A vehicle-to-everything (V2X) big data real-time monitoring and user service optimization platform for intelligent transportation systems

Abstract

This invention belongs to the field of intelligent transportation vehicle networking technology and discloses a real-time monitoring and user service optimization platform for vehicle networking big data for intelligent transportation systems. The platform uses a preprocessing unit to denoise, unify the format, and extract core semantic features from the collected data in real time. It leverages a federated learning framework to achieve encrypted sharing of semantic features across nodes, and combines this with a traffic scenario knowledge graph to complete semantic consistency verification and eliminate conflicting and abnormal data. The transmission unit identifies the causal correlation of data based on 5G network slicing, with highly correlated data transmitted through low-latency slices to ensure data consistency. The data fusion unit can dynamically adjust weights according to traffic event types, such as increasing the weight of LiDAR data in accident scenarios, and combining this with high-precision timestamps and dual-mode positioning to calibrate spatiotemporal offsets, ultimately achieving accurate cross-modal data fusion. The causal inference unit deploys a lightweight causal Markov model on the roadside edge terminal to quickly locate the core causes of traffic anomalies.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation vehicle networking technology, specifically a vehicle networking big data real-time monitoring and user service optimization platform for intelligent transportation systems. Background Technology

[0002] Currently, the integration of vehicle-to-everything (V2X) and intelligent transportation has become an important direction for the development of smart transportation. Through the collaborative interaction of vehicles, roads, people, and the cloud, it improves the efficiency and safety of road networks. Surface technologies such as basic data collection and short-range vehicle-to-infrastructure communication have been implemented in many places. With the deep integration of V2X and intelligent transportation, existing technologies have initially solved surface problems such as data collection, simple transmission, and basic service delivery. However, at the core technology level for complex traffic scenarios, the following technical problems still exist: First, there is insufficient semantic alignment and dynamic association of multi-source heterogeneous data. Data from vehicle, roadside, user, and cloud belong to different dimensions and have obvious deviations in meaning. Existing technologies only standardize data formats but do not solve the deep semantic association of different types of data. This results in the inability to accurately reflect the essence of traffic scenarios after data fusion, leading to monitoring misjudgments and service recommendation disconnects. Secondly, the lack of causal reasoning and insufficient dynamic adaptation in service optimization means that existing service optimization relies on statistical correlation and does not consider the causal logic in traffic scenarios, resulting in weak anti-interference ability of service strategies. At the same time, most service parameters are statically configured and cannot be dynamically adjusted according to sudden traffic events or sudden changes in user behavior, making it difficult to meet the core requirements of real-time response and accurate adaptation of intelligent transportation systems. Summary of the Invention

[0003] The purpose of this invention is to provide a real-time monitoring and user service optimization platform for vehicle-to-everything (V2X) big data for intelligent transportation systems, in order to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a real-time monitoring and user service optimization platform for vehicle-to-everything (V2X) big data in intelligent transportation systems, the platform comprising: Preprocessing unit: preprocesses the collected data, extracts semantic features, and removes abnormal data that conflict with the semantics of the scene; Transmission unit: Identifies the causal relationship between preprocessed data and sets transmission priorities; employs dual encryption technology and dynamically adjusts the transmission rate; Data fusion unit: Adjusts data fusion weights according to traffic event type, corrects data spatiotemporal offset, monitors fusion effect in real time, and adjusts fusion strategy accordingly; Causal inference unit: Based on the fused data, identify the core causes of abnormal events, transmit the inference results back to the regional cloud, calculate the change in causal entropy, establish a scenario-based causal template library and update it regularly; Causal monitoring unit: Maps causal relationships and fused data to a twin scenario, showing the propagation path and impact of the causal chain; predicts road network traffic and vehicle trajectories and generates execution instructions, uncovering the causal relationship between traffic events and user needs; Service generation unit: Builds user profiles, generates optimal travel plans for users, and updates recommendation strategies in real time; Optimization Unit: Collects feedback data, identifies causal relationships and optimizes parameters, while quantitatively evaluating the optimization effect.

[0005] Preferably, the preprocessing unit is specifically as follows: Through the roadside terminal, the data collected by vehicle-mounted sensors, roadside radar and video detectors are denoised and formatted in real time, and the core semantic features of the target are extracted simultaneously. A cross-node semantic feature encryption and sharing mechanism is built using a horizontal federated learning framework, where each participating node only uploads local model gradient parameters; at the same time, a semantic consistency verification module is embedded, which, combined with a traffic scene knowledge graph containing traffic scene semantic rules, removes abnormal data that conflicts with scene semantics.

[0006] Preferably, the transmission unit is specifically as follows: Based on the data and semantic annotations preprocessed by the preprocessing unit, a causal correlation priority scheduling mechanism is designed based on 5G network slicing technology. The edge nodes identify the causal correlation between the preprocessed data in real time, and bundle and transmit highly correlated data through low-latency slices, while non-correlated data is transmitted in batches. A dual encryption scheme of semantic tags and data ontology is adopted, and only authorized nodes can parse semantic tags; at the same time, a vehicle-to-infrastructure communication latency parameter calibration module is added to dynamically adjust the transmission rate to adapt to the communication needs of different traffic scenarios.

[0007] Preferably, the data fusion unit is specifically as follows: After the transmission unit transmits highly correlated semantic data to the target node, the data fusion unit establishes a dynamic weighted fusion model and automatically adjusts the data fusion weights based on the type of traffic event. In accident scenarios, the weight of LiDAR data is increased while the weights of video and millimeter-wave radar data are decreased. A cross-modal data alignment module is deployed to perform calibration based on timestamps and spatial coordinates of GPS and BeiDou dual-mode positioning, correcting the spatiotemporal offset between video and radar data; a data fusion quality assessment module is integrated to monitor the fusion effect in real time and dynamically adjust the fusion strategy through indicators including semantic consistency and key field missing rate; ultimately achieving spatiotemporal alignment of cross-modal data.

[0008] Preferably, the causal inference unit is specifically as follows: Based on the fused data from the data fusion unit, a lightweight causal Markov model is deployed at the roadside edge terminal to compress the parameter scale, thereby adapting to the computing power of the ARM Cortex-A72 architecture, realizing the causal localization of traffic anomalies, and identifying the core causes of abnormal events. The design incorporates a cloud-edge iterative optimization mechanism. The edge terminal transmits the inference results, including event type and cause probability, back to the regional cloud. The cloud then updates the causal network topology by combining the full fusion data and calculates the change in causal entropy using KL divergence. When the difference exceeds a preset threshold, the edge model parameters are corrected, and the optimized model parameters are fed back to the edge terminal to infer the causal relationship between traffic events and user needs. Establish a scenario-based cause-effect template library, pre-store scenario templates, and regularly update it incrementally based on newly added fused data and event data.

[0009] Preferably, the causal monitoring unit is specifically as follows: Based on the causal relationship and the fused data of the data fusion unit, a digital twin road network mapping module using BIM and GIS fusion modeling is established to map the spatiotemporal causal relationship and fused data to the twin scene in real time, and to display the propagation path and degree of influence of the causal chain through a three-dimensional visualization interface. Design a dual-scale prediction engine for macro-level road network traffic and micro-level vehicle trajectory. The macro-level layer uses a GraphCNN model to predict road network traffic, while the micro-level layer uses an LSTM model to predict vehicle trajectories. The two interact in real time based on causal inference results. Macro-level results constrain the micro-level prediction range, and micro-level results correct the macro-level model parameters. It integrates service-oriented output interfaces to directly generate executable commands, connecting to traffic control systems and subsequent user service platforms.

[0010] Preferably, the service generation unit is specifically as follows: Based on the causal relationship between traffic events and user needs inferred by the causal inference unit and the real-time traffic status of the causal monitoring unit, a two-dimensional user profile with semantic and causal characteristics is established. The semantic profile is labeled with semantic tags of core needs, and the causal profile records the causal relationship between user behavior and service experience. The profile data is also synchronously linked to the real-time traffic status monitored by the digital twin. A causal recommendation algorithm is used, which combines fused data, causal inference conclusions and real-time traffic conditions to generate the optimal travel plan for users; at the same time, a service strategy dynamic adjustment module is deployed to update the recommendation strategy in real time based on real-time causal factors and traffic condition changes monitored by twins to ensure service adaptability.

[0011] Preferably, the optimization unit is specifically as follows: After receiving the service plan from the service generation unit and putting it into application, the optimization unit collects the full-link feedback data of the service plan, establishes an optimization process of feedback collection, causal analysis, parameter optimization, and service iteration, deploys a causal analysis module to analyze and identify the causal relationships in the feedback data, deploys a causal adaptive optimization algorithm to dynamically adjust the traffic monitoring model weights, data update frequency, and service recommendation parameters based on the identified causal relationships, and pushes the optimization parameters synchronously to the front-end data fusion unit, causal inference unit, and service generation unit. Establish an optimization effect evaluation module to quantitatively evaluate the optimization effect through indicators including service satisfaction, monitoring accuracy, and response timeliness.

[0012] The beneficial effects of this invention are as follows: 1. This invention uses a preprocessing unit to denoise, unify the format, and extract core semantic features from the collected data in real time. It relies on a horizontal federated learning framework to achieve encrypted sharing of semantic features across nodes, and then combines a traffic scenario knowledge graph to complete semantic consistency verification and eliminate conflicting and abnormal data. The transmission unit identifies the causal correlation of data based on 5G network slices. Highly correlated data is transmitted through low-latency slices to ensure data correlation. The data fusion unit can dynamically adjust the weights according to the type of traffic event. For example, in accident scenarios, the weight of LiDAR data is increased. Combined with high-precision timestamps and dual-mode positioning to calibrate spatiotemporal offset, it ultimately achieves accurate fusion of cross-modal data.

[0013] 2. The causal inference unit of this invention deploys a lightweight causal Markov model at the roadside edge terminal to quickly locate the core causes of traffic anomalies. Then, through cloud and edge iterative optimization mechanisms, it updates the causal network topology with full data, calculates the causal entropy change, and corrects the model parameters. The scenario-based causal template library can also be updated regularly based on new data. The causal monitoring unit uses a digital twin module that integrates BIM and GIS to visualize the causal chain propagation path. It is equipped with a dual-scale engine of GraphCNN macro traffic prediction and LSTM micro trajectory prediction. The two interact in real time to correct parameters, and the generated executable instructions are directly connected to the traffic control system, which greatly improves the accuracy of traffic anomaly early warning and trend prediction.

[0014] 3. The service generation unit of this invention constructs a semantic and causal dual-dimensional user profile based on the causal relationship between traffic events and user needs. It generates the optimal travel plan by combining a causal recommendation algorithm, and the service strategy can be dynamically adjusted according to real-time causal factors. The optimization unit collects feedback data from the entire service chain, identifies the causal relationship in the feedback through the causal analysis module, and then uses a causal adaptive optimization algorithm to adjust parameters such as monitoring model weights and data update frequency, and pushes them to the front-end unit simultaneously. At the same time, it relies on indicators such as service satisfaction and monitoring accuracy to quantitatively evaluate the effect and continuously improve service adaptability and user experience. Attached Figure Description

[0015] Figure 1 This is a flowchart of the vehicle-to-everything (V2X) big data real-time monitoring and user service optimization platform for intelligent transportation systems, as described in this invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] like Figure 1 As shown in the figure, this invention provides a vehicle-to-everything (V2X) big data real-time monitoring and user service optimization platform for intelligent transportation systems. The platform includes: The preprocessing unit, as the starting point of the platform's data flow, performs real-time noise reduction and format unification on the data collected by vehicle-side sensors, roadside radar, and video detectors through the roadside terminal, and simultaneously extracts core semantic features such as target location, speed, and type. The roadside terminal integrates an edge computing chip and a lightweight semantic processing module, which can process 16 heterogeneous data streams in parallel. The edge computing chip integrated in the roadside terminal uses an ARM architecture chip with a computing power of no less than 2 TOPS and power consumption controlled within 15W; the 16 heterogeneous data channels specifically include 4 channels of vehicle-side LiDAR data, 4 channels of roadside millimeter-wave radar data, 4 channels of roadside high-definition video data, 2 channels of vehicle-to-everything (V2X) communication data, and 2 channels of meteorological sensor data, ensuring that the data covers the core monitoring dimensions of traffic scenarios. A cross-node semantic feature encryption and sharing mechanism is built using a horizontal federated learning framework. Each participating node only uploads local model gradient parameters without leaking the original data. At the same time, a semantic consistency verification module is embedded. Combined with a traffic scene knowledge graph containing semantic rules of typical traffic scenarios, abnormal data that conflicts with scene semantics is eliminated, improving data efficiency and semantic feature extraction accuracy, and providing qualified semantic data for the transmission unit.

[0018] The transmission unit receives the preprocessed data and semantic annotations from the preprocessing unit. Based on 5G network slicing technology, a causal correlation priority scheduling mechanism is designed. Edge nodes identify the causal correlation between preprocessed data in real time, such as the correlation between accident data and traffic data of surrounding road sections. Highly correlated data is bundled and transmitted through low-latency slices, while non-correlated data is transmitted in batches. 5G network slicing is divided into low-latency slices and ordinary slices. Low-latency slices have an end-to-end latency of ≤15ms and an uplink bandwidth of ≥20Mbps, and are used to transmit highly correlated data. Causal correlation is calculated by mutual information value. When the mutual information value of two data modules is ≥0.6, it is determined to be highly correlated data. If the mutual information value is <0.6, it is classified as non-correlated data and is transmitted in batches using ordinary slices. A dual encryption scheme of semantic tags and data ontology is adopted, and only authorized nodes can parse semantic tags to avoid leakage of semantic information generated in the preprocessing stage; at the same time, a vehicle-to-infrastructure communication latency parameter calibration module is added to dynamically adjust the transmission rate to adapt to the communication needs of different traffic scenarios.

[0019] The data fusion unit is specifically as follows: After the transmission unit transmits highly correlated semantic data to the target node, the data fusion unit establishes a dynamic weight fusion model and automatically adjusts the data fusion weights based on the traffic event types (such as accidents, congestion, and severe weather) fed back by the causal inference unit. In the event scenario, the weight of LiDAR data is increased from 30% to 65%, while the weights of video and millimeter-wave radar data are reduced. A cross-modal data alignment module is deployed to perform calibration based on high-precision timestamps with an error of ≤1ms and spatial coordinates of GPS and BeiDou dual-mode positioning, correcting the spatiotemporal offset of video and radar data; an integrated data fusion quality assessment module is used to monitor the fusion effect in real time and dynamically adjust the fusion strategy through indicators including a cosine similarity threshold of 0.85 for semantic consistency and a key field missing rate of ≤0.5%; ultimately achieving spatiotemporal alignment of cross-modal data.

[0020] The semantic consistency cosine similarity calculation object is the semantic feature vector of the data source and the corresponding scene rule vector in the traffic scene knowledge graph. For example, in the accident scene, the obstacle position-velocity feature vector extracted by LiDAR data and the obstacle feature rule vector of the accident scene in the knowledge graph must have a similarity of ≥0.85. Key fields include vehicle ID, timestamp, spatial coordinates, and event type to ensure that no core information is missing.

[0021] Multi-source data dynamic weight fusion formula: ; In the formula: This represents the final output result after multi-source semantic data fusion; This indicates the total amount of multi-source data participating in the fusion, such as when fusing vehicle-mounted LiDAR data, roadside millimeter-wave radar data, and roadside video semantic data. =3; Indicates the first The dynamic fusion weights of each data source are related to the event type correction coefficient. The function has weight values ​​in the range [0, 1], and the sum of the weights of all data sources satisfies =1; This represents the correction factor for traffic incident types; different incident types correspond to different factors. Values, such as those in accident scenarios. =0.65, weight of LiDAR data source (0.65) = 0.65; in congested scenarios =0.4, Video semantic data source weight (0.4) = 0.4; Under normal circumstances =0.3, the weights of each data source tend to be balanced.

[0022] Indicates the first Semanticized data from a data source after preprocessing (denoising, semantic annotation, spatiotemporal calibration), for example This can be used to generate semantic feature vectors of obstacle position and velocity acquired by lidar. This can be a semantic feature vector of vehicle quantity-lane occupancy rate extracted from video images.

[0023] The causal inference unit is specifically as follows: Based on the fused data from the data fusion unit, a lightweight causal Markov model is deployed on the roadside edge terminal, compressing the parameter size to 500KB, thereby adapting to the computing power of the ARM Cortex-A72 architecture, realizing the causal localization of traffic anomalies, and quickly identifying the core causes of abnormal events, such as accidents causing congestion or traffic light malfunctions causing traffic chaos. The lightweight causal Markov model has a parameter size of 2MB before compression and is reduced to 500KB after compression. On the ARM Cortex-A72 architecture, the time for a single traffic anomaly causal inference is ≤80ms and the accuracy is ≥92%. The scenario-based causal template library is updated when the amount of newly fused data is ≥10GB and the number of newly added traffic events is ≥50. If the threshold is not reached in the current month, it will be postponed to the next month to ensure that the template library update is based on sufficient and effective data. The design incorporates a cloud-edge iterative optimization mechanism. The edge terminal transmits the inference results, including event type and cause probability, back to the regional cloud. The cloud then updates the causal network topology by combining the full fusion data and calculates the change in causal entropy using KL divergence. When the difference exceeds 0.1, the edge model parameters are corrected, and the optimized model parameters are fed back to the edge terminal to infer the causal relationship between traffic events and user needs. Establish a scenario-based causal template library, pre-store templates for typical scenarios such as rainstorms and holidays, and update it monthly based on newly added fused data and event data to improve the efficiency of scenario matching and inference, and improve the accuracy of causal relationship identification and the efficiency of causal inference for typical scenarios.

[0024] Formula for calculating causal entropy: ; In the formula: Represents conditional causal entropy, used to quantify the conditional causal variables given a known cause variable. Under the premise of outcome variables The lower the value, the greater the uncertainty. and The stronger the causal relationship, the better; when the causal inference results fed back by the edge terminal match the results calculated from the full data in the cloud... When the difference exceeds 0.1, edge model parameter correction is triggered; This represents the set of causal variables for traffic events, including all factors that could potentially trigger traffic abnormalities, such as traffic light malfunctions. 1) Vehicle accidents ( 2) Severe weather ( 3) Temporary traffic control ( 4) etc., each Each type corresponds to a specific trigger type, and all trigger types are mutually exclusive. This represents the set of outcome variables for a traffic event, containing all possible traffic state outcomes, such as road congestion. 1) Slow-moving traffic ( 2) Traffic disruption ( 3) Normal passage ( 4) etc., each Each state corresponds to a specific traffic condition, and the different condition types are mutually exclusive. Describing the cause variable With outcome variable The joint probability of simultaneous occurrence is calculated statistically from historical traffic event data accumulated in the cloud (such as accident-congestion correlation records and traffic light malfunction-traffic interruption correlation records from the past 12 months), for example, P( 2, 1) Represents the probability of a vehicle accident and road congestion occurring simultaneously; Indicating in the cause variable Under the premise that it occurs, the outcome variable The conditional probability of occurrence is calculated using the following formula: ,in Causal variable The marginal probability of an isolated occurrence, for example This represents the probability of traffic congestion occurring on a given road segment, assuming a vehicle accident has occurred.

[0025] The causal monitoring unit is specifically as follows: Based on the causal relationship and the fused data of the data fusion unit, a digital twin road network mapping module using BIM and GIS fusion modeling is established. The spatiotemporal causal relationship (such as the speed of road congestion transmission and the scope of event impact) and the fused data are mapped to the twin scene in real time. The causal chain propagation path and degree of impact are displayed through a three-dimensional visualization interface. The digital twin road network is constructed by fusing BIM models and GIS maps. The twin scene is updated every 2 seconds based on the fused data to ensure synchronization with the actual road network status. The GraphCNN model is trained with 300 iterations and a batch size of 64 to predict road network traffic in the next 30-120 minutes. The LSTM model is trained with 200 iterations and a batch size of 32 to predict vehicle trajectories in the next 10-60 seconds. Both models use the Adam optimizer to improve prediction stability and accuracy. A dual-scale prediction engine for macro-level road network traffic and micro-level vehicle trajectory is designed. The macro-level layer uses a GraphCNN model (inputting traffic data fused at 15-minute intervals) to predict road network-level traffic, while the micro-level layer uses an LSTM model (inputting location data at 1-second intervals) to predict vehicle trajectories. The two interact in real time based on causal inference results. The macro-level results constrain the micro-level prediction range, and the micro-level results correct the macro-level model parameters. It integrates service-oriented output interfaces to directly generate executable commands such as signal timing adjustment and charging pile scheduling, and connects to traffic control systems and subsequent user service platforms.

[0026] Formula for comprehensive evaluation of two-scale prediction error: ; ; ; In the formula: This represents the overall error of the two-scale traffic prediction, used to assess prediction accuracy. A smaller value indicates a better match between the prediction and actual traffic conditions. When the value is ≤0.05, the prediction result is considered to meet the accuracy requirements for traffic control and user services; , The weighting coefficients for macroscopic forecast error and microscopic forecast error are respectively, satisfying the following conditions: + =1, adjusted according to the actual application scenario, such as the control scenario of urban main roads. =0.6 (focusing on macro traffic flow), under the scenario of internal traffic within the park =0.7 (focusing on microscopic trajectories), in the default scenario = =0.5; This represents the macroscopic road network traffic prediction error, used to quantify the prediction deviation of the macroscopic layer in the dual-scale prediction engine of macroscopic road network traffic and microscopic vehicle trajectory. The unit is consistent with the traffic unit. This indicates the number of road segments in the macro-level road network traffic forecast. For example, when forecasting traffic flow for five key monitoring segments of a main road, =5, each road segment corresponds to a traffic forecast value and an actual value; Indicates the first The predicted traffic flow value for each road segment, that is, the traffic flow of the road segment in the next 15 minutes calculated by the GraphCNN model, in units of vehicles / 15 minutes; Indicates the first The actual traffic flow value for each road segment, that is, the real traffic flow of that road segment within the corresponding 15 minutes, collected by the roadside traffic flow detector, in units of... Consistent, used to compare with predicted values ​​to calculate errors; This represents the microscopic vehicle trajectory prediction error, used to quantify the prediction bias of the microscopic layer in the dual-scale prediction engine, and is expressed in meters. This represents the time step for predicting the microscopic vehicle trajectory. For example, when predicting the trajectory of a vehicle over the next 30 seconds... =30, each time step corresponds to a predicted value and an actual value of the trajectory coordinates; , These represent the predicted abscissa and ordinate of the vehicle at the t-th time step, respectively. They are the planar coordinates of the vehicle at that time step calculated by the LSTM model, in meters. The coordinate system is consistent with the spatial coordinates of the digital twin road network. , Let x and y represent the actual x and y coordinates of the vehicle at time step t, respectively. These are the vehicle's true planar coordinates at that time step, collected by the vehicle's GPS / BeiDou positioning module, and are expressed in units of x and y. , Consistent, used to compare with predicted coordinates to calculate trajectory deviation.

[0027] The service generation unit is specifically as follows: Based on the causal relationship between traffic events and user needs inferred by the causal inference unit and the real-time traffic status of the causal monitoring unit, a two-dimensional user profile with semantic and causal characteristics is established. The semantic profile is labeled with semantic tags for users' core needs such as commuting, medical treatment, and tourism. The causal profile records the causal relationship between user behavior and service experience, such as the relationship between route selection and congestion avoidance. The profile data is also synchronously linked to the real-time traffic status monitored by the digital twin. Semantic user profiles include basic dimensions, demand dimensions (commuting / medical / tourism / shopping demand tags, travel time preferences, destination types), and preference dimensions (preferred route types, whether to avoid construction zones). Causal profiles record the correlation data between users' choice of recommended routes and the rate of commuting time reduction and congestion avoidance success rates. For example, choosing recommended route A has an 80% probability of reducing commuting time by 15%. The causal recommendation algorithm calculates the causal correlation strength between route adjustments and improved travel efficiency (correlation strength ≥ 0.7 is prioritized for recommendation), and dynamically adjusts route recommendation weights based on real-time traffic conditions to ensure that the solution matches user needs. Employing a causal recommendation algorithm, which combines fused data, causal inference conclusions, and real-time traffic conditions, the system generates optimal travel options for users, including route planning, parking recommendations, and public transportation transfer suggestions. Simultaneously, a service strategy dynamic adjustment module is deployed to update the recommendation strategy in real time based on real-time causal factors such as sudden congestion and temporary traffic control, as well as changes in traffic conditions monitored by twin monitoring, to ensure service adaptability.

[0028] The optimization unit is specifically as follows: After receiving the service plan from the service generation unit and putting it into application, the optimization unit collects full-link feedback data of the service plan, including user ratings of the service, core causes of poor service performance recorded through semantic questionnaires, and data on the execution effect of control instructions from the causal monitoring unit and feedback data on the accuracy of inference from the causal inference unit. Based on these multi-dimensional feedbacks, the platform algorithm and service strategy are dynamically iterated to continuously improve the system's adaptability. The calculation method for the optimization effect evaluation indicators is as follows: Service satisfaction = (number of 4-5 point evaluations / total number of evaluations) × 100%; Monitoring accuracy = (number of times the monitoring results are consistent with the actual traffic conditions / total number of monitoring) × 100%; Response timeliness = the average time from data collection to service plan generation. By clarifying the calculation method, we can ensure that the optimization effect is quantifiable and comparable. Establish an optimization process for feedback collection, causal analysis, parameter optimization, and service iteration. Deploy a causal analysis module to analyze and identify causal relationships such as service delays and user dissatisfaction, and inference bias and monitoring errors. Deploy a causal adaptive optimization algorithm to dynamically adjust the traffic monitoring model weights, data update frequency, and service recommendation parameters based on the identified causal relationships, and synchronously push the optimized parameters to the front-end data fusion unit, causal inference unit, and service generation unit. Establish an optimization effect evaluation module to quantitatively evaluate the optimization effect through indicators including service satisfaction, monitoring accuracy, and response timeliness, so as to ensure that the iteration direction is correct and improve the service satisfaction and monitoring accuracy after optimization.

[0029] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0030] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A real-time monitoring and user service optimization system for vehicle-to-everything (V2X) big data in intelligent transportation systems, characterized in that: The system includes: Preprocessing unit: preprocesses the collected data and extracts semantic features, and removes abnormal data that conflicts with the semantics of the scene; Transmission unit: Identifies the causal relationship between preprocessed data, sets transmission priority, employs dual encryption technology, and dynamically adjusts the transmission rate; Data fusion unit: Adjusts data fusion weights according to traffic event type, corrects data spatiotemporal offset, monitors fusion effect in real time, and adjusts fusion strategy accordingly; Causal Inference Unit: Based on fused data, a lightweight causal model is deployed on the roadside edge terminal to achieve causal localization of traffic anomalies and identify the core causes of abnormal events; a cloud-edge iterative optimization mechanism is designed, whereby the edge terminal transmits the inference results back to the regional cloud, the cloud updates the causal network topology by combining the full fused data, calculates the change in causal entropy and corrects the edge model parameters, and feeds back the optimized model parameters to the edge terminal; a scenario-based causal template library is established, with pre-stored scenario templates and periodically updated based on new data increments; Causal Monitoring Unit: Based on causal relationships and fused data, a digital twin road network mapping module is established to map spatiotemporal causal relationships and fused data to the twin scene in real time, showing the propagation path and impact of the causal chain; a dual-scale prediction engine for macro-level road network traffic and micro-level vehicle trajectory is designed, and the two interact in real time based on the causal inference results; an integrated service output interface is provided to generate executable instructions and connect to the traffic control system and user service system; Service generation unit: Builds user profiles, generates optimal travel plans for users, and updates recommendation strategies in real time; Optimization Unit: Collects feedback data, identifies causal relationships and optimizes parameters, while quantitatively evaluating the optimization effect.

2. The vehicle-to-everything (V2X) big data real-time monitoring and user service optimization system for intelligent transportation systems according to claim 1, characterized in that: The preprocessing unit is specifically as follows: The data collected by vehicle-mounted sensors, roadside radar, and video detectors through roadside terminals are denoised and formatted in real time, and the core semantic features of the target are extracted simultaneously. A federated learning framework is used to build a cross-node semantic feature encryption and sharing mechanism, where each participating node only uploads local model gradient parameters; at the same time, a semantic consistency verification module is embedded, which, combined with a traffic scene knowledge graph containing traffic scene semantic rules, removes abnormal data that conflicts with scene semantics.

3. The vehicle-to-everything (V2X) big data real-time monitoring and user service optimization system for intelligent transportation systems according to claim 2, characterized in that: The transmission unit is specifically as follows: Based on the preprocessed data and semantic annotation, a causal association priority scheduling mechanism is designed based on network slicing technology. Edge nodes identify the causal association between data in real time, and highly correlated data is bundled and transmitted through low-latency slices, while uncorrelated data is transmitted in batches. It adopts a dual encryption scheme of semantic tags and data ontology, and only authorized nodes can parse semantic tags; it dynamically adjusts the transmission rate by combining communication latency parameters, thereby adapting to the communication needs of different traffic scenarios.

4. The vehicle-to-everything (V2X) big data real-time monitoring and user service optimization system for intelligent transportation systems according to claim 3, characterized in that: The data fusion unit is specifically as follows: Receive highly correlated semantic data from the transmission unit, establish a dynamic weighted fusion model, and automatically adjust the data fusion weights based on traffic event types; Deploy a cross-modal data alignment module to calibrate and correct spatiotemporal offsets based on timestamps and spatial coordinates; integrate a data fusion quality assessment module to monitor the fusion effect in real time and dynamically adjust the fusion strategy through indicators including semantic consistency and key field missing rate, so as to achieve spatiotemporal alignment of cross-modal data.

5. The vehicle-to-everything (V2X) big data real-time monitoring and user service optimization system for intelligent transportation systems according to claim 4, characterized in that: The service generation unit is specifically as follows: Based on the causal relationship between traffic incidents and user needs, as well as real-time traffic conditions, a two-dimensional user profile with semantic and causal dimensions is established, and the profile data is synchronously linked to the real-time traffic conditions monitored by the digital twin. A causal recommendation algorithm is used to generate the optimal travel plan by combining fused data, causal inference conclusions and real-time traffic conditions; at the same time, a service strategy dynamic adjustment module is deployed to update the recommendation strategy in real time according to real-time causal factors and changes in traffic conditions.

6. The vehicle-to-everything (V2X) big data real-time monitoring and user service optimization system for intelligent transportation systems according to claim 5, characterized in that: The optimization unit is specifically as follows: Collect feedback data from the entire service chain, establish an optimization process for feedback collection, causal analysis, parameter optimization, and service iteration, and deploy a causal analysis module to identify causal relationships in the feedback data; A causal adaptive optimization algorithm is adopted to dynamically adjust the traffic monitoring model weights, data update frequency and service recommendation parameters according to causal relationships, and the optimization parameters are synchronously pushed to the front-end data fusion unit, causal inference unit and service generation unit. Establish an optimization effect evaluation module to quantitatively evaluate the optimization effect through indicators including service satisfaction, monitoring accuracy, and response timeliness.

Images