure data below an upper limit,/->Representing the locally attainable density of the ith pressure data below the upper limit, < >>Euclidean distance of the ith pressure data and the jth neighborhood data below the upper limit value,/>Representing the number of neighborhood data of the ith pressure data below the upper limit, +.>The initial abnormality degree of the ith pressure data at the upper limit value is indicated.
5. The real-time processing method of automobile data based on artificial intelligence according to claim 1, wherein the step of obtaining the final abnormality degree of each pressure data under the upper limit value from the neighborhood data of each pressure data under the upper limit value according to the initial abnormality degree of each pressure data under the upper limit value, the correlation between the pressure data sequence and the temperature data sequence comprises the following specific steps:
obtaining a first correlation coefficient and a second correlation coefficient according to the pressure data sequence and the temperature data sequence; obtaining neighborhood correlation of each pressure data under the upper limit value according to the neighborhood data of each pressure data under the upper limit value;
the calculation method for obtaining the final abnormality degree of each pressure data according to the first correlation coefficient, the second correlation coefficient and the neighborhood correlation of each pressure data comprises the following steps:
wherein,a second correlation coefficient representing the ith pressure data, +.>A first correlation coefficient representing the ith pressure data,representing the neighborhood correlation of the ith pressure data below an upper limit,/for the pressure data>The initial degree of abnormality of the ith pressure data under the upper limit value, +.>Represents an exponential function based on natural constants, < ->The final degree of abnormality of the ith pressure data at the upper limit value is indicated.
6. The method for real-time processing of vehicle data based on artificial intelligence according to claim 5, wherein the first correlation coefficient and the second correlation coefficient are obtained according to a pressure data sequence and a temperature data sequence; obtaining the neighborhood correlation of each pressure data under the upper limit value according to the neighborhood data of each pressure data under the upper limit value, comprising the following specific steps:
recording any one pressure data as target pressure data, recording rank correlation coefficients of a pressure data sequence and a temperature data sequence as first correlation coefficients of target pressure data, recording temperature data at the same time as the target pressure data as target temperature data, recording a data sequence obtained by removing the target pressure data from the pressure data sequence as a residual sequence of the target pressure data, recording a data sequence obtained by removing the target temperature data from the temperature data sequence as a residual sequence of the target temperature data, calculating rank correlation coefficients of the residual sequence of the target pressure data and the residual sequence of the target temperature data, recording the rank correlation coefficients of the residual sequence of the target pressure data and the second correlation coefficients of the target pressure data, and obtaining the first correlation coefficients and the second correlation coefficients of each pressure data;
obtaining neighborhood data of target pressure data under an upper limit value, taking temperature data which is the same as the neighborhood data of the target pressure data under the upper limit value as the neighborhood data of the target temperature data under the upper limit value, calculating rank correlation coefficients of the target pressure data and each neighborhood data of the target temperature data under the upper limit value, recording the accumulated sum of the rank correlation coefficients of all the neighborhood data of the target pressure data and the target temperature data under the upper limit value as the neighborhood correlation of the target pressure data under the upper limit value, and obtaining the neighborhood correlation of each pressure data under the upper limit value.
7. The real-time processing method of automobile data based on artificial intelligence according to claim 1, wherein the obtaining of the abnormality detection evaluation value of each pressure data at the upper limit value and the lower limit value according to the difference of the final abnormality degree between the neighborhood data of each pressure data at the lower limit value comprises the following specific steps:
obtaining a plurality of combinations of each pressure under the upper limit value according to the neighborhood data of each pressure under the upper limit value; obtaining a neighborhood radius, an increment neighborhood radius and a decrement neighborhood radius of each pressure data under an upper limit value;
according to a plurality of combinations of each pressure data under the upper limit value, the neighborhood radius, the increment neighborhood radius and the decrement neighborhood radius of each pressure data under the upper limit value are used for obtaining the abnormal detection evaluation value of each pressure data under the upper limit value, and the calculation method comprises the following steps:
wherein,representation ofIncrement neighborhood radius for ith pressure data below upper limit, +.>Neighborhood radius representing the ith pressure data below the upper limit, +.>A decrement neighborhood radius representing the ith pressure data below the upper limit, +.>The ith pressure data is represented by the +.>One of the neighborhood data in the combinations, +.>The ith pressure data is represented by the +.>Another neighborhood data in the combination, +.>Represents the number of combinations of the ith pressure data below the upper limit,/->An abnormality detection evaluation value indicating that the ith pressure data is below the upper limit value.
8. The real-time processing method of automobile data based on artificial intelligence according to claim 7, wherein the neighborhood data under the upper limit value of each pressure data is used for obtaining a plurality of combinations under the upper limit value of each pressure; the method for acquiring the neighborhood radius, the increment neighborhood radius and the decrement neighborhood radius of each pressure data under the upper limit value comprises the following specific steps:
at the upper limit of each pressure dataCombining any two neighborhood data under the value to obtain a plurality of combinations of each pressure data under the upper limit value; acquiring the radius of an M1 distance neighborhood of each pressure data, and recording the radius as the neighborhood radius of each pressure data under an upper limit value; acquiring each pressure dataThe radius of the distance neighborhood is recorded as the increment neighborhood radius of each pressure data under the upper limit value; acquiring +.>The radius of the distance neighborhood is recorded as a decrement neighborhood radius of each pressure data under an upper limit value, and M1 represents a preset upper limit value.
9. The real-time processing method of automobile data based on artificial intelligence according to claim 1, wherein the adjusting the K value according to the anomaly detection evaluation value of each pressure data under the upper limit value to obtain the optimal K value of each pressure data comprises the following specific steps:
comparing the abnormality detection evaluation value of each pressure data under the upper limit value with a preset index threshold Y1, and taking the upper limit value as the optimal K value of the pressure data when the abnormality detection evaluation value of each pressure data under the upper limit value is larger than the preset index threshold;
when the abnormal detection evaluation value of each pressure data under the upper limit value is smaller than or equal to a preset index threshold value, marking the difference value between the upper limit value and one as a first adjustment value, acquiring the abnormal detection evaluation value of each pressure data under the first adjustment value, comparing the abnormal detection evaluation value of each pressure data under the first adjustment value with a preset index threshold value Y1, and when the abnormal detection evaluation value of each pressure data under the first adjustment value is larger than the preset index threshold value, taking the first adjustment value as the optimal K value of the pressure data;
when the abnormal detection evaluation value of each pressure data under the first adjustment value is smaller than or equal to the preset index threshold value, the difference value between the first adjustment value and one is recorded as a second adjustment value, the abnormal detection evaluation value of each pressure data under the second adjustment value is obtained, the abnormal detection evaluation value of each pressure data under the second adjustment value is compared with the preset index threshold value Y1, when the abnormal detection evaluation value of each pressure data under the second adjustment value is larger than the preset index threshold value, the second adjustment value is used as the optimal K value of the pressure data, and the like until the abnormal detection evaluation value of each pressure data under the adjustment value is larger than the preset index threshold value, and the optimal K value of each pressure data is obtained.
10. The real-time processing method of automobile data based on artificial intelligence according to claim 1, wherein the anomaly detection of the pressure data according to the optimal K value of each pressure data to obtain the anomaly data comprises the following specific steps:
and setting the K value of each pressure data to be equal to the optimal K value, and analyzing all the pressure data by using an LOF algorithm based on the K value of each pressure data to obtain abnormal data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410071350.0A CN117591986B (en) | 2024-01-18 | 2024-01-18 | Real-time automobile data processing method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410071350.0A CN117591986B (en) | 2024-01-18 | 2024-01-18 | Real-time automobile data processing method based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117591986A true CN117591986A (en) | 2024-02-23 |
| CN117591986B CN117591986B (en) | 2024-04-05 |
Family
ID=89911863
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410071350.0A Active CN117591986B (en) | 2024-01-18 | 2024-01-18 | Real-time automobile data processing method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117591986B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117828516A (en) * | 2024-03-06 | 2024-04-05 | 湖南大学 | On-site data optimization acquisition system for digital pressure tank production workshop |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214355A (en) * | 2018-09-29 | 2019-01-15 | 西安交通大学 | A kind of machinery monitoring data exception section detection method based on kernel estimates LOF |
| CN110532119A (en) * | 2019-07-26 | 2019-12-03 | 中国船舶重工集团公司第七一九研究所 | Power system operation abnormal point detecting method |
| CN111398886A (en) * | 2020-04-09 | 2020-07-10 | 国网山东省电力公司电力科学研究院 | Detection method and system for automatically detecting online abnormity of epitope of assembly line |
| CN112101174A (en) * | 2020-09-09 | 2020-12-18 | 洛阳师范学院 | LOF-Kurtogram-based mechanical fault diagnosis method |
| CN114880381A (en) * | 2022-05-31 | 2022-08-09 | 上海嘉车信息科技有限公司 | Vehicle abnormity processing method and device, electronic equipment and storage medium |
| CN116429220A (en) * | 2023-06-14 | 2023-07-14 | 济宁安泰矿山设备制造有限公司 | Hydraulic engineering anomaly detection method |
| CN116644373A (en) * | 2023-07-27 | 2023-08-25 | 深圳恒邦新创科技有限公司 | Automobile flow data analysis management system based on artificial intelligence |
| CN116738353A (en) * | 2023-08-15 | 2023-09-12 | 安拓思纳米技术（苏州）有限公司 | Pharmaceutical workshop air filter element performance detection method based on data analysis |
| CN117314020A (en) * | 2023-11-28 | 2023-12-29 | 生态环境部华南环境科学研究所(生态环境部生态环境应急研究所) | Wetland carbon sink data monitoring system of plankton |
-
2024
- 2024-01-18 CN CN202410071350.0A patent/CN117591986B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214355A (en) * | 2018-09-29 | 2019-01-15 | 西安交通大学 | A kind of machinery monitoring data exception section detection method based on kernel estimates LOF |
| CN110532119A (en) * | 2019-07-26 | 2019-12-03 | 中国船舶重工集团公司第七一九研究所 | Power system operation abnormal point detecting method |
| CN111398886A (en) * | 2020-04-09 | 2020-07-10 | 国网山东省电力公司电力科学研究院 | Detection method and system for automatically detecting online abnormity of epitope of assembly line |
| CN112101174A (en) * | 2020-09-09 | 2020-12-18 | 洛阳师范学院 | LOF-Kurtogram-based mechanical fault diagnosis method |
| CN114880381A (en) * | 2022-05-31 | 2022-08-09 | 上海嘉车信息科技有限公司 | Vehicle abnormity processing method and device, electronic equipment and storage medium |
| CN116429220A (en) * | 2023-06-14 | 2023-07-14 | 济宁安泰矿山设备制造有限公司 | Hydraulic engineering anomaly detection method |
| CN116644373A (en) * | 2023-07-27 | 2023-08-25 | 深圳恒邦新创科技有限公司 | Automobile flow data analysis management system based on artificial intelligence |
| CN116738353A (en) * | 2023-08-15 | 2023-09-12 | 安拓思纳米技术（苏州）有限公司 | Pharmaceutical workshop air filter element performance detection method based on data analysis |
| CN117314020A (en) * | 2023-11-28 | 2023-12-29 | 生态环境部华南环境科学研究所(生态环境部生态环境应急研究所) | Wetland carbon sink data monitoring system of plankton |
Non-Patent Citations (2)
| Title |
|---|
| LIANGWEI ZHANG ET AL.: "Adaptive kernel density-based anomaly detection for nonlinear systems", 《KNOWLEDGE-BASED SYSTEMS》, 1 January 2018 (2018-01-01), pages 50 - 63 * |
| 周鹏，程艳云: "一种改进的LOF异常点检测算法", 《计算机技术与发展》, vol. 27, no. 12, 31 December 2017 (2017-12-31), pages 115 - 118 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117828516A (en) * | 2024-03-06 | 2024-04-05 | 湖南大学 | On-site data optimization acquisition system for digital pressure tank production workshop |
| CN117828516B (en) * | 2024-03-06 | 2024-05-24 | 湖南大学 | On-site data optimization acquisition system for digital pressure tank production workshop |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117591986B (en) | 2024-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111832225B (en) | Method for constructing driving condition of automobile | |
| CN111353482B (en) | LSTM-based fatigue factor recessive anomaly detection and fault diagnosis method | |
| CN117591986B (en) | Real-time automobile data processing method based on artificial intelligence | |
| CN111562108A (en) | Rolling bearing intelligent fault diagnosis method based on CNN and FCMC | |
| EP1960853B1 (en) | Evaluating anomaly for one-class classifiers in machine condition monitoring | |
| Mechefske et al. | Fault detection and diagnosis in low speed rolling element bearings Part II: The use of nearest neighbour classification | |
| CN102521534B (en) | Intrusion detection method based on crude entropy property reduction | |
| CN111967535B (en) | Fault diagnosis method and device for temperature sensor of grain storage management scene | |
| CN111062093B (en) | Automobile tire service life prediction method based on image processing and machine learning | |
| CN114170184A (en) | Product image anomaly detection method and device based on embedded feature vector | |
| CN112395382A (en) | Ship abnormal track data detection method and device based on variational self-encoder | |
| CN115098962A (en) | Method for predicting residual life of mechanical equipment in degradation state based on hidden half Markov model | |
| CN113255591A (en) | Bearing fault diagnosis method based on random forest and fusion characteristics | |
| CN114139624A (en) | Method for mining time series data similarity information based on integrated model | |
| CN112836967B (en) | New energy automobile battery safety risk assessment system | |
| CN115958990A (en) | V2G charging pile information diagnosis and safety protection method | |
| CN116714437B (en) | Hydrogen fuel cell automobile safety monitoring system and monitoring method based on big data | |
| CN112199295A (en) | Deep neural network defect positioning method and system based on frequency spectrum | |
| CN109840386B (en) | Damage identification method based on factor analysis | |
| CN117033891A (en) | Traffic accident severity assessment method based on single vehicle traffic accident database | |
| CN116244657A (en) | Train axle temperature abnormality identification method based on generation of countermeasure network and ensemble learning | |
| CN115718890A (en) | Driving behavior prediction method and device, electronic equipment and storage medium | |
| CN112214934A (en) | Suspension system life prediction method based on multiple sensors and related device | |
| CN118154985A (en) | Unsupervised industrial data classification method | |
| CN117237911A (en) | Image-based dynamic obstacle rapid detection method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |