A data processing method and system, cloud server and vehicle

Abstract

The present application provides a data processing method and system, a cloud server and a vehicle, including: receiving a data packet transmitted wirelessly by a vehicle, and decrypting the data packet; wherein the data packet is generated by encapsulating and encrypting traffic data in a traffic road area; performing abnormal clustering of data points based on the decryption result of the data packet, and generating an abnormal area based on the abnormal clustering result; wherein the data point includes the geographic location coordinate information of the data packet; based on the geographic location coordinate information of the vehicle, pushing warning information to vehicles within a preset distance range of the abnormal area. The present application monitors traffic data in real time, and uses wireless transmission technology to realize timely reporting and sharing of abnormal information, shortening the time difference from detection to response, and timely pushing warning information through the cloud server, which enables drivers to take risk avoidance measures in advance and reduce the risk of traffic accidents caused by traffic facility failures.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a data processing method and system, a cloud server and a vehicle. Background Art

[0002] For traffic facilities installed on traffic roads (such as traffic lights, street lights, etc.), the current traffic management plan relies on manual regular fault inspections and active reporting of faults by drivers or pedestrians. This management plan is not only inefficient but also slow to respond and has a lag. It may also result in the corresponding faults not being recorded in a timely manner due to human negligence or non-reporting, increasing the traffic risk of vehicles on the corresponding traffic roads.

[0003] In addition, the current traffic management scheme lacks an effective data sharing and coordination mechanism, which results in the failure of traffic facilities to be disseminated in a timely manner, resulting in information islands, which reduces the overall detection efficiency and response speed of the traffic management system. Moreover, in the face of complex changes in the traffic environment, the current traffic management scheme is often difficult to respond flexibly, and has a weak ability to integrate new technologies and new standards. At the same time, it lacks efficient data capabilities and is difficult to effectively integrate and deeply analyze large amounts of heterogeneous data, affecting the accuracy and timeliness of anomaly detection. Therefore, the current traffic management scheme has obvious deficiencies in terms of automation level, system integration, data processing capabilities, and flexibility in dealing with complex environments. Summary of the invention

[0004] In view of the above-mentioned shortcomings of the prior art, the purpose of the present application is to provide a data processing method and system, a cloud server and a vehicle to solve the technical problems existing in the prior art.

[0005] To achieve the above purpose and other related purposes, the present application provides a data processing method, comprising the following steps:

[0006] Receiving a data packet transmitted wirelessly by a vehicle and decrypting the data packet; wherein the data packet is generated by encapsulating and encrypting traffic data in a traffic road area;

[0007] Performing data point abnormal clustering based on the data packet decryption result, and generating abnormal areas based on the abnormal clustering result; wherein the data point includes the geographic location coordinate information of the data packet;

[0008] Based on the geographic location coordinate information of the vehicle, early warning information is pushed to vehicles within a preset distance range of the abnormal area.

[0009] In one embodiment of the present application, the process of performing abnormal clustering of data points based on the decryption result of the data packet and generating abnormal regions based on the abnormal clustering result includes:

[0010] Performing integrity verification on the data packet based on the data packet decryption result; and,

[0011] After the data packet passes integrity verification, generating the data point based on geographic location coordinate information of the data packet;

[0012] Select a data point as a target data point, and record the minimum number of neighborhood data points of the target data point with the target data point as the center, and mark the target data point as a core data point when the minimum number of neighborhood data points is greater than or equal to a first threshold, and add the corresponding data point to an abnormal cluster; wherein the minimum neighborhood data point includes data points located within a preset neighborhood radius of the target data point;

[0013] The abnormal region is generated by performing regional association on all boundary data points in the abnormal clustering result; wherein the boundary data point is located within a preset neighborhood radius of the core data point and is not a core data point.

[0014] In one embodiment of the present application, after generating the abnormal area, the method further includes:

[0015] Calculate the path length of each data point in the abnormal area using a data point abnormality detection lighting abnormality detection model established in advance or in real time;

[0016] The anomaly score of each data point is calculated based on the path length of each data point, and when the anomaly score is greater than a preset score threshold, the corresponding data point is marked as a false alarm data point; and the associated area corresponding to the false alarm data point is removed from the abnormal area to obtain the real abnormal area.

[0017] In one embodiment of the present application, based on the geographic location coordinate information of the vehicle, the process of pushing warning information to vehicles within a preset distance range of the abnormal area includes:

[0018] Generate a predicted driving trajectory of the current vehicle based on the historical geographic location coordinate information and the real-time geographic location coordinate information of the current vehicle;

[0019] When the predicted driving trajectory intersects the abnormal area, determining the center of the abnormal area, the radius or the side length of the abnormal area according to the geometric shape of the abnormal area;

[0020] Based on the real-time geographic location coordinate information of the current vehicle, the distance value between the current vehicle and the center of the abnormal area is calculated, and when the distance value is less than or equal to the radius or side length of the abnormal area, the current vehicle is marked as a warning vehicle, and warning information is generated and pushed to the warning vehicle; wherein the warning information includes: lighting data of the traffic road area and vehicle driving suggestions.

[0021] The present application also provides a data processing method, which is applied to a vehicle, and the method comprises the following steps:

[0022] Acquire traffic data of a target area, and encapsulate and encrypt the traffic data to generate a data packet; wherein the target area includes a traffic road area determined in advance or in real time;

[0023] Wirelessly transmitting the data packet to a cloud server so that the cloud server decrypts the data packet, performs data point anomaly clustering based on the data packet decryption result, and generates anomaly regions based on the anomaly clustering result;

[0024] When the vehicle is within a preset distance range from the abnormal area, the warning information generated and pushed by the cloud server is received.

[0025] In one embodiment of the present application, if the traffic data includes lighting data, the process of acquiring the lighting data of the target area includes:

[0026] Calculate the light intensity change rate of the target area according to the ambient light intensity of the target area, the average light intensity within the sliding time window length, and the light intensity standard deviation within the sliding time window length; wherein the ambient light intensity is generated by a light intensity detector, and the light intensity detector is arranged on a vehicle traveling in the target area;

[0027] Calculating the flicker frequency of the lighting equipment in the target area through multiple frames of target images; wherein the target image includes the lighting equipment in the target area, and the target image is generated by capturing an image capturing device, and the image capturing device is disposed on a vehicle traveling in the target area;

[0028] The predicted value corresponding to the target feature vector is calculated by a lighting anomaly detection model generated in advance or in real time, and when the predicted value is less than or equal to a fourth threshold, abnormal lighting data of the lighting equipment in the target area is generated; wherein, the target feature vector is generated based on the ambient light intensity of the target area at the current moment, the average light intensity within the sliding time window length, the standard deviation of the light intensity within the sliding time window length, the light intensity change rate, the light intensity fluctuation state and the flickering frequency.

[0029] In one embodiment of the present application, the process of generating a lighting anomaly detection model in advance or in real time includes:

[0030] Performing dimensionality reduction and nonlinear feature extraction on the training feature vector using an autoencoder to obtain a reduced-dimensional feature vector; wherein the training feature vector is generated based on the ambient light intensity of the target area at a historical moment, the average light intensity within the sliding time window length, the standard deviation of the light intensity within the sliding time window length, the light intensity change rate, the light intensity fluctuation state, and the flicker frequency;

[0031] The dimension-reduced feature vector is used as a training sample, and a pre- or real-time established objective function and constraint conditions are added, and classification training is performed on the training sample to generate the lighting anomaly detection model.

[0032] In one embodiment of the present application, the process of determining the light intensity fluctuation state of the target area includes: determining the light intensity fluctuation state of the lighting device in the target area according to the ambient light intensity of the target area, the average light intensity within the sliding time window length, and the light intensity standard deviation within the sliding time window length, and using the light intensity fluctuation state of the lighting device in the target area as the light intensity fluctuation state of the target area;

[0033] Among them, when the difference between the ambient light intensity and the average light intensity within the sliding time window length is greater than or equal to the second threshold, the light intensity fluctuation state of the lighting device in the target area is determined to be abnormal; or, when the light intensity standard deviation within the sliding time window length is greater than or equal to the third threshold, the light intensity fluctuation state of the lighting device in the target area is determined to be abnormal.

[0034] In one embodiment of the present application, the process of encapsulating and encrypting the lighting data, generating a data packet, and wirelessly transmitting the data packet to a cloud server includes:

[0035] Obtaining metadata associated with abnormal lighting data of the lighting device in the target area, and forming an original data packet based on the abnormal lighting data and the metadata; the metadata includes: vehicle number, vehicle geographic location coordinate information, timestamp, abnormal type of the lighting device, abnormal state of the lighting device, and data hash value; wherein the geographic location coordinates of the original data packet include the vehicle geographic location coordinate information in the metadata;

[0036] The original data packet is encrypted to obtain an encrypted data packet, and the encrypted data packet is wirelessly transmitted to a cloud server.

[0037] In one embodiment of the present application, when the encrypted data packet is wirelessly transmitted to the cloud server, the method further includes:

[0038] Calculating signal strength during wireless transmission, and adjusting transmit power and modulation mode based on the signal strength and current channel conditions; and,

[0039] The forward error correction mechanism and automatic retransmission request mechanism are embedded in the wireless transmission process.

[0040] The present application also provides a data processing system, which is applied to a cloud server, and the system includes:

[0041] A data receiving module, used to receive a data packet transmitted wirelessly by a vehicle and decrypt the data packet; wherein the data packet is generated by encapsulating and encrypting traffic data in a traffic road area;

[0042] An abnormal area module, used for performing abnormal clustering of data points according to the decryption result of the data packet, and generating an abnormal area based on the abnormal clustering result; wherein the data point includes the geographical location coordinate information of the data packet;

[0043] The warning push module is used to push warning information to vehicles within a preset distance range of the abnormal area according to the vehicle's geographical location coordinate information.

[0044] The present application also provides a data processing system, applied to a vehicle, the system comprising:

[0045] A data packet generation module, used to obtain traffic data of a target area, and encapsulate and encrypt the traffic data to generate a data packet; wherein the target area includes a traffic road area determined in advance or in real time;

[0046] A wireless transmission module, used for wirelessly transmitting the data packet to a cloud server, so that the cloud server decrypts the data packet, performs abnormal clustering of data points based on the decryption result of the data packet, and generates an abnormal area based on the abnormal clustering result;

[0047] The warning receiving module is used to receive the warning information generated and pushed by the cloud server when the vehicle is within a preset distance range from the abnormal area.

[0048] The present application also provides a cloud server, which is used to execute the data processing method as described above, or to execute the data processing system as described above.

[0049] The present application also provides a vehicle, which is used to execute the data processing method as described above, or to execute the data processing system as described above.

[0050] As described above, the present application provides a data processing method and system, a cloud server and a vehicle, which have the following beneficial effects:

[0051] This application uses wireless transmission technology to realize the timely reporting and sharing of abnormal information, shortens the time difference from detection to response, enables relevant departments to intervene in the repair of traffic facility failures more quickly, improves the emergency response speed, effectively responds to emergencies, and reduces safety hazards. At the same time, this application uses a cloud server to perform abnormal analysis on traffic data, which can not only efficiently screen and verify abnormal traffic data, but also accurately locate the geographical location of abnormal traffic data, thereby providing decision-making support for traffic management departments, optimizing traffic management efficiency, and also helping to formulate more scientific and reasonable traffic maintenance plans to avoid resource waste. At the same time, this application pushes early warning information in a timely manner through a cloud server, allowing drivers to take risk avoidance measures in advance, such as slowing down, changing lanes, or planning alternative routes. This active protection mechanism improves the public's trust in the intelligent traffic management system and promotes social harmony. In addition, this application can detect abnormalities in a timely manner (such as erroneous flashing of traffic lights, inappropriate opening or closing of street lights, etc.) by real-time monitoring of traffic data in traffic road areas, thereby greatly reducing the risk of traffic accidents caused by traffic facility failures and protecting the safety of drivers, passengers and pedestrians. And timely repair of traffic facility failures can avoid unnecessary energy consumption, while reducing additional carbon emissions caused by traffic congestion, in line with energy conservation and emission reduction requirements. Therefore, this application forms a highly automated, real-time responsive traffic data warning and push system, covering the complete process from traffic data collection, analysis and prediction, decision-making to information dissemination. It can reduce traffic accidents, ease traffic congestion, and improve the overall operating efficiency of the traffic management system through intelligent means, which is of great value to promoting the development of intelligent traffic management systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A flowchart of a data processing method provided in an embodiment of the present application;

[0053] Figure 2 A flowchart of a data processing method provided in another embodiment of the present application;

[0054] Figure 3 A schematic diagram of the hardware structure of a data processing system provided in an embodiment of the present application;

[0055] Figure 4 A schematic diagram of the hardware structure of a data processing system provided in another embodiment of the present application;

[0056] Figure 5 A schematic diagram of the framework of a data processing system provided in one embodiment of the present application. DETAILED DESCRIPTION

[0057] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0058] It should be noted that the illustrations provided in this embodiment only illustrate the basic concept of the present application in a schematic manner, and the figures only show the components related to the present application rather than being drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0059] In an exemplary embodiment, if Figure 1 As shown, this embodiment provides a data processing method, which is applied to a vehicle, and the method includes the following steps:

[0060] S110, obtaining traffic data of a target area, and encapsulating and encrypting the traffic data to generate a data packet; wherein the target area includes a traffic road area determined in advance or in real time;

[0061] S120, wirelessly transmitting the data packet to the cloud server, so that the cloud server decrypts the data packet, performs abnormal clustering of data points based on the decryption result of the data packet, and generates an abnormal area based on the abnormal clustering result;

[0062] S130, when the vehicle is within a preset distance range from the abnormal area, receiving warning information generated and pushed by the cloud server; wherein the warning information includes: lighting data of the traffic road area and vehicle driving suggestions.

[0063] In an exemplary embodiment, if the traffic data includes lighting data, the process of acquiring the lighting data of the target area includes:

[0064] The light intensity change rate of the target area is calculated according to the ambient light intensity of the target area, the average light intensity within the sliding time window length, and the light intensity standard deviation within the sliding time window length; wherein the ambient light intensity is generated by a light intensity detector, and the light intensity detector is arranged on a vehicle traveling in the target area. In this embodiment or other embodiments, the specific values ​​of the second threshold value and the third threshold value can be limited according to actual conditions, and this embodiment does not make specific numerical limitations here.

[0065] The flicker frequency of the lighting equipment in the target area is calculated by using multiple frames of target images; wherein the target image includes the lighting equipment in the target area, and the target image is generated by an image capture device, which is set on a vehicle traveling in the target area; the predicted value corresponding to the target feature vector is calculated by a lighting anomaly detection model generated in advance or in real time, and when the predicted value is less than or equal to a fourth threshold, abnormal lighting data of the lighting equipment in the target area is generated; wherein the target feature vector is generated based on the ambient light intensity of the target area at the current moment, the average light intensity within the sliding time window length, the standard deviation of the light intensity within the sliding time window length, the light intensity change rate, the light intensity fluctuation state and the flicker frequency. In this embodiment or other embodiments, the specific value of the fourth threshold can be limited according to actual conditions, and this embodiment does not make a specific numerical limit here.

[0066] Specifically, according to the above records, the process of generating a lighting anomaly detection model in advance or in real time includes:

[0067] The training feature vector is reduced in dimension and nonlinearly extracted by using an autoencoder to obtain a reduced-dimensional feature vector; wherein the training feature vector is generated based on the ambient light intensity of the target area at the historical moment, the average light intensity within the sliding time window length, the standard deviation of the light intensity within the sliding time window length, the light intensity change rate, the light intensity fluctuation state and the flicker frequency;

[0068] The reduced-dimensional feature vectors are used as training samples, and the objective functions and constraints established in advance or in real time are added to perform classification training on the training samples to generate a lighting anomaly detection model.

[0069] According to the above records, the process of determining the light intensity fluctuation state of the target area includes: determining the light intensity fluctuation state of the lighting device in the target area according to the ambient light intensity of the target area, the average light intensity within the sliding time window length, and the standard deviation of the light intensity within the sliding time window length, and taking the light intensity fluctuation state of the lighting device in the target area as the light intensity fluctuation state of the target area; wherein, when the difference between the ambient light intensity and the average light intensity within the sliding time window length is greater than or equal to a second threshold value, the light intensity fluctuation state of the lighting device in the target area is determined to be abnormal; or, when the standard deviation of the light intensity within the sliding time window length is greater than or equal to a third threshold value, the light intensity fluctuation state of the lighting device in the target area is determined to be abnormal.

[0070] As an example, specifically, for the lighting data of the target area, let the ambient light intensity Over time Changes, introducing the average light intensity and light intensity standard deviation , to measure the average difference and fluctuation of the current light intensity compared with the past time period. The time period can be represented by the sliding time window length To express it, we have: , ; In the formula, Indicates time, Indicates the length of the sliding time window. If the ambient light intensity Significant deviation from the average light intensity , that is, the ambient light intensity The average light intensity within the sliding time window length The difference between the values ​​exceeds the second threshold; or the light intensity standard deviation If the third threshold is exceeded, it indicates that there is abnormal light intensity fluctuation in the lighting equipment in the target area.

[0071] Define the average brightness of the target lighting device pixels in each frame of the target image that contains the lighting device in the target area, by continuously Frame target image calculation flicker frequency ,have: ; ; In the formula, Indicates The average brightness of the target lighting device in the frame image, Indicates The average brightness of the target lighting device in the frame image. It means the Frame image The brightness of a pixel, Indicates the total number of pixels, the number of pixels in the coverage area of ​​the target lighting device in the image, Represents a pixel index.

[0072] Constructing feature vectors to enhance sensitivity to abnormal conditions includes: , ; In the formula, is the normalized rate of change of light intensity, represents the training feature vector.

[0073] Use Autoencoder (AE) to train feature vectors Perform dimensionality reduction and nonlinear feature extraction to obtain the reduced dimensionality feature vector ,have: ; In the formula, Represents the encoding and decoding process of the autoencoder AE, Indicates the parameters, Represents the self-encoder AE training feature vector encoding and decoding process.

[0074] The reduced-dimensional feature vector is used as a training sample, and the objective function and constraints established in advance or in real time are added. The support vector machine (SVM) is used to train the reduced-dimensional feature vector output by the hidden layer of the autoencoder AE. Classify and train to generate a lighting anomaly detection model. The objective function is: ; In the formula, Represents the weight vector. In the support vector machine SVM, is a vector that determines the direction of the decision boundary (also known as the maximum margin hyperplane). Each element of corresponds to the reduced dimensionality feature vector A feature in the corresponding feature space, its value indicates the importance of the feature in the classification decision. Represents the bias term. is a scalar that determines the decision boundary on the reduced feature vector The corresponding position in the feature space. Specifically, Controls the distance of the hyperplane from the origin. Indicates In the support vector machine (SVM), slack variables are introduced to allow some sample points to be on the wrong side or within the interval boundary. Each training sample has a corresponding slack variable, which indicates the degree to which the training sample violates the maximum interval constraint. The constraints are: ; In the formula, represents the index of the training sample, is a sample pair, (-1 means abnormal, 1 means normal), is the penalty coefficient, represents the Gaussian kernel function, Indicates Gaussian kernel function value of training samples; ; In the formula, Represents the width parameter, which controls the width or spread of the Gaussian kernel function. When it is larger, it means that the width of the Gaussian kernel function is narrower, the lighting anomaly detection model is more sensitive to the local features of the training samples, and can fit more complex decision boundaries. When it is smaller, the width of the Gaussian kernel function is wider, the lighting anomaly detection model pays more attention to the global structure, and the decision boundary is relatively smooth, which helps prevent overfitting. Represents the center point, which is the weighted average of all training samples.

[0075] Establish judgment criteria for the target feature vector After the dimensionality reduction by the autoencoder AE is input into the lighting anomaly detection model, the prediction value corresponding to the target feature vector is calculated by the lighting anomaly detection model: ; In the formula, represents the predicted value corresponding to the target feature vector, Represents the target feature vector of the autoencoder AE The encoding and decoding process, Represents the target feature vector The result at the end of the encoding and decoding process is used as the Gaussian kernel function value when the training sample is used. , it is determined that the lighting equipment in the target area is abnormal, and abnormal lighting data of the lighting equipment in the target area is generated.

[0076] It can be seen that this embodiment achieves efficient and accurate detection of abnormal conditions of lighting data in traffic road areas through multi-dimensional data analysis and advanced machine learning technology.

[0077] In an exemplary embodiment, the process of encapsulating and encrypting lighting data, generating a data packet, and wirelessly transmitting the data packet to a cloud server includes: obtaining metadata associated with abnormal lighting data of lighting devices in a target area, and forming an original data packet based on the abnormal lighting data and metadata; the metadata includes: vehicle number, vehicle geographic location coordinate information, timestamp, abnormal type of lighting device, abnormal state of lighting device, data hash value; encrypting the original data packet to obtain an encrypted data packet, and wirelessly transmitting the encrypted data packet to a cloud server. In this embodiment or other embodiments, the vehicle geographic location coordinate information in the metadata can constitute the geographic location coordinate information of the original data packet, that is, by encapsulating the vehicle geographic location coordinate information in the metadata, the geographic location coordinate information of the data packet can be obtained.

[0078] As an example, specifically, the detected abnormal lighting data and its related metadata are packaged together into a structured data unit for wireless transmission through a communication network. The purpose of doing so is also to ensure the integrity and structure of the data, so that the receiving end can correctly parse and understand the content of the sent data. If the metadata includes: vehicle number, vehicle geographic location coordinate information, timestamp, abnormal type of lighting equipment, abnormal state of lighting equipment, data hash value; then the packaged data packet of this embodiment Can contain the following fields: : Vehicle ID; : Vehicle GPS coordinates; : timestamp; : Abnormal data field; : Exception data type; : Abnormal data status; : Data hash value, used for integrity verification; then the encapsulated data packet It can be expressed as: . Then use the symmetric encryption algorithm Using Keys Data Packet Encrypt and get the encrypted data packet : . Then use V2X (Vehicle to everything, vehicle-to-everything wireless communication, referred to as V2X) to wirelessly transmit the encrypted data packets to the cloud server.

[0079] In an exemplary embodiment, when the encrypted data packet is wirelessly transmitted to the cloud server, it can also include: calculating the signal strength during wireless transmission, and adjusting the transmission power and modulation method based on the signal strength and current channel conditions; and embedding a forward error correction mechanism and an automatic retransmission request mechanism during the wireless transmission process.

[0080] As an example, specifically, using V2X wireless communication technology such as DSRC (Dedicated Short Range Communications, dedicated short range communication technology, referred to as DSRC), the encrypted data packet is sent to When sending to the cloud, the reliability of wireless transmission is affected by many factors, including signal strength, channel conditions, and interference. Therefore, in order to ensure the encrypted data packets To ensure successful transmission, the transmission power and modulation mode can be adjusted according to the current channel conditions and signal strength. The signal strength calculation process can be: , where Indicates the signal strength of the cloud server (i.e. the receiving end), Indicates the transmission power of the vehicle (i.e. the transmitter), represents the antenna gain, represents the wavelength, Indicates the distance between the vehicle (i.e., the transmitter) and the cloud server (i.e., the receiver).

[0081] In addition, in a wireless transmission environment, data packets may be erroneous due to channel interference. Therefore, in order to improve the reliability of data transmission, this embodiment may use a forward error correction FEC mechanism and an automatic repeat request ARQ mechanism. Among them, the forward error correction FEC mechanism can use error correction codes (such as Turbo codes) to add redundant information, so that the cloud server (i.e., the receiving end) can restore the original data even if some data is lost. Under the automatic repeat request ARQ mechanism, if the cloud server (i.e., the receiving end) detects a data packet error, it will request the vehicle (i.e., the transmitting end) to resend the data packet.

[0082] It can be seen that by encapsulating, encrypting, wirelessly transmitting, and detecting and correcting traffic data, efficient and secure communication between the vehicle and the cloud is achieved, thereby supporting the early warning function of the intelligent transportation system. This not only improves road safety, but also enhances the efficiency of traffic management.

[0083] In another exemplary embodiment, Figure 2 As shown, this embodiment provides a data processing method, which is applied to a cloud server, and the method includes the following steps:

[0084] S210, receiving a data packet transmitted wirelessly by the vehicle, and decrypting the data packet; wherein the data packet may be generated by the vehicle based on the traffic data encapsulation and encryption in the traffic road area. Specifically, the cloud server in this embodiment may receive the encrypted data packet transmitted wirelessly by the vehicle. , and then encrypt the data packet To decrypt, we have: ;in, Is the encryption algorithm The inverse operation of decryption can get the original data packet As an example, the traffic data in this embodiment may be composed of lighting data.

[0085] S220, performing abnormal clustering of data points based on the decryption result of the data packet, and generating abnormal areas based on the abnormal clustering result; wherein the data points are formed based on the geographic location coordinate information of the data packet.

[0086] S230, based on the geographic location coordinate information of the vehicle, push warning information to vehicles within a preset distance range of the abnormal area.

[0087] In an exemplary embodiment, the process of performing abnormal clustering of data points based on the decryption result of a data packet and generating an abnormal area based on the abnormal clustering result includes: performing integrity verification on the data packet based on the decryption result of the data packet; and, after the data packet passes the integrity verification, generating a data point based on the geographic location coordinate information of the data packet; selecting a data point as a target data point, and recording the minimum number of neighborhood data points of the target data point with the target data point as the center, and marking the target data point as a core data point when the minimum number of neighborhood data points is greater than or equal to a first threshold, and adding the corresponding data point to an abnormal cluster; wherein the minimum neighborhood data point includes a data point located within a preset neighborhood radius of the target data point; performing regional association according to all boundary data points in the abnormal clustering result to generate an abnormal area; wherein the boundary data point is located within the preset neighborhood radius of the core data point and is not a core data point. In this embodiment or other embodiments, the specific value of the first threshold can be limited according to actual conditions, and this embodiment does not make a specific numerical limit here.

[0088] As an example, specifically, in order to ensure the accuracy and validity of the data, the cloud server needs to perform data integrity verification on each data packet, that is, compare the previously calculated hash value with the hash value contained in the data packet. If the two hash values ​​match, the data is considered valid and the data packet is deemed to have passed the integrity verification; otherwise, the data is considered damaged and discarded, and the data packet is deemed to have failed the integrity verification. After the data packet passes the integrity verification, this embodiment can use the DBSCAN (Density-Based Spatial Clustering of Applications with Noise, referred to as DBSCAN) algorithm to measure the distance between data points to identify core data points and boundary data points, thereby performing abnormal clustering. That is, the DBSCAN algorithm can be used to identify the clustering area of ​​abnormal events using data packets uploaded from the vehicle to the cloud. Each data packet contains information such as vehicle ID, vehicle GPS coordinates, timestamp, abnormal data field, abnormal data type, and abnormal data status. When clustering using the DBSCAN algorithm, the vehicle GPS coordinate information in these data packets is used as the spatial position of the data point, that is, the data point. Therefore, the relationship between the data point and the data packet is as follows: In the DBSCAN algorithm, the data point is the basic unit of the DBSCAN algorithm operation. In this embodiment, the vehicle GPS coordinate information in each data packet is converted into a point in a two-dimensional space, and this point becomes a data point in the DBSCAN algorithm. The data packet contains a series of metadata about anomaly detection, the most important of which is the vehicle GPS coordinates, which is the basis for determining the spatial position of the data point. In addition, the data packet also includes information such as vehicle ID, timestamp, abnormal data field, abnormal data type and abnormal data status. Although this information is not directly used for clustering, it is very important in subsequent analysis and processing. For example, it can be used for type classification of abnormal events, time series analysis, and determining the source vehicle of abnormal events. Therefore, when the data packet is decrypted and verified to be valid, the vehicle GPS coordinate information therein is extracted as the input of the DBSCAN algorithm. The DBSCAN algorithm identifies core data points, boundary data points and noise data points by measuring the distance between these data points. Specifically, the preset neighborhood radius is recorded as , the minimum number of neighborhood points is recorded as , preset neighborhood radius and the minimum number of neighboring points Determines how data points are classified as core data points, boundary data points, or noise data points. If a data point presets the neighborhood radius At least data points, then this point is marked as a core data point. If a data point itself is not a core data point, but it is in the neighborhood of a core data point, then this point is marked as a boundary data point. If a data point is neither a core data point nor a boundary data point, then it is marked as a noise data point. It can be seen that this embodiment can identify the clustering area of ​​abnormal events, that is, the dense area of ​​abnormal data points, through the DBSCAN algorithm, thereby filtering out isolated false alarms and improving the accuracy and efficiency of anomaly detection. Among them, the DBSCAN algorithm is a density-based spatial clustering algorithm, which divides areas with sufficiently high density into clusters, and can find clusters of arbitrary shapes in noisy data sets.

[0089] According to the above records, the DBSCAN algorithm can identify areas with dense data points, but these areas may correspond to real abnormal events, while isolated data points may be false alarms. Therefore, in an exemplary embodiment, the embodiment can use a pre- or real-time data point abnormality detection lighting abnormality detection model to calculate the path length of each data point in the abnormal area, and calculate the abnormality score of each data point based on the path length of each data point, and when the abnormality score is greater than a preset score threshold, mark the corresponding data point as a false alarm data point; and remove the associated area corresponding to the false alarm data point from the abnormal area to obtain the real abnormal area. In this embodiment or other embodiments, the specific value of the preset score threshold can be limited according to the actual situation, and this embodiment does not make a specific numerical limit here.

[0090] As an example, specifically, abnormal confirmation and elimination of false alarms are one of the core functions based on cloud servers, which aims to accurately distinguish real problem events from false alarms from the abnormal data reported by massive vehicles, and ensure the efficiency and accuracy of the early warning system. Therefore, in this embodiment, the abnormal confirmation stage can use the IF (Isolation Forest, IF for short) algorithm, which is an unsupervised learning method based on a tree structure, specially designed for anomaly detection. The IF algorithm isolates abnormal points by constructing multiple trees and randomly selecting features and data subsets. The higher the anomaly score, the more abnormal the data point. The core advantage of the IF algorithm lies in its efficiency and sensitivity to abnormal data, which is particularly suitable for processing large-scale high-dimensional data sets. The IF algorithm is used to construct a data point anomaly detection lighting anomaly detection model based on historical data, and then the data point anomaly detection lighting anomaly detection model determines whether the data point is abnormal by calculating the "isolation" of the data point, that is, the path length of the data point in the constructed data point anomaly detection lighting anomaly detection model. The longer the path length, the more isolated the data point is, and the more likely it is an abnormal point. The anomaly score of each data point is calculated based on its path length in the data point anomaly detection and lighting anomaly detection model; the higher the score, the more abnormal the data point. If the anomaly score exceeds the preset score threshold, it is confirmed as an anomaly; otherwise, it is considered a false alarm and removed. The selection of the preset score threshold can be based on the distribution of the anomaly score, and the high percentile of the anomaly score distribution is usually selected as the threshold. For example, in this embodiment, the score of 95% can be used as the upper limit of the anomaly point. In this embodiment, false alarm removal is to ensure that the data point anomaly detection and lighting anomaly detection model can effectively identify and remove false alarms, reduce false positive results, and improve the accuracy of anomaly detection by setting reasonable preset score thresholds and parameters. The following is a strategy for false alarm removal: First, threshold optimization is performed, and the threshold of the anomaly score is dynamically adjusted by analyzing historical data to adapt to the dynamic changes of the data. This can be achieved by monitoring the distribution of the anomaly score and regularly updating the threshold. Secondly, parameter fine-tuning: optimize the parameters of the data point anomaly detection and lighting anomaly detection model, such as the number of trees, the maximum depth of the tree, and the sampling ratio, to improve the detection performance of the lighting anomaly detection model and reduce false alarms. Finally, through real-time feedback and learning, feedback information from the anomaly detection process is collected, and the parameters and thresholds of the lighting anomaly detection model are continuously adjusted to adapt to environmental changes and reduce false alarms. In addition, for confirmed abnormal events, the cloud server can accurately locate the location based on the received vehicle GPS coordinates and use the positioning results to push warning information. At the same time, based on historical data analysis and the current traffic flow lighting anomaly detection model, the affected area is evaluated to ensure the effective communication of warning information.

[0091] It can be seen that this embodiment can achieve efficient processing, authenticity verification and precise location determination of abnormal reports by utilizing statistics, data mining and machine learning methods, which is an important part of improving the safety and management efficiency of intelligent transportation systems.

[0092] In an exemplary embodiment, the process of pushing warning information to vehicles within a preset distance range of an abnormal area based on the vehicle's geographic location coordinate information includes: generating a predicted driving trajectory of the current vehicle based on the current vehicle's historical geographic location coordinate information and real-time geographic location coordinate information; when the predicted driving trajectory intersects with the abnormal area, determining the center of the abnormal area, the radius or the side length of the abnormal area according to the geometry of the abnormal area; calculating the distance between the current vehicle and the center of the abnormal area based on the real-time geographic location coordinate information of the current vehicle, and when the distance value is less than or equal to the radius or the side length of the abnormal area, marking the current vehicle as a warning vehicle, generating warning information, and pushing the warning information to the warning vehicle; wherein the warning information includes: lighting data of the traffic road area, and vehicle driving suggestions.

[0093] As an example, specifically, this embodiment can predict which vehicles will enter the confirmed abnormal area by analyzing the historical trajectory data and real-time GPS information of the vehicle, and the vehicle trajectory prediction lighting anomaly detection model can be generated using the Kalman filter lighting anomaly detection model. Assuming the vehicle position In time The prediction of follows a linear dynamic system, then: ; In the formula, is the state transition matrix, is the control input matrix, are control vectors (such as speed and direction), is process noise, which obeys zero-mean Gaussian distribution. The update equation for predicting vehicle position is: ; In the formula, is the optimal estimate at time (t), It's time Time-based Information prediction.

[0094] Assume that the abnormal area can be represented by a geometric shape (such as a circle or rectangle) with its center at , the radius or side length is If the predicted trajectory If the vehicle intersects with the abnormal area, the vehicle should be marked as a warning vehicle. For a circular abnormal area, if the distance between the vehicle position and the center is less than or equal to the radius, the vehicle is located in the abnormal area, and: .

[0095] Generate warning information after determining the warning vehicle and transmit warning information via V2X The warning information can be pushed to the warning vehicle. The warning information can include the abnormal type of traffic road area, the abnormal location of the traffic road area, the timestamp and the vehicle driving advice. The vehicle driving suggestions contained in the warning information can be customized according to the type of anomaly and the current location of the vehicle. For example, if the abnormal data is a traffic light failure, the vehicle driving suggestions may include slowing down, preparing to stop, or finding an alternative route. The transmission rate (Unit: bit / second), data size (Unit: bit), then the transmission time for: Therefore, in order to ensure the timeliness of warning information, the warning information is transmitted through V2X. When sending the warning message to the warning vehicle, the transmission time must be ensured. Short enough to meet real-time requirements.

[0096] It can be seen that this embodiment uses historical trajectory data and real-time GPS information to accurately predict the future position of the vehicle through the Kalman filter algorithm. This prediction method takes into account the dynamically changing driving state and external interference factors (such as process noise), and improves the prediction accuracy; it realizes efficient modeling of vehicle movement patterns, and can accurately predict whether the vehicle will approach or enter the abnormal area under complex road conditions, providing key data support for early warning. In addition, this embodiment defines the abnormal area (such as a circle or rectangle) by geometric methods, and develops an efficient judgment algorithm to identify the intersection of the predicted trajectory and the abnormal area. It not only simplifies the complexity of spatial analysis and quickly locates potential threatened vehicles; it also improves the efficiency and accuracy in identifying potential risk vehicles, ensuring the pertinence and effectiveness of the warning. At the same time, this embodiment can also automatically generate warning information containing abnormal type, location, timestamp and personalized driving suggestions, customize the suggestion content according to the abnormal situation and the actual condition of the vehicle, and enhance the practicality and pertinence of the warning. It promotes drivers to quickly understand and respond to warnings, take effective measures to avoid accidents, and improve the road safety factor. Therefore, this embodiment predicts vehicle trajectories, generates warning information, and pushes it in real time through the V2X network, thereby ensuring a rapid response to abnormal events in the intelligent transportation system and improving road safety and traffic efficiency. At the same time, this embodiment can effectively notify the driver of potential dangers and assist the driver in taking appropriate driving actions by combining the vehicle trajectory prediction lighting anomaly detection model and V2X communication technology.

[0097] In summary, the present application provides a data processing method, which realizes the timely reporting and sharing of abnormal information by adopting wireless transmission technology, shortens the time difference from detection to response, enables relevant departments to intervene in the repair of traffic facility failures more quickly, improves the emergency response speed, effectively responds to emergencies, and reduces safety hazards. At the same time, the data processing method in the present application performs abnormal analysis on traffic data through a cloud server, which can not only efficiently screen and verify abnormal traffic data, but also accurately locate the geographical location of abnormal traffic data, thereby providing decision support for traffic management departments, optimizing traffic management efficiency, and also helping to formulate more scientific and reasonable traffic maintenance plans to avoid resource waste. At the same time, the data processing method in the present application timely pushes warning information through a cloud server, allowing drivers to take risk avoidance measures in advance, such as slowing down, changing lanes, or planning alternative routes. This active protection mechanism improves the public's trust in the intelligent traffic management system and promotes social harmony. In addition, the data processing method in the present application can detect abnormalities in a timely manner (such as erroneous flashing of traffic lights, inappropriate opening or closing of street lights, etc.) by real-time monitoring of traffic data in traffic road areas, thereby greatly reducing the risk of traffic accidents caused by traffic facility failures and protecting the safety of drivers, passengers and pedestrians. And timely repair of traffic facility failures can avoid unnecessary energy consumption, while reducing additional carbon emissions caused by traffic congestion, meeting the requirements of energy conservation and emission reduction. Therefore, the data processing method in this application forms a highly automated, real-time responsive traffic data warning and push system, covering the complete process from traffic data collection, analysis and prediction, decision-making to information dissemination. It can reduce traffic accidents, ease traffic congestion, and improve the overall operating efficiency of the traffic management system through intelligent means, which is of great value to promoting the development of intelligent traffic management systems.

[0098] In another exemplary embodiment of the present application, Figure 3 As shown, this embodiment also provides a data processing system, which is applied to a cloud server and includes:

[0099] The data receiving module 310 is used to receive a data packet transmitted wirelessly by the vehicle and decrypt the data packet; wherein the data packet may be generated by the vehicle based on the traffic data encapsulation and encryption in the traffic road area;

[0100] The abnormal area module 320 is used to perform abnormal clustering of data points according to the decryption results of the data packets, and to generate abnormal areas based on the abnormal clustering results; wherein the data points are formed based on the geographical location coordinate information of the data packets;

[0101] The warning push module 330 is used to push warning information to vehicles within a preset distance range of the abnormal area according to the geographic location coordinate information of the vehicle.

[0102] It should be noted that the data processing system provided in this embodiment and the data processing method applied to the cloud server provided in the above embodiment belong to the same concept, wherein the specific manner in which the data processing method applied to the cloud server performs operations has been described in detail in the above method embodiment, and will not be repeated here. In actual applications, the data processing system provided in this embodiment can allocate the above functions to different functional modules as needed, that is, divide the internal structure of the data processing system into different functional modules, and then implement all or part of the functions of the corresponding functional modules through the data processing method applied to the cloud server described in the above embodiment, and this is not specifically limited here.

[0103] In another exemplary embodiment of the present application, Figure 4 As shown, this embodiment also provides a data processing system, applied to a vehicle, comprising:

[0104] The data packet generation module 410 is used to obtain traffic data of a target area, and encapsulate and encrypt the traffic data to generate a data packet; wherein the target area includes a traffic road area determined in advance or in real time;

[0105] The wireless transmission module 420 is used to wirelessly transmit the data packet to the cloud server so that the cloud server decrypts the data packet, performs abnormal clustering of data points based on the decryption result of the data packet, and generates an abnormal area based on the abnormal clustering result;

[0106] The warning receiving module 430 is used to receive the warning information generated and pushed by the cloud server when the vehicle is within a preset distance range from the abnormal area; wherein the warning information includes: lighting data of the traffic road area and vehicle driving suggestions.

[0107] It should be noted that the data processing system provided in this embodiment and the data processing method applied to a vehicle provided in the above embodiment belong to the same concept, wherein the specific manner in which the data processing method applied to a vehicle performs operations has been described in detail in the above method embodiment, and will not be repeated here. In actual applications, the data processing system provided in this embodiment can distribute the above functions to different functional modules as needed, that is, divide the internal structure of the data processing system into different functional modules, and then implement all or part of the functions of the corresponding functional modules through the data processing method applied to a vehicle described in the above embodiment, and this is not specifically limited here.

[0108] In another exemplary embodiment of the present application, Figure 5 As shown, this embodiment also provides a data processing system, including: a vehicle-mounted abnormality detection module, a V2X communication module, a cloud data analysis and processing module, and an early warning push module.

[0109] Among them, the vehicle-mounted anomaly detection module can perform the following processes:

[0110] 1) Ambient light intensity analysis.

[0111] For the illumination data of the target area, set the ambient light intensity Over time Changes, introducing the average light intensity and light intensity standard deviation , to measure the average difference and fluctuation of the current light intensity compared with the past time period. The time period can be represented by the sliding time window length To express it, we have: , ; In the formula, Indicates time, Indicates the length of the sliding time window. If the ambient light intensity Significant deviation from the average light intensity , that is, the ambient light intensity The average light intensity within the sliding time window length The difference between the values ​​exceeds the second threshold; or the light intensity standard deviation If the third threshold is exceeded, it indicates that there is abnormal light intensity fluctuation in the lighting equipment in the target area.

[0112] 2) Image recognition and flicker frequency analysis.

[0113] The average brightness of the target lighting device pixels in each frame of the target image containing the lighting device in the target area is defined as Frame target image calculation flicker frequency ,have: ; ; In the formula, Indicates The average brightness of the target lighting device in the frame image, Indicates The average brightness of the target lighting device in the frame image. It means the Frame image The brightness of a pixel, Indicates the total number of pixels, the number of pixels in the coverage area of ​​the target lighting device in the image, Represents a pixel index.

[0114] 3) Intelligent judgment logic algorithm.

[0115] Constructing feature vectors to enhance sensitivity to abnormal conditions includes: , ; In the formula, is the normalized rate of change of light intensity, represents the training feature vector.

[0116] Use Autoencoder (AE) to train feature vectors Perform dimensionality reduction and nonlinear feature extraction to obtain the reduced dimensionality feature vector ,have: ; In the formula, Represents the encoding and decoding process of the autoencoder AE, Indicates the parameters, Represents the self-encoder AE training feature vector encoding and decoding process.

[0117] The reduced-dimensional feature vector is used as a training sample, and the objective function and constraints established in advance or in real time are added. The support vector machine (SVM) is used to train the reduced-dimensional feature vector output by the hidden layer of the autoencoder AE. Classify and train to generate a lighting anomaly detection model. The objective function is: ; In the formula, Represents the weight vector. In the support vector machine SVM, is a vector that determines the direction of the decision boundary (also known as the maximum margin hyperplane). Each element of corresponds to the reduced dimensionality feature vector A feature in the feature space, its value indicates the importance of the feature in the classification decision. Represents the bias term. is a scalar that determines the decision boundary on the reduced feature vector The corresponding position in the feature space. Specifically, Controls the distance of the hyperplane from the origin. Indicates In the support vector machine (SVM), slack variables are introduced to allow some sample points to be on the wrong side or within the interval boundary. Each training sample has a corresponding slack variable, which indicates the degree to which the training sample violates the maximum interval constraint. The constraints are: ; In the formula, represents the index of the training sample, is a sample pair, (-1 means abnormal, 1 means normal), is the penalty coefficient, represents the Gaussian kernel function, Indicates Gaussian kernel function value of training samples; ; In the formula, Represents the width parameter, which controls the width or spread of the Gaussian kernel function. When it is larger, it means that the width of the Gaussian kernel function is narrower, the lighting anomaly detection model is more sensitive to the local features of the training samples, and can fit more complex decision boundaries. When it is smaller, the width of the Gaussian kernel function is wider, the lighting anomaly detection model pays more attention to the global structure, and the decision boundary is relatively smooth, which helps prevent overfitting. Represents the center point, which is the weighted average of all training samples.

[0118] Establish judgment criteria for the target feature vector After the dimensionality reduction by the autoencoder AE is input into the lighting anomaly detection model, the prediction value corresponding to the target feature vector is calculated by the lighting anomaly detection model: ; In the formula, represents the predicted value corresponding to the target feature vector, Represents the target feature vector of the autoencoder AE The encoding and decoding process, Represents the target feature vector The result at the end of the encoding and decoding process is used as the Gaussian kernel function value when the training sample is used. , it is determined that the lighting equipment in the target area is abnormal, and abnormal lighting data of the lighting equipment in the target area is generated.

[0119] It can be seen that the on-board anomaly detection module can achieve efficient and accurate detection of abnormal conditions of lighting data in traffic road areas through multi-dimensional data analysis and advanced machine learning technology.

[0120] The V2X communication module can perform the following processes:

[0121] 1) Data encapsulation and encryption.

[0122] The detected abnormal lighting data and its related metadata are packaged together into a structured data unit for wireless transmission over a communication network. This can also ensure the integrity and structure of the data so that the receiving end can correctly parse and understand the content of the sent data. If the metadata includes: vehicle number, vehicle geographic location coordinate information, timestamp, abnormal type of lighting equipment, abnormal state of lighting equipment, data hash value; then the packaged data packet in this embodiment Can contain the following fields: : Vehicle ID; : Vehicle GPS coordinates; : timestamp; : Abnormal data field; : Exception data type; : Abnormal data status; : Data hash value, used for integrity verification; then the encapsulated data packet It can be expressed as: . Then use the symmetric encryption algorithm Using Keys Data Packet Encrypt and get the encrypted data packet : . The encrypted data packet is then wirelessly transmitted to the cloud server using V2X (Vehicle to everything, vehicle wireless communication, referred to as V2X). In this embodiment or other embodiments, the vehicle geographic location coordinate information in the metadata can constitute the geographic location coordinate information of the original data packet, that is, the geographic location coordinate information of the data packet can be obtained by encapsulating the vehicle geographic location coordinate information in the metadata.

[0123] 2) Wireless transmission.

[0124] To ensure that the encrypted data packets To ensure successful transmission, the transmission power and modulation mode can be adjusted according to the current channel conditions and signal strength. The calculation process of signal strength can be: , where Indicates the signal strength of the cloud server (i.e. the receiving end), Indicates the transmission power of the vehicle (i.e. the transmitter), represents the antenna gain, represents the wavelength, Indicates the distance between the vehicle (i.e., the transmitter) and the cloud server (i.e., the receiver).

[0125] 3) Error detection and correction.

[0126] In order to improve the reliability of data transmission, the forward error correction FEC mechanism and the automatic repeat request ARQ mechanism can be used. Among them, the forward error correction FEC mechanism can use error correction codes (such as Turbo codes) to add redundant information, so that the cloud server (i.e., the receiving end) can restore the original data even if some data is lost. Under the automatic repeat request ARQ mechanism, if the cloud server (i.e., the receiving end) detects a data packet error, it will ask the vehicle (i.e., the transmitting end) to resend the data packet.

[0127] It can be seen that the V2X communication module realizes efficient and secure communication between the vehicle and the cloud by encapsulating, encrypting, wirelessly transmitting, and detecting and correcting traffic data, thereby supporting the early warning function of the intelligent transportation system. This not only improves road safety, but also enhances the efficiency of traffic management.

[0128] The cloud data analysis and processing module can perform the following processes:

[0129] 1) Data preprocessing.

[0130] Receive encrypted data packets transmitted wirelessly from the vehicle , and then encrypt the data packet To decrypt, we have: ;in, Is the encryption algorithm The inverse operation of decryption can get the original data packet .

[0131] 2) Abnormal data verification.

[0132] To ensure the accuracy and validity of the data, the cloud data analysis and processing module needs to perform a data integrity check on each data packet, that is, compare the previously calculated hash value with the hash value contained in the data packet. If the two hash values ​​match, the data is considered valid and the data packet is deemed to have passed the integrity verification; otherwise, the data is considered damaged and discarded, and the data packet is deemed to have failed the integrity verification.

[0133] 3) Anomaly clustering and hotspot detection.

[0134] Using the DBSCAN algorithm, the distance between data points is measured to identify core data points and boundary data points, thereby performing abnormal clustering. That is, the DBSCAN algorithm can use the data packets uploaded from the vehicle to the cloud to identify the clustering area of ​​abnormal events. Each data packet contains information such as vehicle ID, vehicle GPS coordinates, timestamp, abnormal data field, abnormal data type and abnormal data status. When clustering using the DBSCAN algorithm, the vehicle GPS coordinate information in these data packets is used as the spatial position of the data point, that is, the data point. Therefore, the relationship between the data point and the data packet is as follows: In the DBSCAN algorithm, the data point is the basic unit of the DBSCAN algorithm operation. In this embodiment, the vehicle GPS coordinate information in each data packet is converted into a point in a two-dimensional space, and this point becomes a data point in the DBSCAN algorithm. The data packet contains a series of metadata about abnormal detection, the most important of which is the vehicle GPS coordinates, which is the basis for determining the spatial position of the data point. In addition, the data packet also includes information such as vehicle ID, timestamp, abnormal data field, abnormal data type, and abnormal data status. Although this information is not directly used for clustering, it is very important in subsequent analysis and processing. For example, it can be used for type classification of abnormal events, time series analysis, and determining the source vehicle of abnormal events. Therefore, when the data packet is decrypted and verified to be valid, the vehicle GPS coordinate information is extracted as the input of the DBSCAN algorithm. The DBSCAN algorithm identifies core data points, boundary data points, and noise data points by measuring the distance between these data points. Specifically, the preset neighborhood radius is recorded as , the minimum number of neighborhood points is recorded as , preset neighborhood radius and the minimum number of neighboring points Determines how data points are classified as core data points, boundary data points, or noise data points. If a data point presets the neighborhood radius At least If a data point is not a core data point, but is in the neighborhood of a core data point, then this point is marked as a boundary data point. If a data point is neither a core data point nor a boundary data point, then it is marked as a noise data point.

[0135] 4) Confirm abnormalities and eliminate false positives.

[0136] The anomaly confirmation stage can use the Isolation Forest (IF) algorithm, which is a tree-based unsupervised learning method specifically designed for anomaly detection. The IF algorithm isolates anomalies by building multiple trees and randomly selecting features and data subsets. The higher the anomaly score, the more abnormal the data point. The core advantage of the IF algorithm lies in its efficiency and sensitivity to abnormal data, which is particularly suitable for processing large-scale high-dimensional data sets. The IF algorithm is used to build a data point anomaly detection and lighting anomaly detection model based on historical data. The data point anomaly detection and lighting anomaly detection model then determines whether the data point is abnormal by calculating the "isolation" of the data point, that is, the path length of the data point in the constructed data point anomaly detection and lighting anomaly detection model. The longer the path length, the more isolated the data point is and the more likely it is to be an anomaly. The anomaly score of each data point is calculated based on its path length in the data point anomaly detection and lighting anomaly detection model; the higher the score, the more abnormal the data point. If the anomaly score exceeds the preset score threshold, it is confirmed as an anomaly; otherwise, it is considered a false alarm and is eliminated. The selection of the preset score threshold can be based on the distribution of the anomaly score, and the high percentile of the anomaly score distribution is usually selected as the threshold. For example, in this embodiment, a score of 95% can be used as the upper limit of the abnormal point. In this embodiment, false alarm elimination is achieved by setting reasonable preset score thresholds and parameters to ensure that the data point abnormality detection lighting anomaly detection model can effectively identify and eliminate false alarms, reduce false positive results, and improve the accuracy of anomaly detection. The following is a strategy for false alarm elimination: First, threshold optimization is performed. By analyzing historical data, the threshold of the abnormality score is dynamically adjusted to adapt to the dynamic changes of the data. This can be achieved by monitoring the distribution of abnormality scores and regularly updating the threshold. Secondly, parameter fine-tuning: optimize the parameters of the data point abnormality detection lighting anomaly detection model, such as the number of trees, the maximum depth of the tree, and the sampling ratio, to improve the detection performance of the lighting anomaly detection model and reduce false alarms. Finally, feedback information of the anomaly detection process is collected through real-time feedback and learning, and the parameters and thresholds of the lighting anomaly detection model are continuously adjusted to adapt to changes in the environment and reduce false alarms.

[0137] 5) Abnormal event location and notification.

[0138] For confirmed abnormal events, the cloud data analysis and processing module can accurately locate the location based on the received vehicle GPS coordinates and use the positioning results to push warning information. At the same time, based on historical data analysis and the current traffic flow lighting anomaly detection model, the affected area is evaluated to ensure the effective communication of warning information.

[0139] The warning push module can execute the following processes:

[0140] 1) Vehicle trajectory prediction.

[0141] By analyzing the historical trajectory data and real-time GPS information of the vehicle to predict which vehicles will enter the confirmed abnormal area, the vehicle trajectory prediction lighting anomaly detection model can be generated using the Kalman filter lighting anomaly detection model. Assuming the vehicle position In time The prediction of follows a linear dynamic system, then: ; In the formula, is the state transition matrix, is the control input matrix, are control vectors (such as speed and direction), is process noise, which obeys zero-mean Gaussian distribution. The update equation for predicting vehicle position is: ; In the formula, is the optimal estimate at time (t), It's time Time-based Information prediction.

[0142] 2) Determination of abnormal areas.

[0143] Assume that the abnormal area can be represented by a geometric shape (such as a circle or rectangle) with its center at , the radius or side length is If the predicted trajectory If the vehicle intersects with the abnormal area, the vehicle should be marked as a warning vehicle. For a circular abnormal area, if the distance between the vehicle position and the center is less than or equal to the radius, the vehicle is located in the abnormal area, and: .

[0144] 3) Generation of early warning information.

[0145] Generate warning information after determining the warning vehicle ,The warning information can include the abnormal type of the traffic road area, the abnormal location of the traffic road area, the timestamp, and vehicle driving suggestions.

[0146] 4) V2X warning information push.

[0147] Use V2X to transmit warning information Push to the warning vehicle. Assume that the warning information The transmission rate (Unit: bit / second), data size (Unit: bit), then the transmission time for: Therefore, in order to ensure the timeliness of the warning information, the warning information is transmitted through V2X. When sending the warning message to the warning vehicle, the transmission time must be ensured. Short enough to meet real-time requirements.

[0148] 5) Safe driving advice.

[0149] Warning information The vehicle driving suggestions contained in the data can be customized according to the type of anomaly and the current location of the vehicle. For example, if the abnormal data is a traffic light failure, the vehicle driving suggestions may include slowing down, preparing to stop, or finding an alternative route.

[0150] It can be seen that this embodiment achieves efficient detection, confirmation and warning of abnormal conditions of road traffic sign facilities through the close cooperation of the vehicle-mounted abnormality detection module, V2X communication module, cloud data analysis and processing module and warning push module. This embodiment includes ambient light intensity analysis, image recognition and flicker frequency analysis, data verification and clustering, time series analysis, Isolation Forest abnormality detection, Kalman filter prediction lighting abnormality detection model, etc., which together constitute a highly intelligent abnormality detection and warning system. Through this abnormality detection and warning system, not only can abnormal situations be discovered and handled in a timely manner, but also false alarms can be effectively reduced, the level of road safety can be improved, and it constitutes an important part of modern intelligent traffic management.

[0151] In another exemplary embodiment of the present application, the embodiment further provides a cloud server, which is used to execute the data processing method applied to the cloud server as described in the above embodiment, or to execute the data processing system applied to the cloud server as described in the above embodiment. It should be noted that since the data processing method applied to the cloud server and the specific manner in which the data processing system performs operations have been described in detail in the embodiments, the technical functions and effects of the vehicle provided in this embodiment can be referred to the above embodiments, and will not be repeated here.

[0152] In another exemplary embodiment of the present application, the embodiment further provides a vehicle, which is used to execute the data processing method applied to the vehicle as described in the above embodiment, or to execute the data processing system applied to the vehicle as described in the above embodiment. It should be noted that since the specific manner in which the data processing method applied to the vehicle and the data processing system perform operations have been described in detail in the embodiments, the technical functions and effects of the vehicle provided in this embodiment can be referred to the above embodiments, and will not be repeated here.

[0153] It should be noted that the above embodiments collect, store, use, process, transmit, provide, disclose, delete, etc. related data (such as traffic data, etc.) with or after obtaining the user's consent. For example, traffic data is authorized with the user's knowledge and consent; or it is actively provided by the user after reading the relevant instructions, or it is actively authorized / provided / uploaded by the user when using some or all of the functions described in the above embodiments, or it is obtained by other means / channels with or after obtaining the user's consent.

[0154] It should be understood that, although the terms first, second, third, etc. may be used to describe thresholds, etc. in the embodiments of the present application, these terms are only used to distinguish thresholds from each other. For example, without departing from the scope of the embodiments of the present application, the first threshold may also be referred to as the second threshold, and similarly, the second threshold may also be referred to as the first threshold.

[0155] The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. A data processing method, characterized in that: Applied to a cloud server, the method comprises the following steps: Receiving a data packet wirelessly transmitted by the vehicle and decrypting the data packet; wherein the data packet is generated by encapsulating and encrypting traffic data in a traffic road area, and the traffic data includes lighting data; Perform abnormal clustering of data points based on the decryption result of the data packet, and generate abnormal areas based on the abnormal clustering result; including: selecting a data point as a target data point, wherein the data point includes the geographic location coordinate information of the data packet; taking the target data point as the center, recording the minimum number of neighborhood data points of the target data point, and when the minimum number of neighborhood data points is greater than or equal to a first threshold, marking the target data point as a core data point, and adding the corresponding data point to an abnormal cluster; wherein the minimum neighborhood data point includes a data point located within a preset neighborhood radius of the target data point; performing regional association according to all boundary data points in the abnormal clustering result to generate the abnormal area; wherein the boundary data point is located within the preset neighborhood radius of the core data point and is not a core data point; Based on the geographic location coordinate information of the vehicle, early warning information is pushed to vehicles within a preset distance range of the abnormal area.

2. The data processing method according to claim 1, characterized in that: The process of performing data point anomaly clustering based on the data packet decryption result and generating anomaly regions based on the anomaly clustering result also includes: Performing integrity verification on the data packet based on the data packet decryption result; and, After the data packet passes integrity verification, the data point is generated based on the geographic location coordinate information of the data packet.

3. The data processing method according to claim 2, characterized in that: After generating the abnormal area, the method further includes: Calculate the path length of each data point in the abnormal area using a data point anomaly detection lighting anomaly detection model established in advance or in real time; The anomaly score of each data point is calculated based on the path length of each data point, and when the anomaly score is greater than a preset score threshold, the corresponding data point is marked as a false alarm data point; and the associated area corresponding to the false alarm data point is removed from the abnormal area to obtain the real abnormal area.

4. The data processing method according to any one of claims 1 to 3, characterized in that: The process of pushing warning information to vehicles within a preset distance range of the abnormal area based on the geographic location coordinate information of the vehicle includes: Generate a predicted driving trajectory of the current vehicle based on the historical geographic location coordinate information and the real-time geographic location coordinate information of the current vehicle; When the predicted driving trajectory intersects the abnormal area, determining the center of the abnormal area, the radius or the side length of the abnormal area according to the geometric shape of the abnormal area; Based on the real-time geographic location coordinate information of the current vehicle, the distance value between the current vehicle and the center of the abnormal area is calculated, and when the distance value is less than or equal to the radius or side length of the abnormal area, the current vehicle is marked as a warning vehicle, and warning information is generated and pushed to the warning vehicle; wherein the warning information includes: lighting data of the traffic road area and vehicle driving suggestions.

5. A data processing method, characterized in that: Applied to a vehicle, the method comprises the following steps: Acquire traffic data of a target area, and encapsulate and encrypt the traffic data to generate a data packet; wherein the target area includes a traffic road area determined in advance or in real time, and the traffic data includes lighting data; The data packet is wirelessly transmitted to a cloud server so that the cloud server decrypts the data packet, performs abnormal data point clustering based on the data packet decryption result, and generates an abnormal area based on the abnormal clustering result; the process of the cloud server generating the abnormal area includes: selecting a data point as a target data point, wherein the data point includes the geographic location coordinate information of the data packet; taking the target data point as the center, recording the minimum number of neighborhood data points of the target data point, and when the minimum number of neighborhood data points is greater than or equal to a first threshold, marking the target data point as a core data point, and adding the corresponding data point to an abnormal cluster; wherein the minimum neighborhood data point includes a data point located within a preset neighborhood radius of the target data point; performing regional association according to all boundary data points in the abnormal clustering result to generate the abnormal area; wherein the boundary data point is located within the preset neighborhood radius of the core data point and is not a core data point; When the vehicle is within a preset distance range from the abnormal area, the warning information generated and pushed by the cloud server is received.

6. The data processing method according to claim 5, characterized in that: If the traffic data includes lighting data, the process of obtaining the lighting data of the target area includes: Calculate the light intensity change rate of the target area according to the ambient light intensity of the target area, the average light intensity within the sliding time window length, and the light intensity standard deviation within the sliding time window length; wherein the ambient light intensity is generated by a light intensity detector, and the light intensity detector is arranged on a vehicle traveling in the target area; Calculating the flicker frequency of the lighting equipment in the target area through multiple frames of target images; wherein the target image includes the lighting equipment in the target area, and the target image is generated by capturing an image capturing device, and the image capturing device is disposed on a vehicle traveling in the target area; The predicted value corresponding to the target feature vector is calculated by a lighting anomaly detection model generated in advance or in real time, and when the predicted value is less than or equal to a fourth threshold, abnormal lighting data of the lighting equipment in the target area is generated; wherein, the target feature vector is generated based on the ambient light intensity of the target area at the current moment, the average light intensity within the sliding time window length, the standard deviation of the light intensity within the sliding time window length, the light intensity change rate, the light intensity fluctuation state and the flickering frequency.

7. The data processing method according to claim 6, characterized in that: The process of generating a lighting anomaly detection model in advance or in real time includes: Performing dimensionality reduction and nonlinear feature extraction on the training feature vector using an autoencoder to obtain a reduced-dimensional feature vector; wherein the training feature vector is generated based on the ambient light intensity of the target area at a historical moment, the average light intensity within the sliding time window length, the standard deviation of the light intensity within the sliding time window length, the light intensity change rate, the light intensity fluctuation state, and the flicker frequency; The dimension-reduced feature vector is used as a training sample, and a pre- or real-time established objective function and constraint conditions are added, and classification training is performed on the training sample to generate the lighting anomaly detection model.

8. The data processing method according to claim 6 or 7, characterized in that: The process of determining the light intensity fluctuation state of the target area includes: Determine the light intensity fluctuation state of the lighting device in the target area according to the ambient light intensity of the target area, the average light intensity within the sliding time window length, and the light intensity standard deviation within the sliding time window length, and use the light intensity fluctuation state of the lighting device in the target area as the light intensity fluctuation state of the target area; Among them, when the difference between the ambient light intensity and the average light intensity within the sliding time window length is greater than or equal to the second threshold, the light intensity fluctuation state of the lighting device in the target area is determined to be abnormal; or, when the light intensity standard deviation within the sliding time window length is greater than or equal to the third threshold, the light intensity fluctuation state of the lighting device in the target area is determined to be abnormal.

9. The data processing method according to claim 5 or 6, characterized in that: If the traffic data includes lighting data, the process of encapsulating and encrypting the lighting data, generating a data packet, and wirelessly transmitting the data packet to the cloud server includes: Obtaining metadata associated with abnormal lighting data of the lighting device in the target area, and forming an original data packet based on the abnormal lighting data and the metadata; the metadata includes: vehicle number, vehicle geographic location coordinate information, timestamp, abnormal type of the lighting device, abnormal state of the lighting device, and data hash value; wherein the geographic location coordinates of the original data packet include the vehicle geographic location coordinate information in the metadata; The original data packet is encrypted to obtain an encrypted data packet, and the encrypted data packet is wirelessly transmitted to a cloud server.

10. The data processing method according to claim 9, characterized in that: When the encrypted data packet is wirelessly transmitted to the cloud server, the method further includes: Calculating signal strength during wireless transmission, and adjusting transmit power and modulation mode based on the signal strength and current channel conditions; and, The forward error correction mechanism and automatic retransmission request mechanism are embedded in the wireless transmission process.

11. A data processing system, characterized in that: Applied to a cloud server, the system includes: A data receiving module, used to receive a data packet transmitted wirelessly by the vehicle and decrypt the data packet; wherein the data packet is generated by encapsulating and encrypting traffic data in a traffic road area, and the traffic data includes lighting data; The abnormal area module is used to perform abnormal clustering of data points according to the decryption result of the data packet, and generate an abnormal area based on the abnormal clustering result; including: selecting a data point as a target data point, wherein the data point includes the geographical location coordinate information of the data packet; taking the target data point as the center, recording the minimum number of neighborhood data points of the target data point, and when the minimum number of neighborhood data points is greater than or equal to a first threshold, marking the target data point as a core data point, and adding the corresponding data point to an abnormal cluster; wherein the minimum neighborhood data point includes a data point located within a preset neighborhood radius of the target data point; performing regional association according to all boundary data points in the abnormal clustering result to generate the abnormal area; wherein the boundary data point is located within the preset neighborhood radius of the core data point and is not a core data point; The warning push module is used to push warning information to vehicles within a preset distance range of the abnormal area according to the vehicle's geographical location coordinate information.

12. A data processing system, characterized in that: Applied to a vehicle, the system comprises: A data packet generation module, used to obtain traffic data of a target area, and encapsulate and encrypt the traffic data to generate a data packet; wherein the target area includes a traffic road area determined in advance or in real time, and the traffic data includes lighting data; A wireless transmission module is used to wirelessly transmit the data packet to a cloud server so that the cloud server decrypts the data packet, performs data point abnormal clustering based on the data packet decryption result, and generates an abnormal area based on the abnormal clustering result; the process of the cloud server generating the abnormal area includes: selecting a data point as a target data point, wherein the data point includes the geographic location coordinate information of the data packet; taking the target data point as the center, recording the minimum number of neighborhood data points of the target data point, and when the minimum number of neighborhood data points is greater than or equal to a first threshold, marking the target data point as a core data point, and adding the corresponding data point to an abnormal cluster; wherein the minimum neighborhood data point includes a data point located within a preset neighborhood radius of the target data point; performing regional association according to all boundary data points in the abnormal clustering result to generate the abnormal area; wherein the boundary data point is located within the preset neighborhood radius of the core data point and is not a core data point; The warning receiving module is used to receive the warning information generated and pushed by the cloud server when the vehicle is within a preset distance range from the abnormal area.

13. A cloud server, characterized in that: The cloud server is used to execute the data processing method as described in any one of claims 1 to 4, or to execute the data processing system as described in claim 11.

14. A vehicle, characterized in that: The vehicle is used to execute the data processing method according to any one of claims 5 to 10, or to execute the data processing system according to claim 12.

Images