An intelligent network connected vehicle data missing and error transmission detection method

Abstract

The present application relates to a kind of intelligent network connection car data leak transmission error detection method, it is related to the field of Internet of Vehicles big data, including: cloud end issues data acquisition configuration information to vehicle end;Cloud end receives the data collected according to data acquisition configuration information by each domain controller in vehicle end, and the daily total quantity count information recorded by vehicle end;Cloud end carries out data statistical characteristics and logical feature identification according to the data collected according to data acquisition configuration information by each domain controller in vehicle end and the daily total quantity count information recorded by vehicle end, then according to the statistical characteristic rule and logical feature rule configured by cloud end, confirm data leak transmission error situation.The present application is adapted to include but not limited to whole vehicle signal, man-machine interaction burying point, image, video, point cloud, intermediate variable, data file and other data types on cloud leak transmission detection;Detection range can cover vehicle networking full period, all business scenarios, support batch data upload, and configure cloud end monitoring to data leak transmission carries out insight and early warning.

Description

Technical Field

[0001] This invention relates to the field of vehicle networking big data technology, and more specifically, to a method for detecting data leakage and mistransmission in intelligent connected vehicles. Background Technology

[0002] As automobiles become increasingly intelligent, smart driving and smart cockpit functions rely more and more on diverse data acquisition and cloud uploading, rather than being limited to traditional CAN data acquisition and cloud uploading. Identifying data that should have been transmitted but wasn't, or data that was missed or over-transmitted due to network signal issues or other abnormal conditions (such as abnormal vehicle power outages), is crucial for data acquisition testing, data cleaning, and tracing the causes of data integrity breaches.

[0003] An automatic detection method for frame loss rate in CAN data upload of vehicle-to-everything (V2X) networks is disclosed in the prior art, but this solution has the following drawbacks:

[0004] 1. This existing technology is only suitable for the scenario of missing frame monitoring of CAN message data upload in the Internet of Vehicles, and it relies heavily on CANID encoding, and cannot support more flexible data collection and cloud uploading methods such as MQTT.

[0005] 2. The existing technology can only determine the missing frame data based on the total number of uploads, and cannot quickly and specifically locate the domain controller that collected the data, which is not conducive to more accurate data cleaning in the cloud and further analysis of the specific modules and causes of missing frames.

[0006] 3. The existing counter only counts within the range of one engine start-stop cycle, and cannot perform frame missing checks on the collected data outside the start-stop period.

[0007] 5. This existing technology does not support monitoring of data that should have been transmitted but was not, based on business settings, and cannot be applied to the field of automated data acquisition testing.

[0008] 6. The counter in this existing technology can only increment one at a time, without considering the case of batch data upload. Summary of the Invention

[0009] The purpose of this invention is to provide a method for detecting data leakage and mistransmission in intelligent connected vehicles, so as to solve the problem that the existing technology only adapts to the scenario of missing frame rate of CAN message data upload and cannot realize the detection and monitoring of leakage and mistransmission of general collected data.

[0010] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0011] A method for detecting data leakage and mistransmission in intelligent connected vehicles includes the following steps:

[0012] The cloud sends data collection configuration information to the vehicle.

[0013] The cloud receives data collected by each domain controller in the vehicle according to the data collection configuration information, as well as the daily total count information recorded by the vehicle.

[0014] The cloud performs data statistical and logical feature identification based on the data collected by each domain controller in the vehicle according to the data collection configuration information and the daily total count information recorded by the vehicle. Then, it confirms the data omission and mistransmission based on the statistical and logical feature rules configured in the cloud.

[0015] Based on the above technical means, the data collected is not parsed at the vehicle end, but only the daily total count information of the collected data is recorded. The collected data and the corresponding daily total count information are transmitted to the cloud, where the cloud performs data parsing and confirms the data omission and mistransmission. It can adapt to cloud omission detection of data types including but not limited to whole vehicle signals, human-machine interaction points, images, videos, point clouds, intermediate variables, and data files. Since the uploaded data is the daily total count information, the detection scope can cover all time periods and all business scenarios of vehicle networking.

[0016] Furthermore, the data acquisition configuration information includes data item definitions, data item acquisition triggering strategies, data collection categories, and the vehicle range covered by the data acquisition configuration. The data item acquisition triggering strategies include periodic acquisition, change acquisition, or event-triggered acquisition.

[0017] Furthermore, the daily total count information recorded on the vehicle includes the timestamp of the last upload time within the day and the number of times each type of data was uploaded. If the data of a certain type is uploaded as a single piece, the count is incremented by 1 after each upload. If N pieces of data are uploaded in batch, the count is incremented by N.

[0018] The daily total count information includes the number of times each type of data is uploaded. Therefore, whether data is uploaded sequentially or in batches, the number of times each type of data is uploaded can be recorded. Batch data upload is supported, and even in the case of batch data upload, it is still possible to confirm the omission or mistransmission of collected data.

[0019] Furthermore, the cloud receives data collected by each domain controller in the vehicle according to the data collection configuration information, as well as the daily total count information recorded by the vehicle, specifically as follows:

[0020] The vehicle receives data collection and configuration information from the cloud, parses it, and distributes it to various domain controllers within the vehicle.

[0021] When each domain controller triggers the data item collection trigger policy, it collects data, records the collected data and collection time, and waits for uploading.

[0022] Check if the intraday counter corresponding to the uploaded data date exists in the vehicle-side cache. If it exists, upload the collected data and collection time, update the intraday counter, and record the last upload time of the day. If it does not exist, create the intraday counter corresponding to the uploaded data date in the vehicle-side cache.

[0023] After the daily settlement is completed, the full count information of the daily settlement recorded on the vehicle is uploaded, and the memory occupied by the daily counter on the vehicle is cleared.

[0024] Furthermore, the criteria for determining daily settlement are as follows:

[0025] If the difference between the current vehicle system time and the last upload time of the day exceeds the preset maximum waiting time, the day will be settled.

[0026] Furthermore, before creating the intraday counter corresponding to the uploaded data date in the vehicle-side cache, it is necessary to determine whether the date of the collection time is consistent with the date of the vehicle-side system. If they are consistent, an intraday counter corresponding to the uploaded data date is created in the vehicle-side cache. If they are inconsistent, the incremental count information of the collected data corresponding to the collection time is sent to the cloud, and the cloud uses the incremental count information to update the daily full count information corresponding to the date of the collection time.

[0027] Furthermore, the data statistical features include the statistical quantity of collected data, collection period, mean, standard deviation, median, quartiles, extreme values, and skewness. The statistical feature rules corresponding to the data statistical features are: whether the statistical quantity of collected data is consistent between the vehicle end and the cloud end; whether the collection period, mean, standard deviation, median, quartiles, extreme values, and skewness of the collected data meet the preset statistical distribution characteristics. If the statistical feature rules are not met, there is a situation of data omission or mistransmission.

[0028] Furthermore, the logical features include business logic, user ID, and data elements. The logical feature rules corresponding to the logical features are as follows: whether the end time of data collection is later than the start time, whether only heartbeat data is sent after the vehicle-end goes into sleep mode, whether the user ID has been registered, whether the collected data is complete, whether the format of the collected data meets the preset format requirements, whether the numerical precision of the collected data meets the preset precision requirements, whether the numerical range of continuous collected data is within the preset range, and whether the numerical value of discrete collected data is taken from the set of set finite elements. If the logical feature rules are not met, there is a situation of data omission or incorrect transmission.

[0029] Furthermore, after confirming that data has been missed or transmitted incorrectly, the following steps are also included:

[0030] The cloud performs data analysis on the collected data and generates a quantitative analysis report, which analyzes and displays data loss or mistransmission.

[0031] Furthermore, the cloud platform performs data analysis on the collected data, specifically employing machine learning anomaly detection algorithms and time series monitoring algorithms to predict data loss or mistransmission. The machine learning anomaly detection algorithm is an isolated forest algorithm, a local outlier algorithm, a support vector machine algorithm, or a K-means clustering algorithm, while the time series monitoring algorithm is an exponential smoothing algorithm, an autoregressive moving average model algorithm, a linear regression algorithm, or a neural network model algorithm.

[0032] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0033] This invention provides a method for detecting data leakage and mistransmission in intelligent connected vehicles. It can adapt to cloud-based leakage detection of various data types, including but not limited to vehicle signals, human-machine interaction data points, images, videos, point clouds, intermediate variables, and data files. The detection scope can cover all time periods and all business scenarios of vehicle networking, supports batch data upload, and can be configured with cloud monitoring to gain insight and early warning of data leakage. Attached Figure Description

[0034] Figure 1 This is a flowchart illustrating a method for detecting data leakage and mistransmission in intelligent connected vehicles proposed in this embodiment of the invention.

[0035] Figure 2 This is a schematic diagram of the data flow in the intelligent connected vehicle data leakage and mistransmission detection method proposed in this embodiment of the invention.

[0036] Figure 3 This is a flowchart illustrating another method for detecting data leakage and mistransmission in intelligent connected vehicles proposed in this embodiment of the invention.

[0037] Figure 4 This is a flowchart illustrating the specific implementation of the intelligent connected vehicle data leakage and mistransmission detection method proposed in this embodiment of the invention. Detailed Implementation

[0038] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0039] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0040] Existing technologies are mainly designed for CAN message data upload missing frame monitoring scenarios in vehicle-to-everything (V2X) networks. They rely on CANID encoding and cannot support more flexible data acquisition and cloud-based methods such as MQTT. They are also unsuitable for detecting other types of data upload missing or incorrect transmissions.

[0041] To address the aforementioned issues, this embodiment of the invention records only the daily total count information of uploaded data on the vehicle side, and parses the data in the cloud to achieve detection of missed or incorrect transmissions of all types of data.

[0042] The following is a detailed explanation:

[0043] This invention proposes a method for detecting data loss and mistransmission in intelligent connected vehicles, such as... Figure 1 As shown, it includes the following steps:

[0044] The cloud sends data collection configuration information to the vehicle.

[0045] The cloud receives data collected by each domain controller in the vehicle according to the data collection configuration information, as well as the daily total count information recorded by the vehicle.

[0046] The cloud performs data statistical and logical feature identification based on the data collected by each domain controller in the vehicle according to the data collection configuration information and the daily total count information recorded by the vehicle. Then, it confirms the data omission and mistransmission based on the statistical and logical feature rules configured in the cloud.

[0047] In specific embodiments, such as Figure 2 The diagram illustrates the data flow of a method for detecting data loss or mistransmission in intelligent connected vehicles. A data acquisition strategy management system is configured in the cloud to publish data acquisition configuration information. A cloud-based transmission detection and monitoring system receives the uploaded data and performs statistical and logical feature identification, then confirms data loss or mistransmission based on the configured statistical and logical feature rules. A cloud gateway establishes the network connection between the cloud and the vehicle. The vehicle includes various domain controllers, a vehicle-side data acquisition and upload module, and a vehicle-side gateway. The vehicle-side data acquisition and upload module uploads data collected by each domain controller and the daily total count information recorded by the vehicle. The vehicle-side gateway and the cloud gateway work together to establish the network connection between the cloud and the vehicle.

[0048] In a specific embodiment, the data acquisition configuration information includes data item definitions, data item acquisition triggering strategies, data collection categories, and the vehicle range covered by the data acquisition configuration. The data item acquisition triggering strategies include periodic acquisition, change-triggered acquisition, or event-triggered acquisition. After the corresponding vehicle is powered on, the vehicle connects to the network through the vehicle-side gateway, retrieves the acquisition configuration information from the cloud, and forwards it to each domain controller on the vehicle. Each domain controller triggers the acquisition strategy and sends the data to the vehicle-side acquisition and upload module through the vehicle-side network. The vehicle-side acquisition and upload module puts the uploaded content into a message queue for uploading.

[0049] In a specific embodiment, the daily total count information recorded by the vehicle includes the timestamp of the last upload time within the day and the number of uploads for each type of data. If the data of a certain type is uploaded as a single piece, the count is incremented by 1 after each upload. If N pieces of data are uploaded in batch, the count is incremented by N.

[0050] In a specific embodiment, the cloud receives data collected by each domain controller in the vehicle terminal according to the data collection configuration information, as well as the daily total count information recorded by the vehicle terminal, specifically as follows:

[0051] The vehicle receives data collection and configuration information from the cloud, parses it, and distributes it to various domain controllers within the vehicle.

[0052] When each domain controller triggers the data item collection triggering policy, it collects data and records the collected data and collection time, waiting to upload. In a specific example, the collected data and collection time record are transmitted to the message upload queue for uploading.

[0053] Check if the intraday counter corresponding to the uploaded data date exists in the vehicle-side cache. If it exists, upload the collected data and collection time, update the intraday counter, and record the last upload time of the day. If it does not exist, create the intraday counter corresponding to the uploaded data date in the vehicle-side cache.

[0054] After the daily settlement is completed, the full count information of the daily settlement recorded on the vehicle is uploaded, and the memory occupied by the daily counter on the vehicle is cleared.

[0055] In a specific embodiment, the criteria for determining the daily settlement are as follows:

[0056] If the difference between the current vehicle system time and the last upload time of the day exceeds the preset maximum waiting time, the day will be settled.

[0057] In a specific embodiment, before creating the intraday counter corresponding to the data upload date in the vehicle-side cache, it is necessary to determine whether the date of the collection time is consistent with the date of the vehicle-side system. If they are consistent, an intraday counter corresponding to the data upload date is created in the vehicle-side cache. If they are inconsistent, the incremental count information of the collected data corresponding to the collection time is sent to the cloud, and the cloud uses the incremental count information to update the daily full count information corresponding to the date of the collection time.

[0058] In a specific embodiment, the data statistical features include the statistical quantity of collected data, collection period, mean, standard deviation, median, quartiles, extreme values, and skewness. The statistical feature rules corresponding to the data statistical features are: whether the statistical quantity of collected data is consistent between the vehicle end and the cloud end; whether the collection period, mean, standard deviation, median, quartiles, extreme values, and skewness of the collected data meet the preset statistical distribution characteristics. If the statistical feature rules are not met, there is a situation of data omission or mistransmission.

[0059] In a specific embodiment, the logical features include business logic, user ID, and data elements. The logical feature rules corresponding to the logical features are as follows: whether the end time of data collection is later than the start time, whether only heartbeat data is sent after the vehicle-end goes into sleep mode, whether the user ID has been registered, whether the collected data is complete, whether the format of the collected data meets the preset format requirements, whether the numerical precision of the collected data meets the preset precision requirements, whether the numerical range of continuous collected data is within the preset range, and whether the numerical value of discrete collected data is taken from the set of set finite elements. If the logical feature rules are not met, there is a situation of data omission or incorrect transmission.

[0060] Table 1 provides explanations of the statistical and logical characteristics of the data, as well as the corresponding statistical and logical characteristic rules.

[0061] Table 1

[0062]

[0063]

[0064] Figure 3 Another method for detecting data loss and mistransmission in intelligent connected vehicles, such as Figure 2 As shown, after confirming that data has been missed or transmitted incorrectly, the following steps are also included:

[0065] The cloud performs data analysis on the collected data and generates a quantitative analysis report, which analyzes and displays data loss or mistransmission.

[0066] In a specific embodiment, the cloud performs data analysis on the collected data, specifically using machine learning anomaly detection algorithms and time series monitoring algorithms to predict data loss or mistransmission. The machine learning anomaly detection algorithm is an isolated forest algorithm, local outlier algorithm, support vector machine algorithm, or K-means clustering algorithm, and the time series monitoring algorithm is an exponential smoothing algorithm, autoregressive moving average model algorithm, linear regression algorithm, or neural network model algorithm.

[0067] In specific embodiments, machine learning anomaly detection algorithms use statistical and logical features of data as input sources to further predict potential missed or incorrect transmissions. Common machine learning anomaly detection algorithms include:

[0068] Isolation Forest Algorithm: This algorithm identifies outliers by constructing a random forest. Random forest is an ensemble learning algorithm based on decision trees. Isolation Forest constructs a random forest by randomly selecting data features and randomly partitioning the data, and determines whether a data point is an outlier based on the path length of the data within the random forest.

[0069] Local Outlier Factor Algorithm: This algorithm identifies outliers by calculating the density around each data point. If the density around a data point is very low, it is likely an outlier.

[0070] Support Vector Machine (SVM) algorithm: This algorithm can be used for binary classification problems to separate normal and outlier data. SVM separates the data by finding a hyperplane, and the further the hyperplane is from outliers, the better.

[0071] K-means clustering algorithm: This algorithm can be used for cluster analysis to divide data into multiple clusters. If a data point does not belong to any cluster, it may be an outlier.

[0072] More specifically, let's take the Support Vector Machine (SVM) algorithm as an example:

[0073] First, historically collected data needs to be analyzed to extract common statistical or logical features, which are then used as training data. The Support Vector Machine (SVM) algorithm can learn from this training data to build an anomaly detection model, which can then be used to identify outliers in unknown data.

[0074] The specific implementation process is as follows:

[0075] 1. Data preprocessing: Transform the historical daily statistical or logical features of the collected data into feature vectors, including frequency, mean, peak value, skewness, distribution, etc.

[0076] 2. Model Training: Using feature vectors, manually label data anomalies, train a support vector machine model, and establish an anomaly detection model.

[0077] 3. Anomaly Detection: Input the feature vector of the day into the trained model and calculate its anomaly score. The higher the anomaly score, the more likely the data is to contain anomalies.

[0078] 4. Output Results: The probability of abnormal output data can be set with a threshold according to specific application requirements. If the threshold is exceeded, it is considered abnormal and appropriate processing will be carried out.

[0079] Furthermore, by integrating the above-mentioned machine learning algorithms and using secondary random forest modeling, the overall robustness and accuracy of the model can be improved.

[0080] In specific embodiments, time series monitoring algorithms perform monitoring based on time sequence and logical correlation to further identify missed or incorrect data transmission from a business perspective. Commonly used time series monitoring algorithms for detecting missed or incorrect data transmission include:

[0081] Exponential smoothing algorithm: This algorithm can be used to predict future values ​​and also to detect data errors. Exponential smoothing predicts future values ​​by weighting and averaging historical data. If there is a large difference between the predicted and actual values, it may indicate missing or incorrect data.

[0082] Autoregressive moving average model algorithm: This algorithm can be used to build statistical models of data, predict future values ​​by fitting historical data, and can also be used to detect data errors. If there is a large difference between the predicted value fitted by the model and the actual value, it may indicate that data has been missed or misrepresented.

[0083] Linear regression algorithm: This algorithm can be used to build a linear model of data, predicting future values ​​by fitting historical data, and can also be used to detect data errors. If there is a large difference between the predicted value fitted by the model and the actual value, it may indicate that data has been missed or misrepresented.

[0084] Neural network model algorithms: These algorithms can be used to build nonlinear models of data, predicting future values ​​by fitting historically collected data. They can also be used to detect missing or incorrect data transmission. If there is a large discrepancy between the predicted value fitted by the model and the actual value, it may indicate missing or incorrect data.

[0085] It is important to note that time series algorithms require a large amount of historical data to build models and detect data errors. They also require data preprocessing and feature engineering to improve the accuracy of the algorithms.

[0086] Taking linear regression as an example, linear regression mainly utilizes the linear growth of connected vehicles after a car model is launched, resulting in a linear increase in the amount of data collected. The specific implementation follows these steps:

[0087] 1. Data preprocessing: Convert the historical daily upload counts of the collected data into feature vectors.

[0088] 2. Model training: Train a linear regression model using historical daily occurrences.

[0089] 3. Anomaly Detection: Input the number of data uploads for the day into the trained model and calculate the error between the predicted and actual values. The larger the error, the more likely the data is to be anomaly-prone.

[0090] 4. Output Results: The probability of abnormal output data can be set with a threshold according to specific application requirements. If the threshold is exceeded, it is considered abnormal and appropriate processing will be carried out.

[0091] Figure 4 The flowchart below shows the specific implementation of the intelligent connected vehicle data leakage and mistransmission detection method proposed in this invention. Figure 4 Specifically:

[0092] The cloud sends the collected configuration information to the vehicle.

[0093] The vehicle-side system parses the collected configuration information and distributes it to each domain controller.

[0094] The domain controller triggers the data item collection strategy and sends the collected data and collection time to the vehicle transmission module;

[0095] The vehicle transmission module puts the collected data and collection time into the upload queue;

[0096] Check if the intraday counter corresponding to the uploaded data date exists in the vehicle-side cache. If it exists, update the intraday counter and record the last upload time of the day.

[0097] If it does not exist, check if the date of the uploaded data is consistent with the date of the vehicle system. If they are consistent, create an intraday counter corresponding to the date of the uploaded data in the vehicle cache. If they are inconsistent, send the incremental technical information corresponding to the date of the uploaded data to the cloud, and the cloud updates the counter information corresponding to that date.

[0098] Check if the difference between the vehicle-side system time and the last upload time of the day does not exceed the preset maximum waiting time. If so, continue to upload data to the cloud. If not, upload the day's counter information to the cloud. After saving it in the cloud, clear the memory occupied by the day's counter on the vehicle side.

[0099] The cloud receives and statistically analyzes data, and formulates rules for missing or incorrect transmission based on the statistical and logical characteristics that should be met by the business rules.

[0100] For complex statistical features involving multiple signals, machine learning modeling is used to identify suspected missed or incorrect transmission phenomena through anomaly monitoring and analysis.

[0101] Data reports are used to display instances of data omissions or errors, and important instances of omissions or errors are communicated to relevant personnel for handling.

[0102] The same or similar labels correspond to the same or similar parts;

[0103] The terms used to describe positional relationships in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent.

[0104] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for detecting data leakage and mistransmission in intelligent connected vehicles, characterized in that, Includes the following steps: The cloud sends data collection configuration information to the vehicle. The cloud receives data collected by each domain controller in the vehicle according to the data collection configuration information, as well as the daily total count information recorded by the vehicle. The cloud performs data statistical and logical feature identification based on the data collected by each domain controller in the vehicle according to the data collection configuration information and the daily total count information recorded by the vehicle. Then, it confirms the data omission and mistransmission based on the statistical and logical feature rules configured in the cloud. The daily total count information recorded on the vehicle includes the timestamp of the last upload time within the day and the number of uploads for each type of data. If the data of a certain type is uploaded as a single piece, the count is incremented by 1 after each upload. If N pieces of data are uploaded in batch, the count is incremented by N. The cloud-based reception system receives data collected by each domain controller in the vehicle terminal according to the data collection configuration information, as well as the daily total count information recorded by the vehicle terminal, specifically as follows: The vehicle receives data collection and configuration information from the cloud, parses it, and distributes it to various domain controllers within the vehicle. When each domain controller triggers the data item collection trigger policy, it collects data, records the collected data and collection time, and waits for uploading. Check if the intraday counter corresponding to the uploaded data date exists in the vehicle-side cache. If it exists, upload the collected data and collection time, update the intraday counter, and record the last upload time of the day. If it does not exist, create a daily counter corresponding to the uploaded data date in the vehicle-side cache; After the daily settlement is completed, the full daily count information recorded on the vehicle is uploaded, and the memory occupied by the daily counter on the vehicle is cleared.

2. The method for detecting data leakage and mistransmission in intelligent connected vehicles according to claim 1, characterized in that, The data acquisition configuration information includes data item definitions, data item acquisition triggering strategies, data collection categories, and the vehicle range covered by the data acquisition configuration. The data item acquisition triggering strategies include periodic acquisition, change acquisition, or event-triggered acquisition.

3. The method for detecting data leakage and mistransmission in intelligent connected vehicles according to claim 1, characterized in that, The criteria for determining the daily settlement are as follows: If the difference between the current vehicle system time and the last upload time of the day exceeds the preset maximum waiting time, the day will be settled.

4. The method for detecting data leakage and mistransmission in intelligent connected vehicles according to claim 1 or 3, characterized in that, Before creating the intraday counter corresponding to the data upload date in the vehicle-side cache, it is necessary to determine whether the date of the collection time is consistent with the date of the vehicle-side system. If they are consistent, an intraday counter corresponding to the data upload date is created in the vehicle-side cache. If they are inconsistent, the incremental count information of the collected data corresponding to the collection time is sent to the cloud. The cloud uses the incremental count information to update the daily full count information corresponding to the date of the collection time.

5. The method for detecting data leakage and mistransmission in intelligent connected vehicles according to claim 1, characterized in that, The statistical features of the data include the number of collected data, the collection period, the mean, the standard deviation, the median, the quartiles, the extreme values, and the skewness. The statistical feature rules corresponding to the statistical features are: whether the number of collected data recorded on the vehicle and the cloud are consistent, and whether the collection period, mean, standard deviation, median, quartiles, extreme values, and skewness of the collected data meet the preset statistical distribution characteristics. If the statistical feature rules are not met, there is a situation of data being missed or transmitted incorrectly.

6. The method for detecting data leakage and mistransmission in intelligent connected vehicles according to claim 5, characterized in that, The logical features include business logic, user ID, and data elements. The logical feature rules corresponding to the logical features are as follows: whether the end time of data collection is later than the start time, whether only heartbeat data is sent after the vehicle-end goes into sleep mode, whether the user ID has been registered, whether the collected data is complete, whether the format of the collected data meets the preset format requirements, whether the numerical precision of the collected data meets the preset precision requirements, whether the numerical range of continuous collected data is within the preset range, and whether the numerical value of discrete collected data is taken from the set of set finite elements. If the logical feature rules are not met, there is a situation of data omission or incorrect transmission.

7. The method for detecting data leakage and mistransmission in intelligent connected vehicles according to any one of claims 1 to 3, 5, and 6, characterized in that, After confirming that data was missed or transmitted incorrectly, the following steps are also included: The cloud performs data analysis on the collected data and generates a quantitative analysis report, which analyzes and displays data loss or mistransmission.

8. The method for detecting data leakage and mistransmission in intelligent connected vehicles according to claim 7, characterized in that, The cloud platform performs data analysis on the collected data, specifically using machine learning anomaly detection algorithms and time series monitoring algorithms to predict data loss or mistransmission. The machine learning anomaly detection algorithms include isolated forest, local outlier, support vector machine, or K-means clustering, while the time series monitoring algorithms include exponential smoothing, autoregressive moving average, linear regression, or neural network models.

Images