AI-based vehicle networking data recognition method and device

Abstract

The present invention discloses an AI-based method and device for identifying Internet of Vehicles (IoV) data, which mainly relates to the field of intelligent vehicle technology. The main purpose is to solve the problem that the identification security and effectiveness of existing IoV data are low, resulting in low vehicle driving safety in the case of intelligent assisted driving. It mainly includes obtaining vehicle operation data and user operation data flowing through the UPF network element, and extracting the first probe in the vehicle operation data and the second probe in the user operation data; when the state positions of the first probe and the second probe are in an aligned state, extracting the operation characteristics and the operation characteristics; if the operation characteristics match the abnormal operation characteristics of the first probe, the operation characteristics match the abnormal operation characteristics of the second probe, and the numerical results of the operation characteristics and the operation characteristics match the preset IoV abnormality threshold, then the IoV data abnormality identification result of the vehicle is generated. It is mainly used to identify IoV data.

Description

Technical Field

[0001] The present invention relates to the field of intelligent vehicle technology, and in particular to an AI-based vehicle networking data recognition method and device. Background Art

[0002] With the rapid development of the automotive industry, intelligent assisted driving functional components have come into focus and have become a key R&D target for automakers. The UPF is the only network element in the 5G core network dedicated to user-plane data processing. Its primary responsibilities include packet routing and forwarding, protocol data unit (PDU) session anchoring, and policy enforcement and traffic management. To ensure the safety of intelligent assisted driving, the UPF, as the core element for user-plane performance in the 5G core network, supports key 5G features such as high bandwidth, low latency, and network slicing.

[0003] Currently, existing methods for identifying IoV data typically involve blacklisting or whitelisting traffic data passing through UPF network elements, or performing threshold determination. However, the network data in the blacklist or whitelist is relatively fixed, which cannot meet the rapid identification requirements of big data that changes at any time. Furthermore, threshold determination cannot meet the high-precision monitoring requirements of flexible data changes in different scenarios. Consequently, this significantly reduces the security and effectiveness of IoV data identification, thereby reducing vehicle driving safety in intelligent assisted driving scenarios. Summary of the Invention

[0004] In view of this, the present invention provides an AI-based vehicle network data recognition method and device, the main purpose of which is to solve the problem that the recognition security and effectiveness of existing vehicle network data are low, resulting in low vehicle driving safety under intelligent assisted driving conditions.

[0005] According to one aspect of the present invention, a method for identifying Internet of Vehicles data based on AI is provided, comprising:

[0006] In response to the Internet of Vehicles data identification instruction, obtain the vehicle operation data and the user operation data flowing through the UPF network element, and extract the first probe in the vehicle operation data and the second probe in the user operation data respectively;

[0007] When the positions of the first probe and the second probe are aligned, extracting the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data;

[0008] If the operating feature matches the abnormal operating feature of the first probe, the operating feature matches the abnormal operating feature of the second probe, and the numerical results of the operating feature and the operating feature match the preset Internet of Vehicles abnormality threshold, then the Internet of Vehicles data abnormality identification result of the vehicle is generated.

[0009] Furthermore, the method further comprises:

[0010] Loading a probe configuration list, and when generating the vehicle operation data and the user operation data, adding probes to the vehicle operation data and the user operation data based on the probe configuration items in the probe configuration list;

[0011] Among them, the probe configuration list includes a first probe configuration item in different vehicle operation data and a second probe configuration item in different user operation data. The vehicle operation data includes vehicle speed, driving direction, vehicle remaining power or vehicle remaining fuel, and vehicle auxiliary services. The user operation data includes network download operations and network auxiliary services.

[0012] Furthermore, when the first probe and the second probe are in an aligned state, before extracting the operating characteristics of the vehicle operating data and the operating characteristics of the user operating data, the method further includes:

[0013] A first probe state sequence and a second probe state sequence are created, each of which includes a combination sequence between at least two state positions, and a probe pointer is moved and pointed in the first probe state sequence and the second probe state sequence to determine the first state position of the first probe and the second state position of the second probe.

[0014] Furthermore, when the first probe and the second probe are in an aligned state, extracting the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data includes:

[0015] When the first state position and the second state position are the same state position, determining that the state positions are in an aligned state;

[0016] The vehicle operation data is subjected to feature extraction based on the operation feature extraction model for which model training has been completed to obtain operation features, and the user operation data is subjected to feature extraction based on the operation feature extraction model for which model training has been completed to obtain operation features.

[0017] Furthermore, the method further comprises:

[0018] If the operating characteristic does not match the abnormal operating characteristic of the first probe, and / or the operating characteristic does not match the abnormal operating characteristic of the second probe, and the numerical results of the operating characteristic and the operating characteristic match a preset vehicle network abnormality threshold, determining a repair item corresponding to the abnormal operating characteristic and / or the abnormal operating characteristic, and performing an operating repair or an operating repair on the vehicle based on the repair item;

[0019] If the operating characteristics do not match the abnormal operating characteristics of the first probe, the operating characteristics do not match the abnormal operating characteristics of the second probe, and the numerical results of the operating characteristics and the operating characteristics do not match the preset Internet of Vehicles abnormality threshold, the next round of Internet of Vehicles data monitoring steps is initiated.

[0020] Furthermore, the method further comprises:

[0021] When the vehicle starts the assisted driving service or the vehicle starts the Internet service, it is determined whether the vehicle's Internet of Vehicles data recognition condition is triggered.

[0022] If the vehicle network data identification condition of the vehicle is triggered, the vehicle network data identification instruction is generated to execute the steps of obtaining the vehicle operation data and user operation data flowing through the UPF network element.

[0023] Furthermore, after generating the vehicle network data anomaly identification result, the method further includes:

[0024] If no repair instruction or change instruction initiated for the abnormal identification result of the Internet of Vehicles data is detected within a preset time interval, an abnormality warning is issued to the client of the vehicle.

[0025] According to another aspect of the present invention, a method and device for identifying Internet of Vehicles data based on AI are provided, comprising:

[0026] An acquisition module, configured to obtain vehicle operation data and user operation data flowing through the UPF network element in response to a vehicle network data identification instruction, and extract a first probe from the vehicle operation data and a second probe from the user operation data respectively;

[0027] an extraction module, configured to extract, when the state positions of the first probe and the second probe are aligned, an operation feature of the vehicle operation data and an operation feature of the user operation data;

[0028] A generation module is used to generate an abnormal vehicle network data recognition result for the vehicle if the operating characteristics match the abnormal operating characteristics of the first probe, the operating characteristics match the abnormal operating characteristics of the second probe, and the numerical results of the operating characteristics and the operating characteristics match the preset vehicle network abnormality threshold.

[0029] Furthermore, the device further comprises:

[0030] a probe adding module, configured to load a probe configuration list and, when generating the vehicle operation data and the user operation data, add probes to the vehicle operation data and the user operation data based on the probe configuration items in the probe configuration list;

[0031] Among them, the probe configuration list includes a first probe configuration item in different vehicle operation data and a second probe configuration item in different user operation data. The vehicle operation data includes vehicle speed, driving direction, vehicle remaining power or vehicle remaining fuel, and vehicle auxiliary services. The user operation data includes network download operations and network auxiliary services.

[0032] Furthermore, the device further comprises:

[0033] A creation module is used to create a first probe state sequence and a second probe state sequence, each of which includes a combination sequence between at least two state positions, and to move and point a probe pointer in the first probe state sequence and the second probe state sequence to determine the first state position of the first probe and the second state position of the second probe.

[0034] Furthermore, the extraction module includes:

[0035] a determining unit, configured to determine that the state positions are in an aligned state when the first state position and the second state position are the same state position;

[0036] An extraction unit is used to perform feature extraction on the vehicle operation data based on an operation feature extraction model that has completed model training to obtain operation features, and to perform feature extraction on the user operation data based on an operation feature extraction model that has completed model training to obtain operation features.

[0037] Furthermore, the device further comprises:

[0038] a first matching module configured to, if the operating characteristic does not match the abnormal operating characteristic of the first probe and / or the operating characteristic does not match the abnormal operating characteristic of the second probe, and the numerical results of the operating characteristic and the operating characteristic match a preset vehicle network abnormality threshold, determine a repair item corresponding to the abnormal operating characteristic and / or the abnormal operating characteristic, and perform an operating repair or an operating repair on the vehicle based on the repair item;

[0039] The second matching module is used to start the next round of Internet of Vehicles data monitoring steps if the operating characteristics do not match the abnormal operating characteristics of the first probe, the operating characteristics do not match the abnormal operating characteristics of the second probe, and the numerical results of the operating characteristics and the operating characteristics do not match the preset Internet of Vehicles abnormality threshold.

[0040] Furthermore, the device further comprises:

[0041] A determination module is used to determine whether the vehicle's Internet of Vehicles data recognition condition is triggered when the vehicle starts the assisted driving service or the vehicle starts the Internet service.

[0042] The generation module is also used to generate the vehicle network data identification instruction if the vehicle network data identification condition is triggered, so as to execute the step of obtaining vehicle operation data and user operation data flowing through the UPF network element.

[0043] Furthermore, the device further comprises:

[0044] The early warning module is used to issue an abnormality early warning to the client of the vehicle if no repair instruction or change instruction initiated for the abnormal identification result of the Internet of Vehicles data is detected within a preset time interval.

[0045] According to another aspect of the present invention, a storage medium is provided, in which at least one executable instruction is stored. The executable instruction enables a processor to perform operations corresponding to the above-mentioned AI-based Internet of Vehicles data recognition method.

[0046] According to another aspect of the present invention, there is provided a terminal, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other via the communication bus;

[0047] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the above-mentioned AI-based Internet of Vehicles data recognition method.

[0048] By means of the above technical solution, the technical solution provided by the embodiment of the present invention has at least the following advantages:

[0049] The present invention provides an AI-based IoV data recognition method and device. First, in response to an IoV data recognition instruction, the method obtains vehicle operation data and user operation data flowing through a UPF network element, and extracts a first probe from the vehicle operation data and a second probe from the user operation data. When the state positions of the first and second probes are aligned, the method extracts the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data. If the operation characteristics match the abnormal operation characteristics of the first probe, the operation characteristics match the abnormal operation characteristics of the second probe, and the numerical results of the operation characteristics and the operation characteristics match a preset IoV anomaly threshold, an IoV data anomaly recognition result for the vehicle is generated. Compared with the prior art, the present invention implements the accurate extraction of vehicle operation data and user operation data by embedding the first and second probes in the vehicle operation data and user operation data, respectively. Furthermore, through state position alignment and feature extraction, the method accurately associates and extracts the vehicle operation characteristics and user operation characteristics, thereby improving adaptability to flexible data changes, meeting data changes in different vehicle usage scenarios, and improving the accurate detection of abnormal data and offensive behaviors, thereby ensuring vehicle driving safety in intelligent assisted driving.

[0050] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0052] Figure 1 A flowchart of an AI-based vehicle networking data recognition method provided by an embodiment of the present invention is shown;

[0053] Figure 2 A schematic diagram of a probe state sequence provided by an embodiment of the present invention is shown;

[0054] Figure 3 The figure shows a block diagram of a vehicle network data recognition device based on AI provided by an embodiment of the present invention;

[0055] Figure 4 A schematic structural diagram of a terminal provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0056] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0057] In order to solve the problem that the existing vehicle network data identification security and effectiveness are low, resulting in low vehicle driving safety in the case of intelligent assisted driving. The embodiment of the present invention provides an AI-based vehicle network data identification method, such as Figure 1 As shown, the method includes:

[0058] 101. In response to an Internet of Vehicles data identification instruction, obtain vehicle operation data and user operation data flowing through a UPF network element, and extract a first probe from the vehicle operation data and a second probe from the user operation data, respectively.

[0059] In this embodiment of the present invention, by parsing the real-time data stream transmitted within the IoV core network element (UPF), probe extraction logic is built for both vehicle operation data and user operation data, achieving precise data feature extraction. In the field of IoV data collection, a probe refers to a structured feature data segment with clear business semantics extracted from the raw data stream. Essentially, it performs periodic diagnostics on container execution. These probes can be implemented using three types of probes: HTTP probes, TCP probes, and command probes. The first type of probe focuses on capturing vehicle operation-related data, such as speed, engine status, GPS location, and driving direction. The second type of probe captures user operation data, such as control commands sent via mobile apps, changes to vehicle settings, and entertainment system operations. This data can reflect user interaction with the vehicle. IoV data recognition commands can be triggered based on user clicks or preset conditions, such as when the target vehicle's mileage reaches a preset value or when the vehicle's continuous driving duration reaches a preset value. This is not specifically limited in this embodiment of the present invention.

[0060] 102. When the positions of the first probe and the second probe are aligned, extract operating characteristics of the vehicle operating data and operating characteristics of the user operating data.

[0061] In this embodiment of the present invention, state and position alignment refers to the situation where the positions or states of two probes (a first probe and a second probe) in the data stream have reached a certain synchronization or matching condition. In IoV data processing, when the state and position of the first probe (the vehicle operation data probe) and the second probe (the user operation data probe) are aligned, it means that the data they capture is temporally or logically consistent or synchronized. For example, the two probes capture data at the same time or within the same time period, ensuring temporal consistency. This is crucial for analyzing the temporal relationship between vehicle operation status and user operations. For example, when a user sends a control command via a mobile app, the user operation data probe captures the command, while the vehicle operation data probe captures the vehicle's response. When these two data are logically synchronized, the relationship between the command and response can be analyzed to evaluate system performance and reliability. When the state and position of the first and second probes are aligned, feature extraction is performed based on the feature extraction models corresponding to the different data types. Specifically, the vehicle operation data is extracted using the operation feature extraction model, while the user operation data is extracted using the operation feature extraction model.

[0062] 103. If the operating feature matches the abnormal operating feature of the first probe, the operating feature matches the abnormal operating feature of the second probe, and the numerical results of the operating feature and the operating feature match the preset Internet of Vehicles abnormality threshold, then an Internet of Vehicles data abnormality identification result of the vehicle is generated.

[0063] In this embodiment of the present invention, after feature extraction, the features are analyzed from two perspectives: feature matching and numerical analysis, to jointly identify anomalies in the IoV data. Feature matching involves matching preset abnormal operating and operational features with the corresponding feature extraction results. Numerical analysis involves normalizing features of different dimensions, assigning feature weights based on their priorities, and calculating the numerical results of the operating and operational features by multiplying the normalized values ​​by the corresponding feature weights. The IoV anomaly threshold can be set based on specific historical empirical data and is not specifically limited in this embodiment. Abnormal operating features can include static and dynamic features. Static features are based on fields or values ​​extracted by the first probe, such as engine status, vehicle speed, and battery voltage. Dynamic features are based on trend deviation detection in historical operating data, such as a 30% decrease in oil pressure compared to the previous 10-minute average for more than two minutes. Abnormal operational features include abnormal operation frequency and abnormal operation combination. Abnormal operation frequency can include thresholds for air conditioning adjustment or entertainment system clicks. Operation combination anomalies are based on the timing correlation or operation conflict of multiple operation instructions. For example, if the seat adjustment command is not detected within 5 seconds after the vehicle is unlocked, there may be a risk of vehicle theft; when the vehicle driving mode is cruise control, the user clicks on the nap mode.

[0064] It should be noted that AI (Artificial Intelligence) technology monitors and analyzes IoV traffic in real time as it passes through UPF network elements and probes. Machine learning algorithms are used to model data characteristics and behavior patterns, enabling accurate detection and identification of abnormal data and attack behaviors. This improves the security and effectiveness of IoV data identification.

[0065] In one embodiment of the present invention, for further illustration and limitation, the method further includes:

[0066] A probe configuration list is loaded, and when generating the vehicle operation data and the user operation data, probes are added to the vehicle operation data and the user operation data based on the probe configuration items in the probe configuration list.

[0067] In an embodiment of the present invention, the probe configuration list includes a first probe configuration item in different vehicle operation data and a second probe configuration item in different user operation data. The vehicle operation data includes vehicle speed, driving direction, vehicle remaining power or vehicle remaining fuel, and vehicle auxiliary services. The user operation data includes network download operations and network auxiliary services. The probe configuration list can be stored in a configuration center for full loading when the vehicle is started or for dynamic loading when a vehicle network data identification instruction is received. The first probe configuration item can be speed, power, auxiliary service status, etc.; the second probe configuration item can be a download operation, network instruction, unlock record, etc. When incremental vehicle operation data and user operation data are monitored, a probe identifier is attached to the corresponding data to form a probe data stream.

[0068] In one embodiment of the present invention, for further explanation and limitation, when the state positions of the first probe and the second probe are in an aligned state, before extracting the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data, the method further includes:

[0069] A first probe state sequence and a second probe state sequence are created.

[0070] In an embodiment of the present invention, the first probe state sequence and the second probe state sequence each include a combination sequence between at least two state positions. The first probe state sequence is a sequence consisting of the states (enabled / disabled) of the first probe on the timeline, for example: [enabled, disabled, enabled, ...]. Similarly, the state sequence of the second probe is, for example: [disabled, enabled, enabled, ...]. Each state position (such as a time point or time period) determines whether the probe activates the data extraction function. When the probe state position is enabled, the probe will perform data extraction, that is, extract the corresponding data. When the probe state position is disabled, the current probe is invalid. After obtaining the first probe state sequence and the second probe state sequence, the probe pointer moves and points in the first probe state sequence and the second probe state sequence to determine the first state position of the first probe and the second state position of the second probe. The pointer slides in the state sequence according to a preset rule (such as time sequence or event triggering) to point to the current state position.

[0071] It should be noted that if Figure 2As shown, each movement of the probe pointer represents a state position. However, the time intervals occupied by different state positions can be equal or unequal. This allows different state positions to occupy different time intervals (rather than fixed equal intervals), which better reflects the dynamic characteristics of real-world scenarios (such as sudden maneuvers or continuous states in driving behavior). In an equal-duration sequence, each state position corresponds to a fixed time slice (e.g., 100ms), and the pointer moves in fixed steps. In an unequal-duration sequence, each state position is accompanied by a time span, and the pointer movement step size is determined by the current state.

[0072] In one embodiment of the present invention, for further explanation and limitation, the step of extracting the operating characteristics of the vehicle operating data and the operating characteristics of the user operation data when the state positions of the first probe and the second probe are in an aligned state includes:

[0073] When the first state position and the second state position are the same state position, determining that the state positions are in an aligned state;

[0074] The vehicle operation data is subjected to feature extraction based on the operation feature extraction model for which model training has been completed to obtain operation features, and the user operation data is subjected to feature extraction based on the operation feature extraction model for which model training has been completed to obtain operation features.

[0075] In an embodiment of the present invention, when the pointer moves, the enabled / disabled state of the probe is dynamically updated based on the state position pointed to by the current probe: if it points to "enabled", the data corresponding to the configuration item (such as vehicle CAN bus data or user operation log) is extracted. If it points to "disabled", data extraction is skipped. When the pointers of the two probes point to the "enabled" state at the same time, it is regarded as an aligned state, triggering joint feature extraction. Alignment conditions include: time synchronization, overlapping enabled time periods of the two probes, and logical associations, such as the vehicle acceleration state and the user's deep accelerator operation being enabled at the same time. Feature extraction is triggered by the joint conditions of vehicle status (such as "stationary charging") and user operation (such as "initiating over-the-air download"), and data is processed only in specific states, reducing the model inference load, ensuring that the extracted features are strongly associated with the current scene, and avoiding invalid data processing.

[0076] Feature extraction is triggered by state and position alignment, and pre-trained models (operation feature extraction model and operation feature extraction model) are used to process vehicle operation data and user operation data, respectively. The operation feature extraction model can be a time series model, such as a long short-term memory network (LSTM) or a convolutional neural network (CNN). It can also be a statistical model constructed based on statistical methods such as sliding window mean and variance. The operation feature extraction model takes vehicle operation data (such as speed, engine speed, gear position, etc.) as input and outputs operation features (such as average acceleration, engine load factor, and frequent gear shifting). To extract features for online download operations (over-the-air upgrades, multimedia downloads) and online auxiliary services (such as real-time navigation and remote diagnosis) in user operation data, the model construction must be based on the temporal nature of the operation behavior, system interactivity, and user intent. This model can be constructed using an extreme gradient boosting (XGBoost) combined with a multimodal feature fusion model, but this is not specifically limited in this embodiment. The running feature model focuses on the time series analysis of physical quantities (such as LSTM prediction of acceleration changes), while the operation feature model focuses on behavioral intention mining. Pre-building and training corresponding feature extraction models for different data characteristics can make the model division of labor clear, thereby improving the accuracy of feature extraction.

[0077] In one embodiment of the present invention, for further illustration and limitation, the method further includes:

[0078] If the operating characteristic does not match the abnormal operating characteristic of the first probe, and / or the operating characteristic does not match the abnormal operating characteristic of the second probe, and the numerical results of the operating characteristic and the operating characteristic match a preset vehicle network abnormality threshold, determining a repair item corresponding to the abnormal operating characteristic and / or the abnormal operating characteristic, and performing an operating repair or an operating repair on the vehicle based on the repair item;

[0079] If the operating characteristics do not match the abnormal operating characteristics of the first probe, the operating characteristics do not match the abnormal operating characteristics of the second probe, and the numerical results of the operating characteristics and the operating characteristics do not match the preset Internet of Vehicles abnormality threshold, the next round of Internet of Vehicles data monitoring steps is initiated.

[0080] In this embodiment of the present invention, if the operating characteristics and abnormal operating characteristics, the operating characteristics and abnormal operating characteristics, or the digitized results of the operating characteristics and the operating characteristics do not fully match the preset IoV abnormality threshold, the vehicle will be repaired based on the mismatch if the digitized results of the operating characteristics and the operating characteristics match the preset IoV abnormality threshold and at least one of the operating characteristics and the abnormal operating characteristics or the operating characteristics and the abnormal operating characteristics matches. For example, the operating characteristics may be an engine speed fluctuation rate of 7% (abnormal operating characteristics of the first probe: engine speed fluctuation rate <5%) and a gear shift jerk intensity of 0.4g (abnormal operating characteristics of the first probe: gear shift jerk intensity <0.3g acceleration change). The operating characteristics may be a real-time navigation signal interruption frequency of 0.3 times / hour (abnormal operating characteristics of the first probe: navigation interruption rate >0.2 times / hour) and a remote diagnostic request failure rate of 2% (abnormal operating characteristics of the first probe: remote diagnostic request failure rate <3%). The numerical combination ("Speed ​​Fluctuation 7% + Navigation Interruption Rate 0.3 Times / Hour") matches the preset IoV abnormality threshold (e.g., "Speed ​​Fluctuation > 6% and Navigation Interruption Rate > 0.2 Times / Hour"). At this point, the numerical results of the operating and operational characteristics match the preset IoV abnormality threshold. The operating characteristics match the abnormal operating characteristics of the first probe, while the operational characteristics do not match the abnormal operating characteristics of the second probe. The repair item corresponding to the abnormal operating characteristics is determined, the fuel injection parameters are adjusted, and the solenoid valve self-test procedure is activated.

[0081] It should be noted that different abnormal operating characteristics and abnormal operation characteristics are mapped to their corresponding repair items in a pre-established relationship. For example, if the abnormal operating characteristic is an excessive engine speed fluctuation rate, the corresponding repair item is to dynamically lower the injection pressure threshold and activate the solenoid valve pulse cleaning program; if the abnormal operation characteristic is navigation data packet transmission congestion, the corresponding repair item is to switch to the backup communication frequency band to enhance signal anti-interference capabilities. In addition, when multiple repair items are triggered simultaneously, different repair items are configured with corresponding execution priorities. Due to the higher safety risks of the power system, the execution priority of the operation characteristic repair item can be configured to be higher than the operation characteristic repair item.

[0082] In one embodiment of the present invention, for further illustration and limitation, the method further includes:

[0083] When the vehicle starts the assisted driving service or the vehicle starts the Internet service, it is determined whether the vehicle's Internet of Vehicles data recognition condition is triggered.

[0084] If the vehicle network data identification condition of the vehicle is triggered, the vehicle network data identification instruction is generated to execute the steps of obtaining the vehicle operation data and user operation data flowing through the UPF network element.

[0085] In the embodiment of the present invention, the Internet of Vehicles data recognition instruction is triggered according to the Internet of Vehicles data recognition condition when the current vehicle starts the assisted driving service or the vehicle starts the Internet service. Different vehicles may have different Internet of Vehicles data recognition conditions according to different services purchased by the user or different user settings. The Internet of Vehicles data recognition condition can be a single trigger condition or a combined trigger condition. When the user turns on L2+ assisted driving (including adaptive cruise control + lane centering),

[0086] Identification conditions for connected vehicle data can include one or more of cumulative mileage, number of passengers, and a time window. Cumulative mileage can be determined by cross-validating the vehicle's CAN bus mileage counter with the GPS trajectory. The number of passengers can be defined as a driver's seat and two rear passengers, with a pressure threshold of ≥50 kg per person. The time window can be configured to trigger between 06:00 and 22:00 (excluding nighttime low-traffic periods to avoid impacting the user experience). The identification condition for cumulative mileage can be greater than 80 kilometers since the last data collection; the identification condition for the number of passengers can be ≥3 passengers identified by the seat pressure sensor and camera. The trigger conditions for connected vehicle data identification commands are configured based on the services purchased by the user and the user's personalized configuration, enabling personalized triggering of connected data identification.

[0087] In one embodiment of the present invention, for further explanation and limitation, after the step of generating the vehicle network data anomaly identification result, the method further includes:

[0088] If no repair instruction or change instruction initiated for the abnormal identification result of the Internet of Vehicles data is detected within a preset time interval, an abnormality warning is issued to the client of the vehicle.

[0089] In an embodiment of the present invention, vehicle network data includes vehicle operation data and vehicle operation data. Detecting an anomaly during the process of identifying anomalies in the connected vehicle data and initiating a repair or modification instruction based on the identified anomaly results are expected to be highly likely events. If no relevant instructions are received for an extended period, it may be that the vehicle's client has experienced a malfunction or has lost its communication connection with the testing center. Therefore, it is necessary to detect the time interval between the current moment and the last time a repair or modification instruction was received based on the identified anomaly results in the connected vehicle data. If the time interval is greater than a preset time interval, an anomaly warning can be generated for the vehicle's client.

[0090] The present invention provides an AI-based IoV data recognition method. First, in response to an IoV data recognition instruction, the method obtains vehicle operation data and user operation data flowing through a UPF network element, and extracts a first probe from the vehicle operation data and a second probe from the user operation data. When the state positions of the first and second probes are aligned, the method extracts the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data. If the operation characteristics match the abnormal operation characteristics of the first probe, the operation characteristics match the abnormal operation characteristics of the second probe, and the numerical results of the operation characteristics and the operation characteristics match a preset IoV anomaly threshold, an IoV data anomaly recognition result for the vehicle is generated. Compared with the prior art, the present invention implements the accurate extraction of vehicle operation data and user operation data by embedding the first and second probes in the vehicle operation data and user operation data, respectively. Furthermore, through state position alignment and feature extraction, the method accurately associates and extracts the vehicle operation characteristics and user operation characteristics, thereby improving adaptability to flexible data changes, meeting data changes in different vehicle usage scenarios, and improving the accurate detection of abnormal data and offensive behaviors, thereby ensuring vehicle driving safety in intelligent assisted driving.

[0091] Furthermore, as a response to the above Figure 1 The embodiment of the present invention provides an AI-based vehicle network data recognition device, such as Figure 3 As shown, the device includes:

[0092] An acquisition module 31 is configured to obtain vehicle operation data and user operation data flowing through the UPF network element in response to a vehicle network data identification instruction, and to extract a first probe from the vehicle operation data and a second probe from the user operation data respectively;

[0093] an extraction module 32 for extracting, when the positions of the first probe and the second probe are aligned, the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data;

[0094] The generation module 33 is used to generate an abnormal vehicle network data recognition result of the vehicle if the operating characteristics match the abnormal operating characteristics of the first probe, the operating characteristics match the abnormal operating characteristics of the second probe, and the numerical results of the operating characteristics and the operating characteristics match the preset vehicle network abnormality threshold.

[0095] Furthermore, the device further comprises:

[0096] a probe adding module, configured to load a probe configuration list and, when generating the vehicle operation data and the user operation data, add probes to the vehicle operation data and the user operation data based on the probe configuration items in the probe configuration list;

[0097] Among them, the probe configuration list includes a first probe configuration item in different vehicle operation data and a second probe configuration item in different user operation data. The vehicle operation data includes vehicle speed, driving direction, vehicle remaining power or vehicle remaining fuel, and vehicle auxiliary services. The user operation data includes network download operations and network auxiliary services.

[0098] Furthermore, the device further comprises:

[0099] A creation module is used to create a first probe state sequence and a second probe state sequence, each of which includes a combination sequence between at least two state positions, and to move and point a probe pointer in the first probe state sequence and the second probe state sequence to determine the first state position of the first probe and the second state position of the second probe.

[0100] Furthermore, the extraction module 32 includes:

[0101] a determining unit, configured to determine that the state positions are in an aligned state when the first state position and the second state position are the same state position;

[0102] An extraction unit is used to perform feature extraction on the vehicle operation data based on an operation feature extraction model that has completed model training to obtain operation features, and to perform feature extraction on the user operation data based on an operation feature extraction model that has completed model training to obtain operation features.

[0103] Furthermore, the device further comprises:

[0104] a first matching module configured to, if the operating characteristic does not match the abnormal operating characteristic of the first probe and / or the operating characteristic does not match the abnormal operating characteristic of the second probe, and the numerical results of the operating characteristic and the operating characteristic match a preset vehicle network abnormality threshold, determine a repair item corresponding to the abnormal operating characteristic and / or the abnormal operating characteristic, and perform an operating repair or an operating repair on the vehicle based on the repair item;

[0105] The second matching module is used to start the next round of Internet of Vehicles data monitoring steps if the operating characteristics do not match the abnormal operating characteristics of the first probe, the operating characteristics do not match the abnormal operating characteristics of the second probe, and the numerical results of the operating characteristics and the operating characteristics do not match the preset Internet of Vehicles abnormality threshold.

[0106] Furthermore, the device further comprises:

[0107] A determination module is used to determine whether the vehicle's Internet of Vehicles data recognition condition is triggered when the vehicle starts the assisted driving service or the vehicle starts the Internet service.

[0108] The generation module 33 is also used to generate the vehicle network data identification instruction if the vehicle network data identification condition is triggered, so as to execute the step of obtaining the vehicle operation data and user operation data flowing through the UPF network element.

[0109] Furthermore, the device further comprises:

[0110] The early warning module is used to issue an abnormality early warning to the client of the vehicle if no repair instruction or change instruction initiated for the abnormal identification result of the Internet of Vehicles data is detected within a preset time interval.

[0111] The present invention provides an AI-based IoV data recognition device. The device first responds to an IoV data recognition instruction, acquires vehicle operation data and user operation data flowing through a UPF network element, and extracts a first probe from the vehicle operation data and a second probe from the user operation data. When the state positions of the first and second probes are aligned, the device extracts the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data. If the operation characteristics match the abnormal operation characteristics of the first probe, the operation characteristics match the abnormal operation characteristics of the second probe, and the numerical results of the operation characteristics and the operation characteristics match a preset IoV anomaly threshold, the device generates an IoV data anomaly recognition result for the vehicle. Compared to the prior art, the present invention implements accurate extraction of vehicle operation data and user operation data by embedding the first and second probes in the vehicle operation data and user operation data, respectively. Furthermore, through state position alignment and feature extraction, the device accurately associates and extracts the vehicle operation characteristics and user operation characteristics. This improves adaptability to flexible data changes, meets data changes in different vehicle usage scenarios, and improves the accurate detection of abnormal data and offensive behaviors, thereby ensuring vehicle driving safety in intelligent assisted driving scenarios.

[0112] According to one embodiment of the present invention, a storage medium is provided, which stores at least one executable instruction. The computer-executable instruction can execute the AI-based Internet of Vehicles data recognition method in any of the above method embodiments.

[0113] Figure 4 A schematic structural diagram of a terminal provided according to an embodiment of the present invention is shown. The specific embodiment of the present invention does not limit the specific implementation of the terminal.

[0114] like Figure 4 As shown, the terminal may include: a processor (processor) 402 , a communications interface (Communications Interface) 404 , a memory (memory) 406 , and a communication bus 408 .

[0115] The processor 402 , the communication interface 404 , and the memory 406 communicate with each other via a communication bus 408 .

[0116] The communication interface 404 is used to communicate with other devices such as clients or other servers.

[0117] The processor 402 is used to execute the program 410, and specifically can execute the relevant steps in the above-mentioned embodiment of the AI-based Internet of Vehicles data recognition method.

[0118] Specifically, the program 410 may include program codes, which include computer operation instructions.

[0119] Processor 402 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The one or more processors included in the terminal may be processors of the same type, such as one or more CPUs, or processors of different types, such as one or more CPUs and one or more ASICs.

[0120] The memory 406 is used to store the program 410. The memory 406 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage.

[0121] The program 410 may be specifically configured to cause the processor 402 to perform the following operations:

[0122] In response to the Internet of Vehicles data identification instruction, obtain the vehicle operation data and the user operation data flowing through the UPF network element, and extract the first probe in the vehicle operation data and the second probe in the user operation data respectively;

[0123] When the positions of the first probe and the second probe are aligned, extracting the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data;

[0124] If the operating feature matches the abnormal operating feature of the first probe, the operating feature matches the abnormal operating feature of the second probe, and the numerical results of the operating feature and the operating feature match the preset Internet of Vehicles abnormality threshold, then the Internet of Vehicles data abnormality identification result of the vehicle is generated.

[0125] Obviously, those skilled in the art will appreciate that the various modules or steps of the present invention described above can be implemented using a general-purpose computing device, centralized on a single computing device, or distributed across a network of multiple computing devices. Alternatively, they can be implemented using program code executable by a computing device, which can then be stored in a storage device and executed by the computing device. In some cases, the steps shown or described can be performed in a different order than that shown, or can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0126] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for identifying Internet of Vehicles data based on AI, characterized in that: include: In response to the Internet of Vehicles data identification instruction, obtain the vehicle operation data and the user operation data flowing through the UPF network element, and extract the first probe in the vehicle operation data and the second probe in the user operation data respectively; When the positions of the first probe and the second probe are aligned, extracting the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data; If the operating feature matches the abnormal operating feature of the first probe, the operating feature matches the abnormal operating feature of the second probe, and the numerical results of the operating feature and the operating feature match the preset IoV abnormality threshold, then generating an IoV data abnormality recognition result for the vehicle; The method further comprises: Loading a probe configuration list, and when generating the vehicle operation data and the user operation data, adding probes to the vehicle operation data and the user operation data based on the probe configuration items in the probe configuration list; The probe configuration list includes a first probe configuration item in different vehicle operation data and a second probe configuration item in different user operation data, wherein the vehicle operation data includes vehicle speed, driving direction, vehicle remaining power or vehicle remaining fuel, and vehicle auxiliary services, and the user operation data includes network download operations and network auxiliary services; When the first probe and the second probe are in an aligned state, before extracting the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data, the method further includes: Creating a first probe state sequence and a second probe state sequence, each of which includes a combination sequence between at least two state positions, and moving and pointing a probe pointer in the first probe state sequence and the second probe state sequence to determine a first state position of the first probe and a second state position of the second probe; The method further comprises: When the vehicle starts the assisted driving service or the vehicle starts the Internet service, it is determined whether the vehicle's Internet of Vehicles data recognition condition is triggered. If the vehicle network data identification condition of the vehicle is triggered, the vehicle network data identification instruction is generated to execute the steps of obtaining the vehicle operation data and user operation data flowing through the UPF network element.

2. The method according to claim 1, characterized in that When the first probe and the second probe are in an aligned state, extracting the operation characteristics of the vehicle operation data and the operation characteristics of the user operation data includes: When the first state position and the second state position are the same state position, determining that the state positions are in an aligned state; The vehicle operation data is subjected to feature extraction based on the operation feature extraction model for which model training has been completed to obtain operation features, and the user operation data is subjected to feature extraction based on the operation feature extraction model for which model training has been completed to obtain operation features.

3. The method according to claim 1, characterized in that The method further comprises: If the operating characteristic does not match the abnormal operating characteristic of the first probe, and / or the operating characteristic does not match the abnormal operating characteristic of the second probe, and the numerical results of the operating characteristic and the operating characteristic match a preset vehicle network abnormality threshold, determining a repair item corresponding to the abnormal operating characteristic and / or the abnormal operating characteristic, and performing an operating repair or an operating repair on the vehicle based on the repair item; If the operating characteristics do not match the abnormal operating characteristics of the first probe, the operating characteristics do not match the abnormal operating characteristics of the second probe, and the numerical results of the operating characteristics and the operating characteristics do not match the preset Internet of Vehicles abnormality threshold, the next round of Internet of Vehicles data monitoring steps is initiated.

4. The method according to any one of claims 1 to 3, characterized in that After generating the vehicle network data anomaly identification result, the method further includes: If no repair instruction or change instruction initiated for the abnormal identification result of the Internet of Vehicles data is detected within a preset time interval, an abnormality warning is issued to the client of the vehicle.

5. An AI-based vehicle networking data recognition device, characterized in that: include: An acquisition module, configured to obtain vehicle operation data and user operation data flowing through the UPF network element in response to a vehicle network data identification instruction, and extract a first probe from the vehicle operation data and a second probe from the user operation data respectively; an extraction module, configured to extract, when the state positions of the first probe and the second probe are aligned, an operation feature of the vehicle operation data and an operation feature of the user operation data; a generating module configured to generate an IoV data anomaly recognition result for the vehicle if the operating characteristic matches an abnormal operating characteristic of the first probe, the operating characteristic matches an abnormal operating characteristic of the second probe, and the numerical results of the operating characteristic and the operating characteristic match a preset IoV anomaly threshold; The device further comprises: a probe adding module, configured to load a probe configuration list and, when generating the vehicle operation data and the user operation data, add probes to the vehicle operation data and the user operation data based on the probe configuration items in the probe configuration list; The probe configuration list includes a first probe configuration item in different vehicle operation data and a second probe configuration item in different user operation data, wherein the vehicle operation data includes vehicle speed, driving direction, vehicle remaining power or vehicle remaining fuel, and vehicle auxiliary services, and the user operation data includes network download operations and network auxiliary services; The device further comprises: a creation module, configured to create a first probe state sequence and a second probe state sequence, each of which includes a combination sequence between at least two state positions, and to determine a first state position of the first probe and a second state position of the second probe by moving and pointing a probe pointer in the first probe state sequence and the second probe state sequence; The device further comprises: A determination module is used to determine whether the vehicle's Internet of Vehicles data recognition condition is triggered when the vehicle starts the assisted driving service or the vehicle starts the Internet service. The generation module is also used to generate the vehicle network data identification instruction if the vehicle network data identification condition is triggered, so as to execute the step of obtaining vehicle operation data and user operation data flowing through the UPF network element.

6. A storage medium storing at least one executable instruction, wherein the executable instruction enables a processor to perform operations corresponding to the AI-based vehicle network data recognition method according to any one of claims 1 to 4.

7. An AI-based Internet of Vehicles terminal, comprising: A processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the AI-based Internet of Vehicles data recognition method as described in any one of claims 1 to 4.

Images