Vehicle condition supervision method and device based on automobile, terminal equipment and medium

Abstract

The embodiment of the invention provides a vehicle condition supervision method and device based on a vehicle, terminal equipment and a medium, and belongs to the technical field of vehicle supervision. The method comprises the following steps: performing data change analysis on initial sensing data of an initial vehicle to obtain a data change point; performing sensor association analysis according to the data change points to obtain a target association relationship between the initial sensors; determining an initial interference parameter according to the target association relationship and the initial sensing data; obtaining target sensing data of the target vehicle under the target sensor, and performing classification according to the target sensing data to obtain a target operation type; determining a target interference parameter from the initial interference parameters according to the target operation type and a preset operation type; determining a data change range of the target vehicle under the sensor according to the target interference parameter; performing anomaly identification on the target sensing data according to the data change range to obtain target anomaly data; and performing condition supervision on the target vehicle according to the target abnormal data to obtain a vehicle supervision result.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle monitoring, and in particular to a vehicle condition monitoring method, device, terminal equipment and medium based on automobiles. Background Art

[0002] With the rapid development of social economy, people's quality of life has been significantly improved, and more and more families and individuals have chosen to buy cars as their daily means of transportation. However, although cars have brought convenience to our lives, it is impossible to effectively monitor and discover the actual condition of the vehicle in a timely manner during driving. This problem directly leads to an increase in driving risks, especially when facing emergencies, it is often difficult for drivers to obtain the vehicle's health status information in the first place, and thus they are unable to make timely and accurate response measures.

[0003] Current technical solutions have introduced machine learning models to monitor vehicle conditions, but these models mainly rely on anomaly classification to detect potential faults or problems. These models often misjudge or miss when identifying abnormal vehicle conditions. For example, some potential fault signals may be ignored, or some normal vehicle conditions may be misjudged as abnormal, leading to unnecessary repairs or safety hazards. Therefore, the vehicle condition monitoring methods in the prior art are prone to misjudgment or miss, and are unable to accurately and timely monitor and discover the actual condition of the vehicle, resulting in increased driving risks. Summary of the invention

[0004] The main purpose of the embodiments of the present invention is to provide a vehicle condition monitoring method, device, terminal equipment and medium based on a car, aiming to solve the problem in related technologies that the actual condition of the vehicle cannot be accurately and timely monitored and discovered, thereby increasing driving risks.

[0005] In a first aspect, an embodiment of the present invention provides a vehicle condition monitoring method based on a vehicle, comprising:

[0006] Obtaining initial sensor data corresponding to an initial sensor of an initial vehicle under a preset operation type, and performing data change analysis on the initial sensor data to obtain data change points corresponding to the initial sensor of the initial vehicle under the preset operation type;

[0007] Performing sensor association analysis according to the data change points to obtain a target association relationship between the initial sensors of the initial vehicle under the preset operation type;

[0008] Determine an initial interference parameter corresponding to the initial sensor according to the target association relationship and the initial sensing data;

[0009] Obtaining target sensing data corresponding to the target vehicle under the target sensor, and classifying the operation type according to the target sensing data to obtain the target operation type corresponding to the target sensing data;

[0010] Determining a target interference parameter corresponding to the target operation type from the initial interference parameter according to the target operation type and the preset operation type;

[0011] Determine the data variation range corresponding to the target vehicle under the target sensor according to the target interference parameter;

[0012] Performing abnormal identification on the target sensor data according to the data variation range to obtain target abnormal data corresponding to the target vehicle;

[0013] The vehicle condition of the target vehicle is monitored according to the target abnormal data to obtain a vehicle supervision result corresponding to the target vehicle.

[0014] In a second aspect, an embodiment of the present invention provides a vehicle condition monitoring device based on a car, comprising:

[0015] A data acquisition module, used to obtain initial sensor data corresponding to an initial sensor of an initial vehicle under a preset operation type, and perform data change analysis on the initial sensor data to obtain data change points corresponding to the initial sensor of the initial vehicle under the preset operation type;

[0016] A data analysis module, configured to perform sensor association analysis according to the data change points to obtain a target association relationship corresponding to the initial sensors of the initial vehicle under the preset operation type;

[0017] A parameter processing module, used for determining an initial interference parameter corresponding to the initial sensor according to the target association relationship and the initial sensing data;

[0018] A data classification module is used to obtain target sensing data corresponding to the target vehicle under the target sensor, and classify the operation type according to the target sensing data to obtain the target operation type corresponding to the target sensing data;

[0019] a parameter determination module, configured to determine a target interference parameter corresponding to the target operation type from the initial interference parameter according to the target operation type and the preset operation type;

[0020] A range determination module, used to determine the data change range corresponding to the target vehicle under the target sensor according to the target interference parameter;

[0021] An abnormality identification module, used for identifying abnormalities of the target sensor data according to the data variation range, and obtaining target abnormality data corresponding to the target vehicle;

[0022] The result determination module is used to perform vehicle condition supervision on the target vehicle according to the target abnormal data to obtain a vehicle supervision result corresponding to the target vehicle.

[0023] In a third aspect, an embodiment of the present invention further provides a terminal device, comprising a processor, a memory, a computer program stored in the memory and executable by the processor, and a data bus for realizing connection and communication between the processor and the memory, wherein when the computer program is executed by the processor, the steps of any one of the vehicle condition supervision methods based on automobiles provided in the specification of the present invention are realized.

[0024] In a fourth aspect, an embodiment of the present invention further provides a storage medium for computer-readable storage, characterized in that the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any one of the vehicle condition supervision methods based on automobiles provided in the specification of the present invention.

[0025] The embodiment of the present invention provides a vehicle condition monitoring method, apparatus, terminal device and medium based on automobiles. The method comprises: obtaining initial sensor data corresponding to an initial sensor of an initial vehicle under a preset operation type, and performing data change analysis on the initial sensor data to obtain data change points corresponding to the initial sensor of the initial vehicle under the preset operation type; performing sensor association analysis according to the data change points to obtain target association relationships corresponding to initial sensors of the initial vehicle under the preset operation type, so that different target association relationships corresponding to initial sensors under different preset operation types can be discovered in time, thereby providing good support for subsequent timely discovery of abnormalities under different preset operation types; and then determining initial interference parameters corresponding to the initial sensors according to the target association relationship and the initial sensor data, so that the sensor parameters can be accurately determined. The initial interference parameters of the sensor are used to reduce the impact of external interference on sensor data and improve the reliability of data, so that abnormalities under different preset operation types can be discovered in time and the corresponding abnormal data can be identified in time; the target sensor data corresponding to the target vehicle under the target sensor is obtained, and the operation type is classified according to the target sensor data to obtain the target operation type corresponding to the target sensor data; the target interference parameters corresponding to the target operation type are determined from the initial interference parameters according to the target operation type and the preset operation type; the data change range corresponding to the target vehicle under the target sensor is determined according to the target interference parameters; the target sensor data is identified according to the data change range to obtain the target abnormal data corresponding to the target vehicle; the vehicle condition of the target vehicle is monitored according to the target abnormal data to obtain the vehicle supervision result corresponding to the target vehicle. By performing detailed analysis and abnormal identification of the sensor data of the target vehicle under the target operation type, the target abnormal data can be discovered in time. This real-time monitoring mechanism can not only accurately monitor the operating status of the target vehicle, but also effectively reduce the risks that the vehicle may face during driving. In this way, driving safety is significantly improved, and the problem that the actual condition of the vehicle cannot be accurately and timely monitored and discovered in the relevant technology, which leads to increased driving risks, is also solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0027] Figure 1 A flow chart of a vehicle condition monitoring method based on a car provided by an embodiment of the present invention;

[0028] Figure 2A schematic diagram of the module structure of a vehicle condition monitoring device based on a car provided by an embodiment of the present invention;

[0029] Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0032] It should be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0033] The embodiment of the present invention provides a vehicle condition monitoring method, apparatus, terminal device and medium based on automobile. The vehicle condition monitoring method based on automobile can be applied to a terminal device, which can be an electronic device such as a tablet computer, a laptop computer, a desktop computer, a personal digital assistant and a wearable device. The terminal device can be a server or a server cluster.

[0034] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0035] Please refer to Figure 1 , Figure 1 A flow chart of a vehicle condition monitoring method based on a car provided in an embodiment of the present invention.

[0036] like Figure 1 As shown, the automobile-based vehicle condition supervision method includes steps S101 to S108.

[0037] Step S101, obtaining initial sensor data corresponding to an initial sensor of an initial vehicle under a preset operation type, and performing data change analysis on the initial sensor data to obtain data change points corresponding to the initial sensor of the initial vehicle under the preset operation type.

[0038] Exemplarily, the preset operation type refers to any operation type of vehicle operations that the initial vehicle can perform. For example, the preset operation type can be any vehicle operation such as vehicle acceleration, deceleration, gear shifting, lane changing, braking, etc.

[0039] Exemplarily, multiple initial sensors are installed at key parts of the initial vehicle, such as the engine, brake system, tire, battery, suspension system, etc. The types of initial sensors include temperature sensors, pressure sensors, acceleration sensors, vibration sensors, etc., and then the initial sensor data corresponding to the initial vehicle is collected, including but not limited to temperature, pressure, vibration, speed, fuel consumption, battery status, tire pressure, brake wear and other data.

[0040] Exemplarily, the initial sensor data is analyzed using a data change analysis algorithm, such as a time series-based analysis method (such as sliding window analysis, differential analysis) or a statistical method (such as variance analysis, trend analysis), so as to identify data change points in the initial sensor data. The data change points are used to indicate a sudden change or trend change in the sensor state of the vehicle under a preset operation type.

[0041] In some embodiments, the data change analysis of the initial sensor data to obtain the data change point corresponding to the initial sensor of the initial vehicle under the preset operation type includes: determining the initial change point, and cutting the initial sensor data according to the initial change point to obtain the first sensor data and the second sensor data; performing statistical data calculation on the first sensor data to obtain the corresponding first mean and first standard deviation, and performing statistical data calculation on the second sensor data to obtain the corresponding second mean and second standard deviation; determining the segmentation quality value corresponding to the initial change point according to the first mean, the first standard deviation, the second mean and the second standard deviation; obtaining the maximum value corresponding to the segmentation quality value as the target quality value, and determining the distribution probability of the target quality value corresponding to the segmentation quality value; determining the segmentation state corresponding to the target quality value according to the distribution probability and the preset value, and determining the data change point from the initial change point according to the segmentation state and the target quality value; wherein the segmentation quality value is obtained according to the following formula:

[0042] ;

[0043] in, represents the segmentation quality value corresponding to the i-th initial change point, represents the first mean value corresponding to the first sensing data corresponding to the i-th initial change point, represents the second mean value corresponding to the second sensing data corresponding to the i-th initial change point; represents the first standard deviation corresponding to the first sensing data corresponding to the i-th initial change point; represents the second standard deviation corresponding to the second sensing data corresponding to the i-th initial change point; represents the number of data corresponding to the first sensing data corresponding to the i-th initial change point, represents the number of data corresponding to the second sensing data corresponding to the i-th initial change point, Indicates taking the absolute value.

[0044] Exemplarily, a change point detection algorithm (such as CUSUM, PELT, Binary Segmentation, etc.) is used to scan the initial sensor data to identify possible initial change points, and then for each identified initial change point, the initial sensor data is cut into two parts: the data before the initial change point is determined as the first sensor data and the data after the initial change point is determined as the second sensor data.

[0045] Exemplarily, a statistical algorithm is used to calculate the mean of the first sensor data to obtain a first mean, and a statistical algorithm is used to calculate the standard deviation of the first sensor data to obtain a first standard deviation, and a statistical algorithm is used to calculate the mean of the second sensor data to obtain a second mean, and a statistical algorithm is used to calculate the standard deviation of the second sensor data to obtain a second standard deviation.

[0046] Exemplarily, the first mean, the first standard deviation, the second mean, and the second standard deviation are used according to the following formula to obtain the segmentation quality value corresponding to the initial change point, and the segmentation quality value measures the degree of difference in data distribution before and after the initial change point:

[0047] ;

[0048] in, represents the segmentation quality value corresponding to the i-th initial change point, represents the first mean value corresponding to the first sensor data corresponding to the i-th initial change point, represents the second mean value corresponding to the second sensor data corresponding to the i-th initial change point; represents the first standard deviation of the first sensor data corresponding to the i-th initial change point; represents the second standard deviation corresponding to the second sensor data corresponding to the i-th initial change point; represents the number of data corresponding to the first sensor data corresponding to the i-th initial change point, represents the number of data corresponding to the second sensing data corresponding to the i-th initial change point, Indicates taking the absolute value.

[0049] Exemplarily, the segmentation quality value can quantify the distribution difference of the data before and after the initial change point by combining the first mean, the first standard deviation, the second mean and the second standard deviation. The first mean and the second mean reflect the central tendency of the data, and the first standard deviation and the second standard deviation reflect the degree of dispersion of the data. By combining the first mean, the first standard deviation, the second mean and the second standard deviation using the above formula, the change in data distribution can be more comprehensively evaluated, thereby providing an objective basis for the subsequent determination of data change points. The above formula comprehensively considers multiple statistical characteristics of data distribution (mean, standard deviation, number of data), rather than relying solely on the mean or standard deviation, so that the segmentation quality value can better reflect the significance of the data before and after the change point. Therefore, by calculating and screening the segmentation quality value, significant change points in the data can be effectively identified and false change points can be excluded. This makes data segmentation more robust and can maintain high reliability in complex real-world scenarios (such as in the presence of noise or large data fluctuations).

[0050] Exemplarily, the segmentation quality value with the largest segmentation quality value among all the initial change points is obtained as the target quality value. At the same time, the distribution probability of the target quality value among all the segmentation quality values ​​is calculated, which can be a probability density function value or a cumulative distribution function value, and then the segmentation state corresponding to the target quality value is determined according to the distribution probability and the preset value. If the distribution probability is higher than or equal to the preset value, it is considered that the initial change point corresponding to the target quality value is a data change point. If the distribution probability is lower than the preset value, it is considered that there is no data change point in the initial sensing data, and then the data change point can be determined as empty or as a preset mark, such as None, so that it can be accurately identified when performing association relationship analysis based on the data change point later.

[0051] Step S102: performing sensor association analysis according to the data change points to obtain target association relationships corresponding to the initial sensors of the initial vehicle under the preset operation type.

[0052] Exemplarily, the associated sensors of the initial vehicle that change or are affected under a preset operation type are obtained based on the data change points. When the data change point is empty or is a preset mark, it means that under the preset operation type, there is no impact on the sensors of the initial vehicle whose data change points are empty or are preset marks. Then, the sensor data corresponding to the sensors whose data change points are not empty or are not preset marks are associated with analysis to obtain the corresponding target association relationship.

[0053] Exemplarily, sensors related to preset operation types are selected through data change points, such as speed sensors, acceleration sensors, steering angle sensors, accelerator pedal position sensors, etc., so as to obtain relevant sensor data corresponding to relevant sensors from initial sensor data, and then calculate the target association relationship between relevant sensors using correlation coefficient analysis (such as Pearson correlation coefficient, Spearman rank correlation coefficient), mutual information analysis, principal component analysis (PCA), cluster analysis (such as K-means), etc. The target association relationship is one of strong positive correlation, strong negative correlation, weak positive correlation, and weak negative correlation.

[0054] In some embodiments, the initial sensor includes at least a first sensor and a second sensor, the initial sensor data includes at least first sensor data corresponding to the first sensor and second sensor data corresponding to the second sensor, the data change point includes at least a first change point corresponding to the first sensor data and a second change point corresponding to the second sensor data, and the sensor association analysis based on the data change points to obtain the target association relationship between the initial sensors of the initial vehicle under the preset operation type includes: obtaining a first acquisition time corresponding to the first change point from the first sensor data and a second acquisition time corresponding to the second change point from the second sensor data; obtaining a first time quantity corresponding to the first acquisition time and a second time quantity corresponding to the second acquisition time; determining an initial association matrix corresponding to the first sensor and the second sensor according to the first time quantity and the second time quantity; adjusting the initial association matrix according to the first change point and the second change point to obtain a target association matrix corresponding to the first sensor and the second sensor of the initial vehicle under the preset operation type; performing sensor association analysis based on the target association matrix to obtain the target association relationship between the first sensor and the second sensor of the initial vehicle under the preset operation type.

[0055] Exemplarily, an initial sensor associated with a preset operation type is selected through data change points, the initial sensor includes at least a first sensor and a second sensor, the initial sensor data includes at least first sensor data corresponding to the first sensor and second sensor data corresponding to the second sensor, and the data change points include at least a first change point corresponding to the first sensor data and a second change point corresponding to the second sensor data.

[0056] Exemplarily, the first sensor data and the second sensor data respectively include multiple data change points, and then first acquisition times corresponding to the multiple first change points are obtained from the first sensor data, and second acquisition times corresponding to the multiple second change points are obtained from the second sensor data, and then data statistics are performed on the first acquisition time to obtain a first time quantity, and data statistics are performed on the second acquisition time to obtain a second time quantity.

[0057] Exemplarily, an initial correlation matrix is ​​constructed using the first time quantity and the second time quantity. The rows of the matrix represent the first sensor, and the columns represent the second sensor. Each element in the matrix represents the quantity corresponding to the data change points of the first sensor and the second sensor at the corresponding time.

[0058] Exemplarily, the initial correlation matrix is ​​adjusted according to the time alignment of the first change point and the second change point. For example, if the time difference between the first change point and the second change point is large, it may be necessary to add a time offset adjustment in the matrix. The elements in the matrix are dynamically adjusted using the time characteristics of the first change point and the second change point (such as the change amplitude, trend, etc.). For example, the sensor pair with a large change amplitude can increase its association weight, and then correct the adjusted initial correlation matrix to ensure that the elements in the matrix can accurately reflect the true association relationship between the two sensors under the preset operation type, and then standardize or normalize the elements in the matrix to ensure that the association strength between different sensors is comparable.

[0059] Exemplarily, the target association matrix is ​​used to perform further association analysis, such as calculating the correlation coefficient, mutual information or other statistical indicators, to quantify the association strength between the two sensors, and then compare the association strength with a preset strength, so as to determine the target association relationship between the first sensor and the second sensor based on the comparison result.

[0060] In some embodiments, the sensor association analysis performed according to the target association matrix to obtain the target association relationship between the first sensor and the second sensor of the initial vehicle under the preset operation type includes: obtaining a target characterization vector between the first sensor and the second sensor according to the target association matrix; weighting the target characterization vector through a modulation function to obtain an initial characterization value corresponding to the first sensor and the second sensor; data-adjusting the initial characterization value through an adaptive filter to obtain a target characterization value corresponding to the first sensor and the second sensor; and determining the target association relationship between the first sensor and the second sensor according to the target characterization value.

[0061] Exemplarily, each row (or column) in the target association matrix represents an association relationship between a sensor and other sensors. The association rows (or columns) corresponding to the first sensor and the second sensor are extracted from the matrix to form two representation vectors. The elements in each representation vector represent the strength of association between the first sensor (or the second sensor) and other sensors (including the second sensor or the first sensor).

[0062] Exemplarily, a modulation function is designed to perform weighted processing on the elements in the characterization vector. The input of the modulation function is the elements of the characterization vector, and the output is the weighted numerical value. The modulation function can dynamically adjust the weight according to the importance of the sensor, the strength of association, or other factors. The modulation function is applied to the characterization vector of the first sensor and the characterization vector of the second sensor respectively to obtain two weighted initial characterization numerical values. The weighted numerical values ​​can better reflect the correlation characteristics between the first sensor and the second sensor.

[0063] Exemplarily, an adaptive filter is used to perform data adjustment on the initial characterization value. The adaptive filter can dynamically adjust the filter parameters according to the input data. The input of the filter is the weighted initial characterization value, and the output is the adjusted target characterization value. The initial characterization value of the first sensor and the initial characterization value of the second sensor are filtered respectively. The filter can automatically optimize the filtering effect according to data fluctuations, noise levels or other factors, so that the target characterization value is more stable and accurate.

[0064] Exemplarily, the target representation values ​​of the first sensor and the second sensor are analyzed to evaluate the strength and consistency of the association between them. According to the degree of difference or similarity of the target representation values, the target association relationship between the first sensor and the second sensor is determined. If the target representation values ​​are close or highly correlated, it is determined that there is a strong association relationship between them; otherwise, it is determined to be a weak association relationship.

[0065] Specifically, the application of modulation functions and adaptive filters makes the calculation of association relationships more dynamic and flexible, and can adapt to different operation types and data characteristics. The data conditioning step can effectively eliminate the influence of noise and outliers, making the target representation value more stable and reliable, thereby improving the accuracy of association relationship analysis. Further, through the determination of target association relationships, it can provide strong support for more accurate monitoring of vehicle state changes and improving the efficiency and reliability of multi-sensor collaborative work.

[0066] Step S103: determining an initial interference parameter corresponding to the initial sensor according to the target association relationship and the initial sensing data.

[0067] Exemplarily, the initial sensor includes at least a first sensor and a second sensor, and a sensor associated with the first sensor is found from the previously determined target association relationship, namely, an associated sensor. Data corresponding to the associated sensor is extracted from the initial sensor data, which is called the first data. At the same time, the sensor data corresponding to the first sensor is extracted, which is called the second data. Ensure that the first data and the second data are aligned in time, that is, correspond to the same time point. Preprocess the first data and the second data, including denoising, filtering, normalization and other operations to eliminate noise and outliers, ensure the accuracy of the data, and then identify the interference factors of the associated sensor on the first sensor, such as cross sensitivity, environmental impact, etc., and select a suitable data adjustment method, such as linear regression, partial least squares method, Kalman filtering, etc., for adjusting the second data according to the first data, and then using the first data as an adjustment factor to adjust the second data to eliminate or reduce the influence of the associated sensor on the first sensor. For example, by establishing a mathematical model between the first data and the second data, calculating the interference correction amount, and subtracting the correction amount from the second data, the interference degree of the associated sensor on the first sensor, that is, the initial interference parameter, is calculated during the data adjustment process. The initial interference parameter can be an interference coefficient, an interference percentage, or other quantitative indicators, reflecting the influence of the associated sensor on the first sensor. Similarly, the initial interference parameter corresponding to the second sensor is obtained according to the above process.

[0068] In some embodiments, the initial interference parameters include at least a first interference parameter corresponding to the first sensor and a second interference parameter corresponding to the second sensor, and the determining the initial interference parameters corresponding to the initial sensor according to the target association relationship and the initial sensor data includes: determining a first associated sensor corresponding to the first sensor and a second associated sensor corresponding to the second sensor from the initial sensor according to the target association relationship; obtaining third sensor data corresponding to the first associated sensor and fourth sensor data corresponding to the second associated sensor; calculating the first data difference corresponding to the adjacent time data in the first sensor data and calculating the second sensor data The method comprises the steps of: calculating a first data difference corresponding to the adjacent time data in the third sensor data and calculating a fourth data difference corresponding to the adjacent time data in the fourth sensor data; calculating a fifth data difference corresponding to the first sensor data and the third sensor data at the same time and calculating a sixth data difference corresponding to the second sensor data and the fourth sensor data at the same time; calculating a first sample mean and a first sample variance for the first data difference, and calculating a third sample mean and a third sample variance for the third data difference; determining the first sensor data corresponding to the first sensor data according to the first sample mean and the first sample variance combined with a normal distribution. The method comprises the steps of: calculating a first mean range of the first sensor data and calculating a third mean range of the first sensor data according to the third sample mean and the third sample variance in combination with a normal distribution; calculating a fifth sample mean and a fifth sample variance for the fifth data difference, and determining a fifth mean range corresponding to the first sensor data and the first associated sensor according to the fifth sample mean and the fifth sample variance in combination with a normal distribution; adjusting the first mean range according to the third mean range and the fifth mean range to obtain the first interference parameter corresponding to the first sensor; calculating a second sample mean and a second sample variance for the second data difference, and calculating a fourth sample mean for the fourth data difference. and fourth sample variance; determining a second mean range corresponding to the second sensor data according to the second sample mean and the second sample variance in combination with a normal distribution; determining a fourth mean range corresponding to the fourth sensor data according to the fourth sample mean and the fourth sample variance in combination with a normal distribution; calculating a sixth sample mean and a sixth sample variance for the sixth data difference, and determining a sixth mean range corresponding to the second sensor data and the second associated sensor according to the sixth sample mean and the sixth sample variance in combination with a normal distribution; adjusting the second mean range according to the fourth mean range and the sixth mean range to obtain the second interference parameter corresponding to the second sensor.

[0069] Exemplarily, the initial interference parameters include at least a first interference parameter corresponding to the first sensor and a second interference parameter corresponding to the second sensor.

[0070] Exemplarily, a first associated sensor associated with the first sensor and a second associated sensor associated with the second sensor are extracted from the target association relationship, and then third sensor data corresponding to the first associated sensor are extracted from the initial sensor data and fourth sensor data corresponding to the second associated sensor are extracted from the initial sensor data.

[0071] Exemplarily, the differences in data at adjacent moments are calculated for the first sensor data, the second sensor data, the third sensor data, and the fourth sensor data, respectively, to obtain the first data difference, i.e., the difference in data at adjacent moments in the first sensor data, the second data difference, i.e., the difference in data at adjacent moments in the second sensor data, the third data difference, i.e., the difference in data at adjacent moments in the third sensor data, and the fourth data difference, i.e., the difference in data at adjacent moments in the fourth sensor data.

[0072] Exemplarily, for the first sensing data and the third sensing data, a fifth data difference at the same time point is calculated, and for the second sensing data and the fourth sensing data, a sixth data difference at the same time point is calculated.

[0073] Exemplarily, the first sample mean is obtained by calculating the mean of the first data difference using a statistical algorithm, and the first sample variance is obtained by calculating the variance of the first data difference using a statistical algorithm. The third sample mean is obtained by calculating the mean of the third data difference using a statistical algorithm, and the third sample variance is obtained by calculating the sample variance of the third data difference using a statistical algorithm.

[0074] Exemplarily, the second sample mean is obtained by calculating the mean of the second data difference using a statistical algorithm, and the second sample variance is obtained by calculating the variance of the second data difference using a statistical algorithm. Also, the fourth sample mean is obtained by calculating the mean of the fourth data difference using a statistical algorithm, and the fourth sample variance is obtained by calculating the variance of the fourth data difference using a statistical algorithm.

[0075] Exemplarily, the fifth sample mean is obtained by calculating the mean of the fifth data difference using a statistical algorithm, and the fifth sample variance is obtained by calculating the variance of the fifth data difference using a statistical algorithm. The sixth sample mean is obtained by calculating the mean of the sixth data difference using a statistical algorithm, and the sixth sample variance is obtained by calculating the variance of the sixth data difference using a statistical algorithm.

[0076] Exemplarily, the confidence interval corresponding to the first sample mean in the first sensor data is determined according to the first sample mean and the first sample variance in combination with the normal distribution, thereby determining the confidence corresponding to the first sample mean according to the confidence interval, thereby determining the first mean range corresponding to the first sensor data according to the confidence in combination with the first sample mean and the first sample variance. Similarly, the third mean range corresponding to the third sensor data is determined according to the third sample mean and the third sample variance in combination with the normal distribution. The second mean range corresponding to the second sensor data is determined according to the second sample mean and the second sample variance in combination with the normal distribution. The fourth mean range corresponding to the fourth sensor data is determined according to the fourth sample mean and the fourth sample variance in combination with the normal distribution. The fifth mean range between the first sensor data and the first associated sensor is determined according to the fifth sample mean and the fifth sample variance in combination with the normal distribution. The sixth mean range between the second sensor data and the second associated sensor is determined according to the sixth sample mean and the sixth sample variance in combination with the normal distribution.

[0077] Exemplarily, an adjustment factor is determined according to the relationship between the fifth mean range and the third mean range. For example, if the fifth mean range is larger, it means that the first sensor is more strongly interfered by the first associated sensor, and the adjustment factor should be larger; otherwise, it should be smaller. The first mean range is combined with the adjustment factor, and linear or nonlinear adjustment is performed, so that the interference parameter is calculated according to the difference between the adjusted first mean range and the original first mean range. The first interference parameter is a quantitative indicator that describes the degree of interference of the first sensor by the first associated sensor, so as to adjust the first mean range according to the third mean range and the fifth mean range, and obtain the first interference parameter corresponding to the first sensor. Similarly, the second mean range is adjusted according to the fourth mean range and the sixth mean range to obtain the second interference parameter corresponding to the second sensor.

[0078] Step S104: obtaining target sensor data corresponding to the target vehicle under the target sensor, and classifying the operation type according to the target sensor data to obtain the target operation type corresponding to the target sensor data.

[0079] Exemplarily, target sensor data corresponding to a target vehicle in motion under a target sensor is obtained, and then an operation type is classified according to the target sensor data using a data classification model, thereby obtaining a target operation type corresponding to the target vehicle under the target sensor data. The target operation type can be any vehicle operation such as vehicle acceleration, deceleration, gear shifting, lane changing, braking, etc.

[0080] Step S105: determining a target interference parameter corresponding to the target operation type from the initial interference parameter according to the target operation type and the preset operation type.

[0081] Exemplarily, the target operation type is compared with the preset operation type, so as to obtain the target interference parameter corresponding to the case where the target operation type is the same as the preset operation type from the initial interference parameter.

[0082] Step S106: determining the data variation range corresponding to the target vehicle under the target sensor according to the target interference parameter.

[0083] Exemplarily, the expected sensor data of the target vehicle under the target sensor corresponding to the target operation type is obtained, and then the maximum interference parameter and the minimum interference parameter corresponding to the target operation type are obtained from the target interference parameter, so as to obtain the data volume corresponding to the target sensor data and the standard deviation corresponding to the target sensor data, and then the data volume is squared to obtain the square root value, and then the standard deviation is divided by the square root value to obtain the target ratio, and then the target ratio and the minimum interference parameter are multiplied to obtain the minimum value and the target ratio and the maximum interference parameter are multiplied to obtain the maximum value, so as to subtract the minimum value from the expected sensor data to obtain the lower limit value of the data variation range, and add the maximum value to the expected sensor data to obtain the upper limit value of the data variation range.

[0084] Step S107: performing abnormality identification on the target sensor data according to the data variation range to obtain target abnormality data corresponding to the target vehicle.

[0085] Exemplarily, it is determined whether the target sensor data is within the data variation range. When the target sensor data is not within the data variation range, the corresponding target sensor data is determined as target abnormal data corresponding to the target vehicle.

[0086] In some embodiments, the abnormality identification of the target sensor data according to the data variation range to obtain the target abnormality data corresponding to the target vehicle includes: determining the target data expectation corresponding to the target sensor, and performing difference calculation based on the target sensor data and the target data expectation to obtain the deviation value corresponding to the target sensor data; determining the degree of deviation corresponding to the target sensor data according to the deviation value and the data variation range; determining initial abnormal data from the target sensor data according to the degree of deviation, and obtaining the remaining sensor data after removing the initial abnormal data from the target sensor data; performing distance calculation on the initial abnormal data to obtain the distance information corresponding to the initial abnormal data and performing local density calculation on the initial abnormal data to obtain the density information corresponding to the initial abnormal data; The method comprises the steps of: performing a data distribution analysis on the distance information to obtain the corresponding first probability distribution information and performing a data distribution analysis on the density information to obtain the corresponding second probability distribution information; fusing the first probability distribution information and the second probability distribution information to determine the target abnormality probability corresponding to the initial abnormal data; determining the first abnormal data from the initial abnormal data according to the target abnormality probability; classifying the remaining sensor data by using a support vector machine to obtain the first classified data and the second classified data corresponding to the remaining sensor data; calculating the first quantity corresponding to the first classified data and the second quantity corresponding to the second classified data, and determining the second abnormal data corresponding to the remaining sensor data according to the first quantity and the second quantity; and merging the first abnormal data and the second abnormal data to obtain the target abnormal data corresponding to the target vehicle.

[0087] Exemplarily, the target data expectation corresponding to the target vehicle under the target sensor is obtained based on historical data, and then the target sensor data and the target data expectation are differentially calculated to obtain the deviation value corresponding to the target sensor data, thereby calculating the difference between the upper limit value and the lower limit value in the data change range to obtain the change difference, and then the deviation value and the change difference are ratio-calculated to obtain the degree of deviation corresponding to the target sensor data.

[0088] Exemplarily, a preset degree is determined, and when the deviation degree is greater than a preset threshold, the target sensor data corresponding to the deviation degree is determined as initial abnormal data.

[0089] Exemplarily, after obtaining the initial abnormal data, the initial abnormal data is removed from the target sensor data so that the remaining data is determined as the remaining sensor data, and then the distance calculation is performed on the initial abnormal data to obtain the distance information of each initial abnormal data. The distance information reflects the difference between the distribution of the initial abnormal data and the normal data. The local density calculation is performed on the initial abnormal data to obtain the density information of each initial abnormal data. The density information reflects the concentration degree of the initial abnormal data in its neighborhood.

[0090] Exemplarily, data distribution analysis is performed on the distance information to obtain first probability distribution information corresponding to the initial abnormal data. Data distribution analysis is performed on the density information to obtain second probability distribution information corresponding to the initial abnormal data. The first probability distribution information and the second probability distribution information are merged to determine the target abnormal probability of the initial abnormal data. According to the target abnormal probability, the first abnormal data with a higher target abnormal probability is screened out.

[0091] Exemplarily, a support vector machine (SVM) is used to classify the remaining sensor data to obtain first classified data and second classified data. Then, a first quantity corresponding to the first classified data and a second quantity corresponding to the second classified data are calculated. According to the first quantity and the second quantity, it is determined whether there is second abnormal data in the remaining sensor data. When the difference between the first quantity and the second quantity is less than or equal to a preset quantity, it means that there is no second abnormal data in the remaining sensor data. When the difference between the first quantity and the second quantity is greater than the preset quantity, the first distance between the first classified data and the first abnormal data and the second distance between the second classified data and the second abnormal data are calculated, and then the classified data corresponding to the smallest of the first distance and the second distance is determined as the second abnormal data.

[0092] Exemplarily, the first abnormal data and the second abnormal data are combined to obtain target abnormal data corresponding to the target vehicle.

[0093] Step S108: Perform vehicle condition monitoring on the target vehicle according to the target abnormal data to obtain a vehicle monitoring result corresponding to the target vehicle.

[0094] Exemplarily, a machine learning model or a deep learning model is used to perform vehicle condition supervision on target abnormal data to obtain a target vehicle condition corresponding to the target vehicle, for example, the target vehicle condition is one of low risk, medium risk and high risk. Thus, a corresponding text is generated according to the target vehicle condition to remind the target associated user corresponding to the target vehicle, and the generated text is determined as the vehicle supervision result corresponding to the target vehicle.

[0095] For example, the abnormal data of the target vehicle is collected, including various sensor data, and each piece of data is labeled to indicate its risk level (low risk, medium risk, high risk). The label can be manually annotated based on historical data or generated by an expert system. Therefore, a suitable machine learning model or deep learning model, such as logistic regression, support vector machine, random forest, neural network, etc., is selected, and the model is trained using the abnormal data and the corresponding labels, and the model's hyperparameters are adjusted to optimize performance. Therefore, the target abnormal data collected in real time is input into the trained model to output the target vehicle status, such as low risk, medium risk, and high risk, so that different text generation rules are defined according to the target vehicle status. For example, low risk generates the text "the vehicle is running normally and no special attention is required". Medium risk generates the text "there are some problems with the vehicle operation, it is recommended to check as soon as possible". High risk generates the text "there are serious problems with the vehicle operation, stop and check immediately", so that the generated text is determined as the vehicle supervision result corresponding to the target vehicle.

[0096] In some embodiments, the vehicle condition supervision of the target vehicle according to the target abnormal data to obtain a vehicle supervision result corresponding to the target vehicle includes: using the data analysis layer of the abnormal recognition model to perform abnormal characterization on the target abnormal data to obtain an initial characterization vector corresponding to the target sensor; using the data fusion layer of the abnormal recognition model to perform data fusion on the initial characterization vector using the target association relationship to obtain a target characterization vector corresponding to the target vehicle; using the data classification layer of the abnormal recognition model to perform condition classification on the target characterization vector to obtain a target classification result corresponding to the target vehicle; and determining the vehicle supervision result corresponding to the target vehicle according to the target classification result and the target operation type.

[0097] Exemplarily, the anomaly recognition model includes a data analysis layer, a data fusion layer, and a data classification layer. The data analysis layer of the anomaly recognition model is used to perform anomaly characterization on the target anomaly data to obtain an initial characterization vector corresponding to the target sensor. The target association relationship is then used to integrate the initial characterization vectors from different sensors based on the data fusion layer, taking into account the correlation between them, to form a comprehensive target characterization vector. For example, the data fusion layer uses stacking, weighted averaging, attention mechanisms, etc. to capture the relationship and synergy between sensors.

[0098] Exemplarily, the data classification layer makes a classification decision based on the target representation vector, determines the condition category of the vehicle to obtain a target classification result, and the target classification result is one of low risk, medium risk and high risk.

[0099] Exemplarily, different text generation rules are defined according to the target classification results, such as low risk, medium risk, and high risk. For example, the target classification result is combined with the target operation type, and the target operation type is acceleration. When the target classification result is low risk, the text "The acceleration of the vehicle is normal and no special attention is required" is generated. When the target classification result is medium risk, the text "There are some problems with the acceleration of the vehicle, and it is recommended to check as soon as possible" is generated. When the target classification result is high risk, the text "There are serious problems with the acceleration of the vehicle, stop and check immediately" is generated, so that the text is determined as the vehicle supervision result corresponding to the target vehicle.

[0100] For example, after the corresponding text is jointly generated according to the target classification result and the target operation type, the vehicle supervision result can also be sent to the target associated user corresponding to the target vehicle via email or SMS to achieve the purpose of real-time supervision of the target vehicle. Alternatively, the vehicle supervision result can be converted into corresponding voice, and then the voice is broadcasted to achieve the purpose of real-time supervision of the target vehicle.

[0101] See also Figure 2 , Figure 2A vehicle condition monitoring device 200 based on a car is provided in an embodiment of the present application. The vehicle condition monitoring device 200 based on a car includes a data acquisition module 201, a data analysis module 202, a parameter processing module 203, a data classification module 204, a parameter determination module 205, a range determination module 206, an abnormality recognition module 207, and a result determination module 208, wherein the data acquisition module 201 is used to obtain initial sensor data corresponding to an initial sensor of an initial vehicle under a preset operation type, and perform data change analysis on the initial sensor data to obtain data change points corresponding to the initial sensor of the initial vehicle under the preset operation type; the data analysis module 202 is used to perform sensor association analysis based on the data change points to obtain target association relationships corresponding to the initial sensors of the initial vehicle under the preset operation type; the parameter processing module 203 is used to obtain target association relationships corresponding to the initial sensors of the initial vehicle under the preset operation type based on the target association relationships and the initial sensor data. The invention relates to a method for determining an initial interference parameter corresponding to the initial sensor based on the initial sensor data; a data classification module 204 is used to obtain the target sensor data corresponding to the target vehicle under the target sensor, and classify the operation type according to the target sensor data to obtain the target operation type corresponding to the target sensor data; a parameter determination module 205 is used to determine the target interference parameter corresponding to the target operation type from the initial interference parameter according to the target operation type and the preset operation type; a range determination module 206 is used to determine the data change range corresponding to the target vehicle under the target sensor according to the target interference parameter; an abnormality identification module 207 is used to identify the abnormality of the target sensor data according to the data change range, and obtain the target abnormality data corresponding to the target vehicle; a result determination module 208 is used to perform vehicle condition supervision on the target vehicle according to the target abnormality data to obtain the vehicle supervision result corresponding to the target vehicle.

[0102] In some implementations, the automobile-based vehicle condition monitoring device 200 may be applied to a terminal device.

[0103] It should be noted that, those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the automobile-based vehicle condition monitoring device 200 described above can refer to the corresponding process in the aforementioned automobile-based vehicle condition monitoring method embodiment, and will not be repeated here.

[0104] See also Figure 3 , Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention.

[0105] like Figure 3As shown, the terminal device 300 includes a processor 301 and a memory 302 , and the processor 301 and the memory 302 are connected via a bus 303 , such as an I2C (Inter-integrated Circuit) bus.

[0106] Specifically, the processor 301 is used to provide computing and control capabilities to support the operation of the entire terminal device. The processor 301 can be a central processing unit (CPU), and the processor 301 can also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0107] Specifically, the memory 302 may be a Flash chip, a read-only memory (ROM) disk, an optical disk, a USB flash drive, or a mobile hard disk.

[0108] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the embodiment of the present invention, and does not constitute a limitation on the terminal device to which the embodiment of the present invention is applied. The specific server may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0109] The processor is used to run a computer program stored in the memory, and implement any one of the vehicle condition supervision methods based on automobiles provided in the embodiments of the present invention when executing the computer program.

[0110] In one embodiment, the processor is used to run a computer program stored in the memory, and implements the following steps when executing the computer program:

[0111] Obtaining initial sensor data corresponding to an initial sensor of an initial vehicle under a preset operation type, and performing data change analysis on the initial sensor data to obtain data change points corresponding to the initial sensor of the initial vehicle under the preset operation type;

[0112] Performing sensor association analysis according to the data change points to obtain a target association relationship between the initial sensors of the initial vehicle under the preset operation type;

[0113] Determine an initial interference parameter corresponding to the initial sensor according to the target association relationship and the initial sensing data;

[0114] Obtaining target sensing data corresponding to the target vehicle under the target sensor, and classifying the operation type according to the target sensing data to obtain the target operation type corresponding to the target sensing data;

[0115] Determining a target interference parameter corresponding to the target operation type from the initial interference parameter according to the target operation type and the preset operation type;

[0116] Determine the data variation range corresponding to the target vehicle under the target sensor according to the target interference parameter;

[0117] Performing abnormal identification on the target sensor data according to the data variation range to obtain target abnormal data corresponding to the target vehicle;

[0118] The vehicle condition of the target vehicle is monitored according to the target abnormal data to obtain a vehicle supervision result corresponding to the target vehicle.

[0119] It should be noted that technicians in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the terminal device described above can refer to the corresponding process in the aforementioned automobile-based vehicle condition supervision method embodiment, and will not be repeated here.

[0120] An embodiment of the present invention also provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any one of the vehicle condition supervision methods based on automobiles provided in the description of the embodiment of the present invention.

[0121] The storage medium may be an internal storage unit of the terminal device described in the foregoing embodiment, such as a hard disk or memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc., equipped on the terminal device.

[0122] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware embodiment, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0123] It should be understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0124] The serial numbers of the embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. The above description is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A vehicle condition supervision method based on a car, characterized in that: The method comprises: Obtaining initial sensor data corresponding to an initial sensor of an initial vehicle under a preset operation type, and performing data change analysis on the initial sensor data to obtain data change points corresponding to the initial sensor of the initial vehicle under the preset operation type; Performing sensor association analysis according to the data change points to obtain a target association relationship between the initial sensors of the initial vehicle under the preset operation type; Determine an initial interference parameter corresponding to the initial sensor according to the target association relationship and the initial sensing data; Obtaining target sensing data corresponding to the target vehicle under the target sensor, and classifying the operation type according to the target sensing data to obtain the target operation type corresponding to the target sensing data; Determining a target interference parameter corresponding to the target operation type from the initial interference parameter according to the target operation type and the preset operation type; Determine the data variation range corresponding to the target vehicle under the target sensor according to the target interference parameter; Performing abnormal identification on the target sensor data according to the data variation range to obtain target abnormal data corresponding to the target vehicle; The vehicle condition of the target vehicle is monitored according to the target abnormal data to obtain a vehicle supervision result corresponding to the target vehicle.

2. The method according to claim 1, characterized in that: The performing data change analysis on the initial sensor data to obtain the data change point corresponding to the initial sensor of the initial vehicle under the preset operation type includes: Determine an initial change point, and cut the initial sensor data according to the initial change point to obtain first sensor data and second sensor data; Performing statistical data calculation on the first sensor data to obtain a corresponding first mean and a first standard deviation, and performing statistical data calculation on the second sensor data to obtain a corresponding second mean and a second standard deviation; determining a segmentation quality value corresponding to the initial change point according to the first mean, the first standard deviation, the second mean, and the second standard deviation; Obtaining a maximum value corresponding to the segmentation quality value as a target quality value, and determining a distribution probability of the target quality value corresponding to the segmentation quality value; Determining a segmentation state corresponding to the target quality value according to the distribution probability and a preset value, and determining the data change point from the initial change point according to the segmentation state and the target quality value; The segmentation quality value is obtained according to the following formula: ； in, represents the segmentation quality value corresponding to the i-th initial change point, represents the first mean value corresponding to the first sensing data corresponding to the i-th initial change point, represents the second mean value corresponding to the second sensing data corresponding to the i-th initial change point; represents the first standard deviation corresponding to the first sensing data corresponding to the i-th initial change point; represents the second standard deviation corresponding to the second sensing data corresponding to the i-th initial change point; represents the number of data corresponding to the first sensing data corresponding to the i-th initial change point, represents the number of data corresponding to the second sensing data corresponding to the i-th initial change point, Indicates taking the absolute value.

3. The method according to claim 1, characterized in that The initial sensor includes at least a first sensor and a second sensor, the initial sensor data includes at least first sensor data corresponding to the first sensor and second sensor data corresponding to the second sensor, the data change point includes at least a first change point corresponding to the first sensor data and a second change point corresponding to the second sensor data, and performing sensor association analysis according to the data change points to obtain the target association relationship corresponding to the initial sensors of the initial vehicle under the preset operation type includes: Obtaining a first acquisition time corresponding to the first change point from the first sensor data and obtaining a second acquisition time corresponding to the second change point from the second sensor data; Obtaining a first time quantity corresponding to the first acquisition time and a second time quantity corresponding to the second acquisition time; determining an initial correlation matrix corresponding to the first sensor and the second sensor according to the first time quantity and the second time quantity; The initial correlation matrix is ​​adjusted according to the first change point and the second change point to obtain a target correlation matrix corresponding to the first sensor and the second sensor of the initial vehicle under the preset operation type; A sensor association analysis is performed according to the target association matrix to obtain a target association relationship corresponding to the first sensor and the second sensor of the initial vehicle under the preset operation type.

4. The method according to claim 3, characterized in that: The performing sensor association analysis according to the target association matrix to obtain the target association relationship between the first sensor and the second sensor of the initial vehicle under the preset operation type includes: Obtaining a target representation vector between the first sensor and the second sensor according to the target association matrix; Performing weighted processing on the target characterization vector by a modulation function to obtain an initial characterization value corresponding to the first sensor and the second sensor; Performing data adjustment on the initial characterization value by using an adaptive filter to obtain a target characterization value corresponding to the first sensor and the second sensor; The target association relationship corresponding to the first sensor and the second sensor is determined according to the target representation value.

5. The method according to claim 3, characterized in that: The initial interference parameters at least include a first interference parameter corresponding to the first sensor and a second interference parameter corresponding to the second sensor, and determining the initial interference parameters corresponding to the initial sensor according to the target association relationship and the initial sensing data includes: Determine, from the initial sensors, a first associated sensor corresponding to the first sensor and a second associated sensor corresponding to the second sensor according to the target association relationship; Obtaining third sensor data corresponding to the first associated sensor and fourth sensor data corresponding to the second associated sensor; Calculating a first data difference value corresponding to adjacent time data in the first sensing data and calculating a second data difference value corresponding to adjacent time data in the second sensing data; Calculating a third data difference value corresponding to adjacent time data in the third sensing data and calculating a fourth data difference value corresponding to adjacent time data in the fourth sensing data; Calculating a fifth data difference value corresponding to the first sensor data and the third sensor data at the same time, and calculating a sixth data difference value corresponding to the second sensor data and the fourth sensor data at the same time; Calculating a first sample mean and a first sample variance for the first data difference, and calculating a third sample mean and a third sample variance for the third data difference; Determine a first mean value range corresponding to the first sensor data according to the first sample mean value and the first sample variance in combination with a normal distribution; Determine a third mean range corresponding to the third sensor data according to the third sample mean and the third sample variance in combination with a normal distribution; Calculating a fifth sample mean and a fifth sample variance for the fifth data difference, and determining a fifth mean range corresponding to the first sensing data and the first associated sensor according to the fifth sample mean and the fifth sample variance in combination with a normal distribution; The first mean value range is adjusted according to the third mean value range and the fifth mean value range to obtain the first interference parameter corresponding to the first sensor; Calculating a second sample mean and a second sample variance for the second data difference, and calculating a fourth sample mean and a fourth sample variance for the fourth data difference; Determine a second mean range corresponding to the second sensor data according to the second sample mean and the second sample variance in combination with a normal distribution; Determine a fourth mean range corresponding to the fourth sensor data according to the fourth sample mean and the fourth sample variance in combination with a normal distribution; Calculating a sixth sample mean and a sixth sample variance for the sixth data difference, and determining a sixth mean range corresponding to the second sensing data and the second associated sensor according to the sixth sample mean and the sixth sample variance in combination with a normal distribution; The second mean value range is adjusted according to the fourth mean value range and the sixth mean value range to obtain the second interference parameter corresponding to the second sensor.

6. The method according to claim 1, characterized in that The step of performing abnormality identification on the target sensor data according to the data variation range to obtain target abnormality data corresponding to the target vehicle includes: Determine the target data expectation corresponding to the target sensor, and perform difference calculation based on the target sensor data and the target data expectation to obtain a deviation value corresponding to the target sensor data; Determining a degree of deviation corresponding to the target sensor data according to the deviation value and the data variation range; Determining initial abnormal data from the target sensor data according to the degree of deviation, and obtaining remaining sensor data after removing the initial abnormal data from the target sensor data; Performing distance calculation on the initial abnormal data to obtain distance information corresponding to the initial abnormal data and performing local density calculation on the initial abnormal data to obtain density information corresponding to the initial abnormal data; Performing data distribution analysis on the distance information to obtain corresponding first probability distribution information and performing data distribution analysis on the density information to obtain corresponding second probability distribution information; Fusing the first probability distribution information and the second probability distribution information to determine a target abnormality probability corresponding to the initial abnormal data; determining first abnormal data from the initial abnormal data according to the target abnormal probability; Classifying the remaining sensor data using a support vector machine to obtain first classified data and second classified data corresponding to the remaining sensor data; Calculating a first quantity corresponding to the first classified data and a second quantity corresponding to the second classified data, and determining second abnormal data corresponding to the remaining sensor data according to the first quantity and the second quantity; The first abnormal data and the second abnormal data are combined to obtain the target abnormal data corresponding to the target vehicle.

7. The method according to claim 1, characterized in that The step of performing vehicle condition monitoring on the target vehicle according to the target abnormal data to obtain a vehicle monitoring result corresponding to the target vehicle includes: Using the data analysis layer of the anomaly recognition model to perform anomaly characterization on the target anomaly data to obtain an initial characterization vector corresponding to the target sensor; Using the data fusion layer of the abnormal recognition model to perform data fusion on the initial representation vector using the target association relationship to obtain a target representation vector corresponding to the target vehicle; Using the data classification layer of the abnormal recognition model to classify the target representation vector, the target classification result corresponding to the target vehicle is obtained; The vehicle supervision result corresponding to the target vehicle is determined according to the target classification result and the target operation type.

8. A vehicle condition monitoring device based on a car, characterized in that: include: A data acquisition module, used to obtain initial sensor data corresponding to an initial sensor of an initial vehicle under a preset operation type, and perform data change analysis on the initial sensor data to obtain data change points corresponding to the initial sensor of the initial vehicle under the preset operation type; A data analysis module, configured to perform sensor association analysis according to the data change points to obtain a target association relationship corresponding to the initial sensors of the initial vehicle under the preset operation type; A parameter processing module, used for determining an initial interference parameter corresponding to the initial sensor according to the target association relationship and the initial sensing data; A data classification module is used to obtain target sensing data corresponding to the target vehicle under the target sensor, and classify the operation type according to the target sensing data to obtain the target operation type corresponding to the target sensing data; a parameter determination module, configured to determine a target interference parameter corresponding to the target operation type from the initial interference parameter according to the target operation type and the preset operation type; A range determination module, used to determine the data change range corresponding to the target vehicle under the target sensor according to the target interference parameter; An abnormality identification module, used for identifying abnormalities of the target sensor data according to the data variation range, and obtaining target abnormality data corresponding to the target vehicle; The result determination module is used to perform vehicle condition supervision on the target vehicle according to the target abnormal data to obtain a vehicle supervision result corresponding to the target vehicle.

9. A terminal device, characterized in that: The terminal device includes a processor and a memory; The memory is used to store computer programs; The processor is configured to execute the computer program and implement the vehicle condition supervision method based on a vehicle according to any one of claims 1 to 7 when executing the computer program.

10. A computer storage medium for computer storage, characterized in that: The computer storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the automobile-based vehicle condition supervision method described in any one of claims 1 to 7.

Images