Internet of vehicles data processing method, device and equipment

Abstract

The application relates to a vehicle networking data processing method, device and equipment. The vehicle networking data processing method comprises the following steps: determining a link state and calculating a prediction confidence of the link state based on a communication link state parameter; calculating an initial sending priority of each to-be-sent data packet based on the link state and the prediction confidence; encapsulating the initial sending priority of the data packet and a data packet generation timestamp into a head identifier of the data packet, and sending the data packet carrying the head identifier to an edge node. The application can actively trigger priority promotion before the link state deteriorates, compensates for key data scheduling, significantly reduces the loss and timeout risk of event data packets in a dynamic vehicle networking environment, guarantees reliable data transmission, and solves the technical problem that important event data packets are prone to be lost or timed out in the related art, and reliable transmission of safety warning data cannot be guaranteed.

Description

Technical Field

[0001] This application relates to the fields of vehicle networking and edge computing technology, specifically to a vehicle networking data processing method, apparatus, and device. Background Technology

[0002] Currently, the Internet of Vehicles (IoV), as a core component of intelligent transportation systems, enables various applications such as driving safety warnings, traffic efficiency optimization, and in-vehicle information services through real-time communication between vehicles, between vehicles and roadside infrastructure, and between vehicles and cloud platforms. With the continuous development of technologies such as cellular IoV and dedicated short-range communication, the amount of data generated by vehicle terminals is experiencing explosive growth. However, problems such as dynamic changes in network topology caused by high-speed vehicle movement, fluctuations in communication link quality, and disordered data packet timing pose serious challenges to the real-time acquisition, reliable transmission, and efficient processing of IoV data.

[0003] In related technologies, vehicle-to-everything (V2X) data typically uses periodic or static communication modes. However, when vehicles move at high speeds, causing link deterioration, important event data packets are easily lost or timed out, making it impossible to guarantee the reliable transmission of safety warning data.

[0004] Therefore, it is necessary to design a new method for processing vehicle network data to overcome the above problems. Summary of the Invention

[0005] This application provides a vehicle network data processing method, apparatus, and device, which can solve the technical problem in related technologies that important event data packets are easily lost or timed out, making it impossible to guarantee the reliable transmission of safety warning data.

[0006] In a first aspect, embodiments of this application provide a vehicle-to-everything (V2X) data processing method, applied to a vehicle, the V2X data processing method comprising: The link state is determined based on the communication link state parameters, and the prediction confidence of the link state is calculated. The initial transmission priority of each data packet to be sent is calculated based on the link state and the predicted confidence level. The initial sending priority of the data packet and the data packet generation timestamp are encapsulated in the header identifier of the data packet, and the data packet carrying the header identifier is sent to the edge node.

[0007] In conjunction with the first aspect, in one implementation, determining the link state based on communication link state parameters and calculating the prediction confidence of the link state includes: Based on network latency Data packet loss rate Signal strength and bandwidth fluctuations Calculate the overall link quality index; An online recursive prediction model with a forgetting factor is constructed, which includes an exponentially weighted moving average recursive model and a prediction confidence function. The link status is determined based on the calculated comprehensive link quality index and the exponentially weighted moving average recursive model. The prediction confidence is calculated based on the calculated comprehensive link quality index and the prediction confidence function.

[0008] In conjunction with the first aspect, in one embodiment, the vehicle-to-everything (V2X) data processing method further includes: It receives feedback signals from edge nodes and adjusts the parameters of the exponentially weighted moving average recursive model based on these feedback signals.

[0009] In conjunction with the first aspect, in one implementation, determining the link status based on the calculated comprehensive link quality index and the exponentially weighted moving average recursive model includes: The smoothed link quality value at the current moment is calculated based on the comprehensive link quality index and the exponentially weighted moving average recursive model. If the quality value of the smooth link is detected to have decreased by more than the set quality value in a cumulative manner across multiple consecutive sampling points, it is determined that the link is deteriorating and the link status is recorded as deteriorating; otherwise, the link is considered to be in a stable state and the link status is recorded as stable.

[0010] In conjunction with the first aspect, in one implementation, calculating the initial transmission priority of each data packet to be sent based on link state and predicted confidence includes: The original priority of each data packet to be sent is calculated based on the criticality coefficient and the priority increment driven by the link state; whereby the criticality coefficient is predefined according to the event type. The original priority is smoothed by using a low-pass filter to obtain the initial transmission priority of each data packet to be sent.

[0011] In conjunction with the first aspect, in one embodiment, the vehicle-to-everything (V2X) data processing method further includes: If the calculated initial transmission priorities of multiple data packets to be sent are the same, a fine-tuning factor is added to the initial transmission priorities of the multiple data packets to be sent, and the fine-tuning factor is different for different data packets to be sent.

[0012] Secondly, embodiments of this application provide a vehicle-to-everything (V2X) data processing method applied to edge nodes, comprising: Receive data packets with header identifiers sent by the vehicle end. The header identifiers of the data packets contain the initial transmission priority of the data packets and the data packet generation timestamp. Data packets are reordered based on their initial sending priority and generation timestamp, and the data stream is output according to the logical timing. Perform anomaly detection and real-time aggregation on the data stream and output the final processing result.

[0013] In conjunction with the second aspect, in one embodiment, the vehicle-to-everything (V2X) data processing method further includes: Data packets in the data stream are grouped according to event type, and data packets from the same vehicle end and belonging to the same event type are defined as the same event stream; When the first data packet of an event stream is received, computing resources are allocated for the event and the state container is initialized. When subsequent data packets belonging to the same event stream arrive, the aggregation result in the state container is updated. If no data packets of the same event stream are received for a set period of time, the event is considered to have ended and the state container is released, and the final processing result is output.

[0014] In conjunction with the second aspect, in one embodiment, the vehicle-to-everything (V2X) data processing method further includes: The timestamp error variance introduced by reordering is calculated based on the packet generation timestamp, the time when the edge node outputs the packet, and the estimated average transmission delay of the packet from generation to output by the edge node. The packet out-of-order rate is calculated based on the number of packets that were out of order before entering the sorting window and the total number of packets in the window. The timestamp error variance and data packet out-of-order rate introduced by reordering are encapsulated as feedback signals and fed back to the vehicle.

[0015] Thirdly, embodiments of this application provide a vehicle-to-everything (V2X) data processing device, the V2X data processing device comprising: The first calculation module is used to determine the link state based on the communication link state parameters and calculate the prediction confidence of the link state. The second calculation module is used to calculate the initial transmission priority of each data packet to be sent based on the link status and prediction confidence. The encapsulation module is used to encapsulate the initial sending priority of the data packet and the data packet generation timestamp into the header identifier of the data packet, and send the data packet carrying the header identifier to the edge node.

[0016] Fourthly, this application provides a vehicle-to-everything (V2X) data processing device, which includes a processor, a memory, and a V2X data processing program stored in the memory and executable by the processor. When the V2X data processing program is executed by the processor, it implements the steps of the above-described V2X data processing method.

[0017] The beneficial effects of the technical solutions provided in this application include: The link status can be determined by the communication link status parameters, and the predicted confidence level of the link status can be calculated. Based on the link status and the predicted confidence level, the initial transmission priority of each data packet to be sent is calculated. The initial transmission priority of the data packet and the data packet generation timestamp are encapsulated in the header identifier of the data packet, and the data packet carrying the header identifier is sent to the edge node. This embodiment can proactively trigger priority upgrades before the link status deteriorates, and perform scheduling compensation for critical data. It significantly reduces the risk of loss and timeout of event data packets in dynamic vehicle networking environments, ensures reliable data transmission, and solves the technical problem in related technologies that important event data packets are easily lost or timed out, and cannot guarantee the reliable transmission of safety warning data. Attached Figure Description

[0018] Figure 1 This application provides a schematic diagram of a vehicle network data processing flow. Figure 2 This is a diagram of the 12-byte priority identifier encapsulation structure of this application; Figure 3 This is a schematic diagram of another vehicle network data processing procedure provided in an embodiment of this application; Figure 4 This is a flowchart illustrating the sorting engine workflow of an embodiment of this application. Figure 5 This is a flowchart of the event flow state machine processing in an embodiment of this application; Figure 6 This is a schematic diagram of the hardware structure of the vehicle network data processing device involved in the embodiments of this application. Detailed Implementation

[0019] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present application.

[0020] First, some of the technical terms used in this application will be explained to help those skilled in the art understand this application.

[0021] Link Quality Composite Index (Q): A multi-dimensional scalar index used to quantify the communication link status between vehicle terminals and edge nodes. This index is calculated by normalizing and weighting four heterogeneous physical parameters: network latency, data packet loss rate, signal strength, and bandwidth fluctuation. The value ranges from 0 to 1, with a higher value indicating better current link quality. This provides a unified quantitative input for subsequent link status prediction and priority scheduling.

[0022] Exponentially Weighted Moving Average (EWMA) Recursive Model: A recursive smoothing model used to extract long-term trends in link quality. This model introduces a forgetting factor λ to weight the current link quality index with historical smoothed values, effectively suppressing the interference of instantaneous noise on link status judgment, thus more stably reflecting the true trend of link quality changes.

[0023] Prediction confidence (C): A quantitative metric used to assess the reliability of current link quality prediction results.

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0025] In a first aspect, embodiments of this application provide a method for processing vehicle network data.

[0026] In one embodiment, reference is made to Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the vehicle-to-everything (V2X) data processing method of this application. Figure 1 As shown, the vehicle-to-everything (V2X) data processing methods include: S100: Determine the link state based on the communication link state parameters and calculate the prediction confidence of the link state.

[0027] S200: Calculate the initial transmission priority of each data packet to be sent based on link status and prediction confidence.

[0028] S300: Encapsulates the initial transmission priority of the data packet and the data packet generation timestamp into the header identifier of the data packet, and sends the data packet carrying the header identifier to the edge node.

[0029] In this embodiment, during step S100, the communication link status parameters can be collected in real time by monitoring modules deployed on the vehicle terminal (i.e., the vehicle end) and edge nodes. To reduce random errors caused by instantaneous jitter, each parameter uses the average of the sliding window of the most recent 10 sampling points as the current status value, and the system sampling interval is fixed at 20ms.

[0030] In step S200 above, the link status is based on the output of step S100. With prediction confidence Calculate the initial transmission priority for each data packet to be sent. This enables dynamic transmission scheduling that adapts to changes in link state.

[0031] In step S300 above, in the packet header of the application layer or custom adaptation layer, the system defines a new fixed-length field to store the priority identifier. This field has a total length of 12 bytes. See also Figure 2 As shown, the priority identifier in this embodiment includes two subfields: the first 4 bytes are the calculated dynamic priority value. It uses the IEEE 754 single-precision floating-point format for storage, with a value range of 0.1 to 1.3; the last 8 bytes are the timestamp when the data packet was generated. The data packet is stored as a 64-bit unsigned integer, with units of μs, recording the generation time of the data packet on the vehicle terminal's local clock. The two subfields are arranged consecutively in order of priority followed by the timestamp, forming a globally unique 12-byte identifier. This identifier is written into the custom extension area of ​​the protocol header after the application layer data packet is constructed and before it is passed to the lower-level protocol stack for low-level encapsulation. If an internationally standardized application layer protocol such as SAE J2735 is used, the system utilizes its reserved user-defined field bits to carry this identifier. If a custom adaptation layer is used, the system inserts a thin layer between the transport layer and the network layer. This thin layer only adds a 12-byte prefix as a priority identifier to the original data packet header; the rest of the encapsulation process is unaffected.

[0032] This embodiment determines the link status and calculates the predicted confidence level of the link status through communication link status parameters. Based on the link status and predicted confidence level, it calculates the initial transmission priority of each data packet to be sent, encapsulates the initial transmission priority of the data packet and the data packet generation timestamp into the header identifier of the data packet, and sends the data packet carrying the header identifier to the edge node. This embodiment can proactively trigger priority upgrades before the link status deteriorates, perform scheduling compensation for critical data, significantly reduce the risk of loss and timeout of event data packets in dynamic vehicle networking environments, ensure reliable data transmission, and solve the technical problem in related technologies that important event data packets are easily lost or timed out, making it impossible to guarantee the reliable transmission of safety warning data; it can also solve the problem of disordered timing of event data packets in dynamic network environments.

[0033] Further, in one embodiment, determining the link state based on communication link state parameters and calculating the prediction confidence of the link state includes: S101: Based on network latency Data packet loss rate Signal strength and bandwidth fluctuations Calculate the overall link quality index.

[0034] S102: Construct an online recursive prediction model with a forgetting factor, wherein the online recursive prediction model with a forgetting factor includes an exponentially weighted moving average recursive model and a prediction confidence function.

[0035] S103: Determine the link status based on the calculated comprehensive link quality index and the exponentially weighted moving average recursive model.

[0036] S104: Calculate the prediction confidence based on the calculated comprehensive link quality index and the prediction confidence function.

[0037] In this embodiment, the collected communication link status parameters include network latency. Data packet loss rate Signal strength and bandwidth fluctuations In step S101 above, in order to achieve a future timeframe of no more than... To predict link stability within a time window, the system first constructs a comprehensive link quality index. It is used to uniformly quantize multidimensional communication states, and is defined as follows:

[0038] in, , , , The preset weighting coefficients satisfy the following conditions: ,in, and These represent the maximum reference values ​​for signal strength and bandwidth, respectively. Through the above normalization process, the system can uniformly map heterogeneous physical parameters such as delay, packet loss rate, signal strength, and bandwidth into a single scalar index, and The value range is [0,1], with a larger value indicating better current link quality. This embodiment uses monitoring modules deployed at vehicle terminals and edge nodes to collect network latency data in real time. Packet loss rate Signal strength and bandwidth fluctuations The link quality comprehensive index is calculated using a normalized weighted fusion formula. This index maps heterogeneous physical parameters to a scalar in the range [0,1], providing a standardized and quantifiable input basis for subsequent link state prediction and priority scheduling.

[0039] In step S102 above, after obtaining the comprehensive link quality index... Subsequently, the system further constructs an online recursive prediction model with a forgetting factor. This model includes an exponentially weighted moving average recursive model and a prediction confidence function. The former is used to extract the long-term trend of link quality, and the latter is used to evaluate the reliability of the current prediction results. The specific definitions are as follows: 1) Exponentially weighted moving average recursive model, used to extract long-term trends in link quality:

[0040] in, For discrete-time indexing, Forgetting factor, used to control the rate of decay of historical data. This represents the smooth link quality value at the current moment.

[0041] 2) The prediction confidence function is used to determine whether the current prediction result is reliable. Its definition is:

[0042] in, This represents the variance of the link quality values ​​from the most recent 6 sampling points. This embodiment sets a prediction confidence threshold as the maximum allowable variance threshold. = 0.5, this threshold is the preferred median value within the normalized interval [0,1], used to classify the prediction confidence level. When the prediction results are reliable, an active priority enhancement strategy is implemented; when... When the prediction reliability is generally low, a conservative priority-increasing strategy is implemented; when If the prediction is invalid, the default stable link scheduling will be used.

[0043] Further, in one embodiment, determining the link status based on the calculated comprehensive link quality index and the exponentially weighted moving average recursive model includes: S1031: Calculate the smoothed link quality value at the current moment based on the comprehensive link quality index and the exponentially weighted moving average recursive model.

[0044] S1032: If the quality value of the smooth link is detected to have decreased by more than the set quality value in a cumulative manner at multiple consecutive sampling points, it is determined that the link has a deterioration trend and the link state is recorded as deterioration; otherwise, the link is considered to remain stable and the link state is recorded as stable.

[0045] In this embodiment, after calculating the comprehensive link quality index, the comprehensive link quality index is substituted into the above-mentioned exponentially weighted moving average recursive model to calculate the smoothed link quality value at the current time. The exponentially weighted moving average recursive model is mainly used to suppress the interference of instantaneous noise on link status judgment, thus reflecting the true trend of link quality changes more stably. For example, when the system detects... When the cumulative decrease of three consecutive sampling points exceeds 0.05, the link is considered to be deteriorating, and this is recorded as... = Deterioration, otherwise the link is considered to remain stable, denoted as =Stable.

[0046] To avoid amplifying prediction errors in high-speed moving scenarios, the system also considers the vehicle's speed. The prediction time window is dynamically adjusted. Under low-speed driving conditions, the prediction window can be maintained at [value missing]. Under high-speed movement conditions, the prediction window is automatically compressed to approximately [size missing]. This ensures that the extrapolation error of the prediction model remains within an acceptable range.

[0047] Furthermore, in one embodiment, the vehicle-to-everything (V2X) data processing method may further include: receiving feedback signals transmitted from edge nodes, and adjusting the parameters of the exponentially weighted moving average recursive model based on the feedback signals. In this embodiment, the feedback signals may include the timestamp error variance introduced by reordering. With packet out-of-order rate timestamp error variance With packet out-of-order rate Encapsulated into a feedback vector The feedback is sent to the vehicle end, and the aforementioned exponentially weighted moving average recursive model utilizes the feedback vector. Adjust its core parameters: for example, when When the value is greater than 0.1, the forgetting factor λ is reduced by a set value to enhance the model's response to rapidly changing links. <50 The forgetting factor λ is increased by a set value to smooth steady-state noise. This set value can be adjusted as needed, for example, it can be 0.08. This feedback loop ensures that the accuracy of macroscopic decisions is not contaminated by link prediction errors, and only time-series physical indicators drive the adaptive calibration of the prediction model.

[0048] Furthermore, in some embodiments, calculating the initial transmission priority of each data packet to be sent based on link state and predicted confidence may include: S201: Calculate the original priority of each data packet to be sent based on the criticality coefficient and the priority increment driven by the link state; wherein, the criticality coefficient is predefined according to the event type.

[0049] S202: Use a low-pass filter to smooth the original priority and obtain the initial transmission priority of each data packet to be sent.

[0050] In this embodiment, in step S201 above, the system first predefines a criticality coefficient based on the event type. In this embodiment, corresponding security warning events can be set. Traffic efficiency events correspond =0.6, corresponding to information service events =0.3. In other embodiments, other values ​​can be set as needed; the hierarchical mechanism is used to reflect the differences in importance of different business data, so that security-related data has a higher scheduling weight.

[0051] Define the original priority Calculation method: .

[0052] in, The priority increment is driven by the link state, and its value is dynamically determined based on the output of step S100. When the link state... The condition is judged to be deteriorating, and the prediction confidence level meets the following criteria. When this occurs, it indicates that the system can reliably predict that the link is about to deteriorate. At this point, the following settings should be applied: =0.2, thereby prioritizing the transmission of critical data; when the link status remains stable, set =0 indicates no additional priority boost; when the link state is determined to be deteriorating, but the prediction confidence meets the following: When this occurs, it indicates that the prediction result has significant uncertainty. In this case, the system adopts a conservative boosting strategy, setting only the desired boost level. =0.1 to avoid excessive preemption of network resources due to misjudgment.

[0053] Meanwhile, to avoid the scheduling stability being affected by priority values ​​that are too large or too small, the system further optimizes... The range of values ​​is truncated and restricted: The lower bound of the truncation limit is used to ensure that non-urgent services are not completely blocked, while the upper bound allows security-related events to obtain scheduling rights beyond the highest level of normal services. When the calculated original priority is greater than 1.3, it is taken as 1.3; when the calculated original priority is less than 0.1, it is taken as 0.1.

[0054] In step S202, to avoid network scheduling oscillations caused by drastic priority changes due to instantaneous fluctuations in link status, the system introduces a first-order low-pass filter mechanism to smooth out the priority change process: After obtaining the original priority, the system further uses a first-order low-pass filter to generate a smoothed initial transmission priority. Its recursive formula is: .

[0055] in, Indicates the sequence number of data packets generated under the same event source. These are the filter coefficients. The initial transmission priority of the preceding data packet is used as the basis for filtering. The main function of the filtering mechanism is to suppress drastic priority jumps between consecutive data packets. For example, when the link experiences a momentary disturbance... When the price drops rapidly from a high level, the filtered result is as follows: It will only decrease slowly, thus avoiding abrupt changes in priority that could cause subsequent schedulers to misjudge the network status.

[0056] Preferably, to ensure scheduling consistency within the same event stream, the system employs a state machine mechanism to control the priority inheritance of consecutive data packets. Once the first data packet of a certain event source is completed... If the priority value is calculated, subsequent data packets of the same type will directly inherit this priority value until either of the following conditions is met, at which point it will be recalculated: ① the current event ends; ② the link state remains stable for more than [time period missing]. This constraint mechanism effectively prevents frequent priority jittering within the same event stream. For data streams of different event types, the system... The numerical value determines the cross-stream scheduling. Larger data packets are placed earlier in the queue of edge nodes, thus gaining higher transmission priority.

[0057] Further, preferably, the vehicle network data processing method further includes: if the calculated initial transmission priorities of multiple data packets to be sent are the same, then a fine-tuning factor is added to the initial transmission priorities of the multiple data packets to be sent, and the fine-tuning factor is different for different data packets to be sent. This embodiment addresses the issue of multiple data packets having the same initial transmission priority. To address the sorting conflict issue during numerical calculations, the system further introduces a fine-tuning factor. This fine-tuning term is appended to the end of the priority value. It is only used to break the tie in ordering when all priorities are exactly the same. It does not change the actual scheduling weight, but it can ensure that later arriving data packets will not affect the overall scheduling stability due to the random ordering mechanism. Different data packets to be sent can have different fine-tuning factors.

[0058] After step S200 is completed, each data packet carries the corresponding initial transmission priority. and the timestamp generated by the data packet They proceed together to step S300 for subsequent packet header encapsulation and scheduling processing.

[0059] In step S300 above, in the data packet header of the application layer or custom adaptation layer, the system defines a new fixed-length field to store the priority identifier. This field has a total length of 12 bytes. The priority identifier includes two subfields; the first 4 bytes are the dynamic priority value calculated in step S200. It uses the IEEE 754 single-precision floating-point format for storage, with a value range of 0.1 to 1.3, which can accurately express the priority after filtering and retain the fine-tuning factor. The impact; the last 8 bytes are the timestamp when the data packet was generated. The data packet, stored as a 64-bit unsigned integer in μs, records the generation time of the data packet on the vehicle terminal's local clock. The two subfields are arranged consecutively in order of priority (first, then timestamp), forming a globally unique 12-byte identifier. This identifier is written into the custom extension area of ​​the protocol header after the application layer data packet is constructed and before being passed to the lower-level protocol stack for low-level encapsulation. If an internationally standardized application layer protocol such as SAE J2735 is used, the system utilizes its reserved user-defined field bits to carry this identifier; if a custom adaptation layer is used, the system inserts a thin layer between the transport and network layers. This thin layer only adds a 12-byte prefix as a priority identifier to the original data packet header; the rest of the encapsulation process remains unaffected.

[0060] After receiving a data packet, the edge node first parses the header of the application layer or adaptation layer to extract the 12-byte identifier and restores it to its original form. and Two independent values ​​are then used to encapsulate the data packet, and the original packet content is then delivered to the upper-layer application. The underlying physical layer and MAC layer protocol headers, such as IEEE 802.11p or C-V2X, are not modified in any way, fully adhering to the standard definitions to ensure interoperability with existing communication equipment. The presence of this 12-byte identifier will not trigger abnormal behavior in the underlying retransmission mechanism or congestion control algorithm because the underlying protocol treats the entire application layer and its header as a transparent payload. After encapsulation is completed in step S300, the data packet, carrying its digital identity, enters the network for transmission.

[0061] This embodiment uses monitoring modules deployed at vehicle terminals and edge nodes to collect network latency, packet loss rate, signal strength, and bandwidth fluctuations in real time, and constructs a comprehensive link quality index. The online recursive prediction model with a forgetting factor can proactively trigger priority escalation before the link state deteriorates. Compared to existing passive retransmission or fixed priority mechanisms, this method schedules and compensates for critical data approximately 80-200ms in advance, significantly reducing the risk of event data packet loss and timeout in dynamic vehicular networking environments.

[0062] To address the issue that instantaneous fluctuations in the link can easily lead to drastic priority jumps, this embodiment introduces a first-order low-pass filtering mechanism in the priority calculation stage, and bases it on the prediction confidence level. Priority increment It is divided into three levels: high (0.2), medium (0.1), and zero. This design avoids excessive network resource contention and scheduling jitter caused by misjudgment or noise, and allows the transmission priority of various events such as security warnings, traffic efficiency, and information services to be smoothly adjusted, maintaining stable output of the system under different link states.

[0063] This embodiment defines a 12-byte fixed-length priority identifier (including dynamic priority) in the application layer or adaptation layer. With the generation timestamp (and based on latency jitter statistics) Dynamically calculate the sorting window length The sorting engine performs a two-level comparison based on priority and timestamp, combined with a timeout forced output mechanism. This effectively solves the problem of disordered timing of event data packets without making any modifications to the underlying communication protocol (such as IEEE 802.11p / C-V2X).

[0064] Secondly, this application also provides another vehicle-to-everything (V2X) data processing method, applied to edge nodes, see [link to relevant documentation]. Figure 3 As shown, the vehicle-to-everything (V2X) data processing method includes: S400: Receives data packets carrying header identifiers sent by the vehicle end. The header identifier of the data packet contains the initial transmission priority of the data packet and the data packet generation timestamp.

[0065] S500: Based on the initial transmission priority of the data packet and the data packet generation timestamp, the data packets are reordered in time sequence and the data stream is output according to the logical time sequence.

[0066] S600: Performs anomaly detection and real-time aggregation on the data stream and outputs the final processing result.

[0067] In this embodiment, the edge node sets a sorting window based on the maximum expected latency jitter statistics. The length of this window is not a fixed value but is driven by the real-time network status. In step S500 above, the edge node continuously calculates the difference between the timestamp of the received data packet and the local arrival time of the edge node, defining latency jitter:

[0068] In the formula, Indicates the first The latency jitter value corresponding to each data packet; Represented as the first The local clock time when each data packet arrives at the edge node; Indicates the first The time when each data packet arrives at the local clock of the edge node. Generate a timestamp for the data packet embedded in step S300. The first embedded step S300 Each data packet generates a timestamp. This formula measures the deviation between the sending and receiving intervals of adjacent data packets, and is used when network path differences or node processing delays cause packets to arrive later than expected. Significantly positive values ​​will appear. See also Figure 4 As shown, to determine the sorting window length, edge nodes maintain the jitter history of the most recent N=50 data packets, and define the maximum expected latency jitter statistic. ,in This represents the average of historical fluctuations. This represents the standard deviation of historical jitter. The formula, based on Chebyshev's inequality, ensures that approximately 99% of jitter samples fall within the range of... Within the range, sorting window length Pick The window is set to the larger of the fixed lower limit of 20ms and the absolute upper limit of 500ms to prevent it from growing indefinitely. The window's dynamic update cycle is recalculated every 20 data packets received. .

[0069] Complete the sorting window length After dynamic calculation and periodic updates, edge nodes can initiate the sorting engine to perform packet time-order rearrangement based on the network state-driven window parameters mentioned above. The sorting engine uses priority and timestamps as its core criteria, combined with a dynamic window waiting mechanism, to correct out-of-order packets and output them in an ordered manner. The specific operation process is as follows: Edge nodes maintain a queue of packets to be ordered for each packet. When a packet arrives, its 12-byte priority identifier is parsed to reconstruct the packet. and After a data packet enters the queue, it does not immediately participate in the sorting process; instead, a sorting process is initiated. A timer of length is used. While the timer is active, newly arriving data packets are continuously inserted into the queue, and a sorting operation is triggered periodically every 5ms or whenever a new data packet arrives. The sorting operation employs a two-level comparison rule: the first level of comparison... The values ​​are arranged in order of magnitude, with larger values ​​placed at the front of the queue; the second level is... Compare when they are equal The timestamp with the smaller value, indicating an earlier generation time, is ranked higher. This rule is consistent with the fine-tuning factor in step S200. The design forms a closed loop. The fine-tuning factor carried in the queue ensures that two different event sources can be distinguished even if their priority calculations happen to be equal. When a data packet has been waiting in the queue for more than [a certain amount of time]... Subtracting the already waited time, the data packet is marked as timed out and forcibly output, no longer waiting for potentially arriving higher-priority packets. Simultaneously, after each packet is output, the wait time of the current queue header packet is checked; if it has exceeded... If the value is 0.9 times the original value, the queue output will be triggered in advance to reduce the risk of blocking at the head of the queue. After the data stream is reordered in step S500, it will be output strictly according to the logical timing and passed to the distributed computing engine in step S600 for intelligent processing.

[0070] Furthermore, in one embodiment, in step S600, the data packets in the data stream can be grouped according to event type, and data packets from the same vehicle end and belonging to the same event type can be defined as the same event stream; when the first data packet of an event stream is received, computing resources are allocated for the event and the state container is initialized, and the aggregation result in the state container is updated when subsequent data packets belonging to the same event stream arrive; if no data packets of the same event stream are received for a continuous set time, the event is determined to be over and the state container is released, and the final processing result is output.

[0071] In this embodiment, step S600 strictly limits the processing paradigm of edge nodes to an event-driven mode, rather than the traditional periodic or batch processing mode. Each data packet in the reordered data stream output in step S500 carries the timestamp embedded in step S300. Priority calculated in step S200 These data packets are grouped according to event type. Consecutive data packets from the same vehicle terminal and belonging to the same event type, such as sudden braking or lane change warnings, constitute a complete event stream. See also Figure 5 As shown, the distributed computing engine employs an event-driven triggering mechanism: the system allocates computing resources and initializes a state container only when the first data packet is received for an event stream; subsequent data packets belonging to the same event directly update the aggregation result within the container; when no new data packets for the same event are received for 200ms consecutively, the system determines the event has ended, releases the container, and outputs the final processing result. This event-driven mechanism avoids the idle overhead of periodic polling, ensuring that computing resources are consumed only when an event actually occurs.

[0072] In step S600, the specific logic of anomaly detection and real-time aggregation processing is bound to the criticality coefficient K of the event type. For safety warning events, K=0.9, the aggregation processing performs multi-source data fusion: edge nodes collect data from different vehicle terminals within a 50-millisecond window. Data packets of the same type with a difference of less than 30ms are... Calculate the weighted average confidence level for the weights:

[0073] in For the first Each data packet contains the raw sensor confidence value. Anomaly detection compares the confidence scores of consecutive data packets within the same event stream. Interval and step S500 statistics If the interval exceeds 3 These are then marked as time-series abnormal events, and the time of occurrence of the abnormality is recorded. For traffic efficiency events (K=0.6) and service events (K=0.3), the aggregation processing adopts counting and statistics and data caching and forwarding strategies, respectively, with a computational load far lower than that of security events.

[0074] Furthermore, in one embodiment, the vehicle-to-everything (V2X) data processing method may further include: S700: Calculates the timestamp error variance introduced by reordering based on the packet generation timestamp, the time when the edge node outputs the packet, and the estimated average transmission delay from packet generation to edge node output.

[0075] S800: Calculates the packet out-of-order rate based on the number of packets that were out of order before entering the sorting window and the total number of packets in the window.

[0076] S900: Encapsulates the timestamp error variance and data packet out-of-order rate introduced by reordering into a feedback signal and feeds it back to the vehicle.

[0077] In this embodiment, after anomaly detection and real-time aggregation processing are completed, the system extracts two feedback signals directly related to step S100. The first feedback signal is the timestamp error variance introduced by reordering:

[0078] in The variance of the reordering timestamp error is used to quantify the fidelity of the reordered data stream in restoring the original time sequence. The smaller the variance, the more accurate the time sequence restoration. This represents the total number of data packets output in the current batch. Output the first edge node The moment of each data packet For the first Each data packet generates a timestamp locally on the vehicle terminal. The variance is an estimate of the average transmission delay from generation to output, and it quantifies the fidelity of the reordering process in restoring the original temporal order.

[0079] The second feedback signal is the out-of-order rate. , This represents the number of data packets that were out of order before entering the sorting window. The total number of data packets within the window is used. The out-of-order rate is calculated without relying on the reordering results, but rather by directly determining the data packet pairs with decreasing timestamps in the original arrival sequence. and Encapsulated into a feedback vector After processing 200 event packets or averaging every few seconds, the data is returned to the prediction model in step S100.

[0080] The above-mentioned vehicle-to-everything (V2X) data processing method is applicable to C... Vehicle-to-everything (V2X) or DSRC communication protocols can be deployed between edge computing nodes and on-board units (OBUs / vehicle terminals). They can complete data acquisition, transmission scheduling, timing reordering, and event-driven intelligent processing while complying with existing communication standards, and have good protocol compatibility and scenario applicability.

[0081] This embodiment strictly limits the processing paradigm of edge nodes to event-driven, allocating computing resources and state containers only for the event of receiving the first data packet, and releasing resources and outputting results after 200ms without new packets. Simultaneously, for security alert events (K=0.9), multi-source data fusion and weighted average confidence calculation are performed within a 50ms window, based on timestamp intervals and... The relationship marking timing anomalies are addressed. Compared to periodic batch processing, this method significantly reduces invalid polling and concentrates the computational load on high-criticality events.

[0082] Furthermore, this embodiment incorporates the timestamp error variance introduced by reordering. With disorder rate Encapsulated as a feedback vector And it is fed back to the prediction model in step S100. When When the value is >0.1, the forgetting factor is reduced to 0.92 to enhance the response to rapidly changing links. <50 The forgetting factor is increased to 0.98 to smooth steady-state noise. This closed-loop mechanism relies solely on time-series physical indicators to drive adaptive model calibration, avoiding the contamination of macroscopic decisions by link prediction errors and enabling the system to maintain stable performance under different vehicle speeds and network environments.

[0083] Thirdly, embodiments of this application also provide a vehicle network data processing device.

[0084] In one embodiment, the vehicle-to-everything (V2X) data processing device includes: a first calculation module, which is used to determine the link status based on communication link status parameters and calculate the predicted confidence of the link status; a second calculation module, which is used to calculate the initial transmission priority of each data packet to be sent based on the link status and the predicted confidence; and an encapsulation module, which is used to encapsulate the initial transmission priority of the data packet and the data packet generation timestamp into the header identifier of the data packet, and send the data packet carrying the header identifier to the edge node.

[0085] Furthermore, in one embodiment, another vehicle-to-everything (V2X) data processing device may be provided, comprising: a receiving module for receiving data packets carrying header identifiers sent by a vehicle, wherein the header identifier of the data packets includes the initial transmission priority of the data packets and the data packet generation timestamp; a reordering module for reordering the data packets in time based on the initial transmission priority of the data packets and the data packet generation timestamp, and outputting the data stream according to the logical time sequence; and a processing module for performing anomaly detection and real-time aggregation processing on the data stream and outputting the final processing result.

[0086] The functions of each module in the above-mentioned vehicle network data processing device correspond to the steps in the above-mentioned vehicle network data processing method embodiment, and their functions and implementation processes will not be described in detail here.

[0087] Fourthly, embodiments of this application provide a vehicle-to-everything (V2X) data processing device, which may be a personal computer (PC), laptop computer, server, or other device with data processing capabilities.

[0088] Reference Figure 6 , Figure 6 This is a schematic diagram of the hardware structure of the vehicle-to-everything (V2X) data processing device involved in the embodiments of this application. In the embodiments of this application, the V2X data processing device may include a processor, a memory, a communication interface, and a communication bus.

[0089] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.

[0090] Communication interfaces include input / output (I / O) interfaces, physical interfaces, and logical interfaces used for interconnecting devices within the vehicle-to-everything (V2X) data processing equipment, as well as interfaces used for interconnecting the V2X data processing equipment with other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0091] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0092] The processor can be a general-purpose processor, which can call the vehicle-to-everything (V2X) data processing program stored in the memory and execute the V2X data processing method provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the V2X data processing program is called can be referred to in the various embodiments of the V2X data processing method of this application, and will not be repeated here.

[0093] Those skilled in the art will understand that Figure 6 The hardware structure shown does not constitute a limitation of this application and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0094] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0095] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0096] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0097] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0098] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0099] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0100] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A vehicle-to-everything (V2X) data processing method, applied to the vehicle end, characterized in that, The vehicle-to-everything (V2X) data processing method includes: The link state is determined based on the communication link state parameters, and the prediction confidence of the link state is calculated. The initial transmission priority of each data packet to be sent is calculated based on the link state and the predicted confidence level. The initial sending priority of the data packet and the data packet generation timestamp are encapsulated in the header identifier of the data packet, and the data packet carrying the header identifier is sent to the edge node.

2. The vehicle network data processing method as described in claim 1, characterized in that, The process of determining the link state based on communication link state parameters and calculating the prediction confidence of the link state includes: Based on network latency Data packet loss rate Signal strength and bandwidth fluctuations Calculate the overall link quality index; An online recursive prediction model with a forgetting factor is constructed, which includes an exponentially weighted moving average recursive model and a prediction confidence function. The link status is determined based on the calculated comprehensive link quality index and the exponentially weighted moving average recursive model. The prediction confidence is calculated based on the calculated comprehensive link quality index and the prediction confidence function.

3. The vehicle network data processing method as described in claim 2, characterized in that, The vehicle-to-everything (V2X) data processing method further includes: It receives feedback signals from edge nodes and adjusts the parameters of the exponentially weighted moving average recursive model based on these feedback signals.

4. The vehicle network data processing method as described in claim 2, characterized in that, The determination of link status based on the calculated comprehensive link quality index and the exponentially weighted moving average recursive model includes: The smoothed link quality value at the current moment is calculated based on the comprehensive link quality index and the exponentially weighted moving average recursive model. If the quality value of the smooth link is detected to have decreased by more than the set quality value in a cumulative manner across multiple consecutive sampling points, it is determined that the link is deteriorating and the link status is recorded as deteriorating; otherwise, the link is considered to be in a stable state and the link status is recorded as stable.

5. The vehicle network data processing method as described in claim 1, characterized in that, The calculation of the initial transmission priority for each data packet to be sent based on link state and predicted confidence includes: The original priority of each data packet to be sent is calculated based on the criticality coefficient and the priority increment driven by the link state; whereby the criticality coefficient is predefined according to the event type. The original priority is smoothed by using a low-pass filter to obtain the initial transmission priority of each data packet to be sent.

6. A vehicle-to-everything (V2X) data processing method, applied to edge nodes, characterized in that, It includes: Receive data packets with header identifiers sent by the vehicle end. The header identifiers of the data packets contain the initial transmission priority of the data packets and the data packet generation timestamp. Data packets are reordered based on their initial sending priority and generation timestamp, and the data stream is output according to the logical timing. Perform anomaly detection and real-time aggregation on the data stream and output the final processing result.

7. The vehicle network data processing method as described in claim 6, characterized in that, The vehicle-to-everything (V2X) data processing method further includes: Data packets in the data stream are grouped according to event type, and data packets from the same vehicle end and belonging to the same event type are defined as the same event stream; When the first data packet of an event stream is received, computing resources are allocated for the event and the state container is initialized. When subsequent data packets belonging to the same event stream arrive, the aggregation result in the state container is updated. If no data packets of the same event stream are received for a set period of time, the event is considered to have ended and the state container is released, and the final processing result is output.

8. The vehicle network data processing method as described in claim 6, characterized in that, The vehicle-to-everything (V2X) data processing method further includes: The timestamp error variance introduced by reordering is calculated based on the packet generation timestamp, the time when the edge node outputs the packet, and the estimated average transmission delay of the packet from generation to output by the edge node. The packet out-of-order rate is calculated based on the number of packets that were out of order before entering the sorting window and the total number of packets in the window. The timestamp error variance and data packet out-of-order rate introduced by reordering are encapsulated as feedback signals and fed back to the vehicle.

9. A vehicle-to-everything (V2X) data processing device, characterized in that, The vehicle network data processing device includes: The first calculation module is used to determine the link state based on the communication link state parameters and calculate the prediction confidence of the link state. The second calculation module is used to calculate the initial transmission priority of each data packet to be sent based on the link status and prediction confidence. The encapsulation module is used to encapsulate the initial sending priority of the data packet and the data packet generation timestamp into the header identifier of the data packet, and send the data packet carrying the header identifier to the edge node.

10. A vehicle-to-everything (V2X) data processing device, characterized in that, The vehicle-to-everything (V2X) data processing device includes a processor, a memory, and a V2X data processing program stored in the memory and executable by the processor, wherein when the V2X data processing program is executed by the processor, it implements the steps of the V2X data processing method as described in any one of claims 1 to 5 or claims 6 to 8.

Images