A method of associating a vehicle with a terminal and a medium

Abstract

The embodiment of the application provides a kind of vehicle and terminal association method and medium, method includes: the association of vehicle and terminal is determined anchor in current observation event, determines the current co-occurrence feature of vehicle and its adjacent terminal using the observation data of current observation event, and then determines the association of vehicle and its adjacent terminal based on current co-occurrence feature and the scene feature of the scene where vehicle is located, without waiting for vehicle to complete complete trip or collect full trajectory data, improve the real-time of vehicle and terminal association.

Description

Technical Field

[0001] This application belongs to the field of communication technology, and in particular relates to a method and medium for associating a vehicle with a terminal. Background Technology

[0002] In fields such as intelligent traffic management, road network operation monitoring, and autonomous driving collaboration, the technology of associating vehicles with terminals is the foundation for realizing traffic situation awareness and multi-entity collaborative control. This technology binds vehicles with nearby terminals (such as mobile phones and wearable devices) to clarify the ownership relationship between terminals and vehicles, thereby providing reliable association and location data for applications such as vehicle-road collaboration, traffic scheduling, and safety management.

[0003] Currently, the common approach is to compare and analyze the historical spatiotemporal trajectories of vehicles and terminals to correlate vehicles and terminals whose historical spatiotemporal trajectories highly overlap. However, this comparative analysis method based on historical spatiotemporal trajectories requires waiting for the vehicle to complete a full journey and for all data to be collected before correlation calculations can be performed, resulting in poor real-time performance. Summary of the Invention

[0004] This application provides a method for associating a vehicle with a terminal and a computer-readable storage medium, which can improve the real-time performance of the association between the vehicle and the terminal.

[0005] In a first aspect, embodiments of this application provide a method for associating a vehicle with a terminal, including: In response to a current observation event of a vehicle, determine one or more current neighboring terminals of the vehicle within a preset spatiotemporal proximity range; For each of the current neighboring terminals, based on the observation data of the current observation event, determine the current co-occurrence feature of the current neighboring terminal and the vehicle being observed simultaneously; For each of the current neighboring terminals, the association between the current neighboring terminal and the vehicle is determined based on the current co-occurrence features and the scene features of the scene in which the vehicle is located.

[0006] In one feasible implementation, for each current neighboring terminal, the association between the current neighboring terminal and the vehicle is determined based on the current co-occurrence characteristics and the scene characteristics of the scene in which the vehicle is located, specifically including: Based on the scene features and the preset threshold mapping relationship, a confidence threshold is determined. The scene features include at least one of road type features, traffic time period features, and traffic density features. For each of the current neighboring terminals, based on the current co-occurrence features, the current association confidence between the current neighboring terminal and the vehicle is obtained through a preset association confidence model; For each of the current neighboring terminals, if the current association confidence level is greater than the confidence level threshold, then the current neighboring terminal is associated with the vehicle; if the current association confidence level is not greater than the confidence level threshold, then the current neighboring terminal is not associated with the vehicle.

[0007] In one feasible implementation, based on the observation data of the current observation event, the current co-occurrence feature of the current neighboring terminal and the vehicle being observed simultaneously is determined, specifically including: Determine whether the current neighboring terminal exists in a pre-built candidate terminal dictionary, which is used to store information on all terminals that are observed simultaneously with the vehicle. If so, then based on the observation data, update the historical feature data of the current neighboring terminal and the vehicle being observed simultaneously to obtain the current feature data; If not, add the current neighboring terminal to the candidate terminal dictionary and initialize the current feature data based on the observation data of the current observation event; Based on the current feature data, determine the current co-occurrence feature.

[0008] In one feasible implementation, the current co-occurrence feature includes at least one of co-occurrence intensity, spatial proximity stability, speed consistency, role characteristics, route constraint rationality, and group consistency; the co-occurrence intensity characterizes the frequency at which the terminal and the vehicle are observed simultaneously; the spatial proximity stability characterizes the degree of fluctuation in the spatial distance between the terminal and the vehicle; the speed consistency characterizes the correlation between the movement speeds of the terminal and the vehicle; the role characteristics characterize the role of the terminal within the vehicle; the route constraint rationality characterizes the feasibility of the terminal and the vehicle under preset road network constraints; and the group consistency characterizes the similarity of movement behavior patterns between the terminal and associated terminals of multiple vehicles.

[0009] In one feasible implementation, the historical feature data includes the historical mean distance and historical cumulative squared difference between the current neighboring terminal and the vehicle; the current co-occurrence feature includes spatial proximity stability, which characterizes the degree of fluctuation in the spatial distance between the terminal and the vehicle; Based on the observation data of the current observation event, determine the current co-occurrence characteristics of the current neighboring terminal and the vehicle being observed simultaneously, specifically including: Determine the current distance between the current nearest terminal and the vehicle in the current observation event; An online recursive algorithm is used to update the historical average distance based on the distance increment between the current distance and the historical average distance, so as to obtain the current average distance. Based on the distance increment and the current distance mean, update the historical cumulative squared difference to obtain the current cumulative squared difference; The spatial proximity stability is determined based on the current mean distance and the current cumulative squared difference.

[0010] In one feasible implementation, the historical feature data includes the last co-occurrence time of the current neighboring terminal and the vehicle, the first co-occurrence position of the current neighboring terminal at the last co-occurrence time, and the second co-occurrence position of the vehicle at the last co-occurrence time; the current co-occurrence feature includes route constraint rationality, which characterizes the feasibility of the terminal and the vehicle under preset road network constraints; Based on the observation data of the current observation event, determine the current co-occurrence characteristics of the current neighboring terminal and the vehicle being observed simultaneously, specifically including: In cases where the current neighboring terminal and the vehicle are observed simultaneously multiple times, the third co-occurrence position of the current neighboring terminal at the current co-occurrence time and the fourth co-occurrence position of the vehicle at the current co-occurrence time are obtained. Determine the time interval between the current co-occurrence time and the previous co-occurrence time; Under preset route constraints, the shortest path distance and shortest travel time of the current neighboring terminal are determined based on the first co-occurrence location and the third co-occurrence location; and the shortest path distance of the vehicle is determined based on the second co-occurrence location and the fourth co-occurrence location. The feasibility of passage is determined based on the time interval and the shortest passage time of the current neighboring terminal; The distance matching degree is determined based on the difference between the shortest path distance of the current nearest terminal and the shortest path distance of the vehicle; The rationality of the route constraints is determined based on the described feasibility of passage and the described distance matching degree.

[0011] In one feasible implementation, the historical feature data includes the vehicle observation time and the first behavioral data of the associated terminals of the vehicle at the vehicle observation time; the current co-occurrence feature includes group consistency, which characterizes the similarity of the movement behavior patterns between the terminal and the associated terminals of the multiple vehicles. Based on the observation data of the current observation event, determine the current co-occurrence characteristics of the current neighboring terminal and the vehicle being observed simultaneously, specifically including: Based on the first behavior data, determine the group behavior template; Obtain the second behavioral data of the currently nearby terminal at the time of observation of the vehicle; The group consistency is determined based on the similarity between the second behavioral data and the group behavior template.

[0012] In one feasible implementation, after obtaining the current association confidence between the current neighboring terminal and the vehicle based on the current co-occurrence characteristics and a preset association confidence model, the method further includes: Obtain the true association confidence level between the current nearest terminal and the vehicle; Determine the prediction error between the current association confidence level and the true association confidence level; The association confidence model is updated based on the prediction error and the current co-occurrence features.

[0013] In one feasible implementation, the method further includes: The transfer hypothesis is triggered when the confidence level of the association between the target terminal and the first mode of transportation decreases, the confidence level of the association between the target terminal and the second mode of transportation increases, and the first mode of transportation and the second mode of transportation have a spatiotemporal intersection point. Based on the historical co-occurrence characteristics of the target terminal with the first mode of transportation and the second mode of transportation, verify whether the transfer hypothesis is valid. If the transfer assumption holds true, the target terminal, the first vehicle, and the second vehicle are marked as a transfer event; If multiple vehicles are detected to be within a distance of a preset distance threshold, or if the similarity between the group behavior templates of the multiple vehicles is greater than a preset similarity threshold, the multiple vehicles will be marked as a collaborative vehicle fleet event.

[0014] Secondly, embodiments of this application provide a device for associating a vehicle with a terminal, the device comprising: The event triggering module is used to respond to the current observation event of the vehicle and determine one or more current nearby terminals of the vehicle within a preset spatiotemporal proximity range; The co-occurrence feature module is used to determine, for each of the current neighboring terminals, the current co-occurrence feature of the current neighboring terminal and the vehicle being observed simultaneously, based on the observation data of the current observation event; The association determination module is used to determine the association relationship between each current neighboring terminal and the vehicle based on the current co-occurrence features and the scene features of the scene in which the vehicle is located.

[0015] Thirdly, embodiments of this application provide a device for associating a vehicle with a terminal. The device includes a processor and a memory storing computer program instructions. The processor reads and executes the computer program instructions to implement any of the aforementioned methods for associating a vehicle with a terminal.

[0016] Fourthly, embodiments of this application provide a computer storage medium storing computer program instructions, which, when executed by a processor, implement any of the above-described methods for associating a vehicle with a terminal.

[0017] Fifthly, embodiments of this application provide a computer program product in which instructions, when executed by the processor of an electronic device, cause the electronic device to perform any of the aforementioned methods for associating a vehicle with a terminal.

[0018] This application provides a method and medium for associating a vehicle with a terminal, which can anchor the determination of the association between the vehicle and the terminal to the current observation event. It uses the observation data of the current observation event to determine the current co-occurrence characteristics of the vehicle and its neighboring terminals, and then determines the association relationship between the vehicle and its neighboring terminals based on the current co-occurrence characteristics and the scene characteristics of the scene in which the vehicle is located. This does not require waiting for the vehicle to complete a complete journey or collect full trajectory data, thus improving the real-time performance of the association between the vehicle and the terminal.

[0019] In addition, this application can also introduce features such as co-occurrence intensity, spatial proximity stability, speed consistency, role characteristics, route constraint rationality, and group consistency to comprehensively determine the association between the terminal and the vehicle from multiple levels, thereby improving the accuracy and reliability of the association. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating a method for associating a vehicle with a terminal, as provided in an embodiment of this application. Figure 2 This is a schematic diagram of the event triggering and current nearby terminal retrieval process provided in the embodiments of this application; Figure 3This is a schematic diagram of the process for calculating the association confidence level provided in an embodiment of this application; Figure 4 This is a flowchart illustrating the process of determining the association relationship and optimizing parameters online, as provided in an embodiment of this application. Figure 5 This is a schematic diagram of the structure of a vehicle-terminal association device provided in an embodiment of this application; Figure 6 This is a schematic diagram of the structure of a vehicle-terminal association device provided in an embodiment of this application. Detailed Implementation

[0022] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0023] It should be noted that, in this document, relational terms such as "first" and "second" are used merely 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 a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0024] In fields such as intelligent traffic management, road network operation monitoring, and autonomous driving collaboration, the technology of associating vehicles with terminals is the foundation for realizing traffic situation awareness and multi-entity collaborative control. This technology binds vehicles with nearby terminals (such as mobile phones and wearable devices) to clarify the ownership relationship between terminals and vehicles, thereby providing reliable association and location data for applications such as vehicle-road collaboration, traffic scheduling, and safety management.

[0025] Currently, the common approach is to compare and analyze the historical spatiotemporal trajectories of vehicles and terminals to correlate vehicles and terminals whose historical spatiotemporal trajectories highly overlap. However, this comparative analysis method based on historical spatiotemporal trajectories requires waiting for the vehicle to complete a full journey and for all data to be collected before correlation calculations can be performed, resulting in poor real-time performance.

[0026] To address the problems of the prior art, embodiments of this application provide a method, apparatus, device, and medium for associating a vehicle with a terminal.

[0027] In practical applications, the execution subject of the method for associating vehicles and terminals in this application embodiment can be a terminal device, such as a desktop computer or laptop computer, or a remote device similar to a server. Of course, this application embodiment can also adopt an execution subject in the form of software, such as a client or software program installed on a terminal device. The specific type of execution subject corresponding to the technical solution provided in this application embodiment is not strictly limited here, and can be flexibly selected according to the actual application scenario and actual needs.

[0028] The following describes specific embodiments of a method, apparatus, medium, and device for associating a vehicle with a terminal, as provided in this application. First, a method for associating a vehicle with a terminal will be introduced.

[0029] Figure 1 This illustration shows a flowchart of a method for associating a vehicle with a terminal according to an embodiment of this application. Figure 1 As shown, the method includes steps S100 to S102: S100: In response to a current observation event of the vehicle, determine one or more current neighboring terminals of the vehicle within a preset spatiotemporal proximity range.

[0030] In one or more embodiments of this application, in order to determine the association relationship between terminals that may be associated with vehicles in subsequent steps, this application may, in response to the current observation event of the vehicle, filter out the current neighboring terminals of the vehicle within a preset spatiotemporal proximity range.

[0031] Specifically, in response to a current observation event of a vehicle, this application can determine one or more currently nearby terminals of the vehicle within a preset spatiotemporal proximity range. The observation data of the current observation event includes, but is not limited to, the unique identifier of the vehicle, the observation time of the vehicle, and the location of the vehicle.

[0032] It should be noted that this application does not limit the specific type of vehicle, such as any one of private cars, taxis, ride-hailing vehicles, buses, freight vehicles, rail transit vehicles, water transport vehicles, or air transport vehicles. This application does not limit the specific source of the observation data, which can be set according to actual needs. For example, observation data such as license plate number, location, speed, vehicle observation time, and heading angle can be obtained through roadside units (RSUs); observation data such as license plate number, vehicle type, location, and vehicle observation time can be obtained through traffic checkpoints / cameras; observation data such as license plate number, vehicle trajectory, and bus data can be obtained through on-board units (OBUs); and observation data can be captured through mobile acquisition devices (such as sensing devices mounted on patrol cars or drones). The above data sources can be used individually or in combination to improve the coverage and reliability of the observation data. This application does not limit the data source used when filtering nearby terminals. It can be set according to actual needs, such as cellular network signaling data, Wi-Fi / Bluetooth probe data, roadside C-V2X messages, etc. These data sources can be used individually or in combination to improve the coverage and reliability of terminal retrieval. When multiple data sources provide conflicting observation data for the same vehicle or terminal at the same time / time period, this application can resolve conflicts and select reliable data based on the following principles: spatiotemporal accuracy priority, i.e., prioritizing data sources with higher positioning accuracy; timestamp nearest, i.e., for dynamic observation data such as vehicle speed, selecting data closest to the vehicle observation time of the current observation event; source reliability weight, i.e., assigning static or dynamic reliability weights to each data source and performing a weighted average during fusion; consistency voting, i.e., enhancing the reliability of observation data when multiple independent data sources report the same observation data, which can be supplemented by observation consistency logs for judgment.

[0033] Furthermore, this application does not limit the specific size of the spatiotemporal proximity range, which can be set according to actual needs. For example, the spatial radius can be 50 meters, 100 meters, 200 meters, etc.; the time window can be 5 seconds, 10 seconds, 30 seconds, etc. In highway scenarios, the spatial radius can be appropriately increased to cover more distant terminals; in densely populated urban scenarios, the radius can be reduced to reduce interference with terminals. This application does not limit the specific form representing the vehicle and the terminal, which can be set according to actual needs. For example, the hash value of IMEI, IMSI, or MAC address can be used as the identifier of the terminal; the license plate number and / or visual features of the vehicle can be used as the identifier of the vehicle.

[0034] For example, taking the position of the vehicle at observation time t as the center, with a spatial radius of 100 meters and a time window of 10 seconds, from Wi... The Fi probe data extracts the MAC address hash values ​​of all terminals detected within the same spatiotemporal proximity, and identifies these terminals as current neighboring terminals. Simultaneously, if cellular network signaling data is available, the IMSI hash values ​​of current neighboring terminals located within the same spatiotemporal proximity by the base station can be merged into the list, and duplicate identifiers removed.

[0035] Figure 2 This is a schematic diagram illustrating the event triggering and current nearby terminal retrieval process provided in an embodiment of this application. For example... Figure 2 As shown, the process of event triggering and current neighboring terminal retrieval includes steps S200 to S204, as follows: S200: The arrival of the current observed event for the vehicle.

[0036] S201: Extract observation data such as the vehicle identification number (VID), observation time, location, and speed of the vehicle.

[0037] S202: Update the spacetime log of the vehicle.

[0038] S203: Retrieve the current nearest terminal of a vehicle within the spatiotemporal proximity range.

[0039] S204: Determine the current set of neighboring terminals.

[0040] S101: For each of the current neighboring terminals, based on the observation data of the current observation event, determine the current co-occurrence feature of the current neighboring terminal and the vehicle being observed simultaneously.

[0041] In one or more embodiments of this application, in order to evaluate the correlation between the terminal and the vehicle based on unified and standardized quantitative indicators in subsequent steps, this application needs to use the observation data provided by the current observation event to construct the current co-occurrence features that can be directly used by the correlation confidence model.

[0042] It should be noted that this application does not limit the specific type of co-occurrence features, which can be set according to actual needs, such as the consistency of the relative azimuth angle between the terminal and the vehicle, the matching degree between the trend of terminal signal strength change and the trend of vehicle distance change, the time synchronization of the terminal and the vehicle entering or leaving a specific area simultaneously, and the statistical difference between the terminal's movement direction and the vehicle's heading angle. In order to comprehensively determine the credibility of the association between the terminal and the vehicle from multiple dimensions, in one or more embodiments of this application, the current co-occurrence features include at least one of co-occurrence intensity, spatial proximity stability, speed consistency, role characteristics, route constraint rationality, and group consistency. Among them, co-occurrence intensity characterizes the frequency of the terminal and the vehicle being observed simultaneously; spatial proximity stability characterizes the degree of fluctuation in the spatial distance between the terminal and the vehicle; speed consistency characterizes the correlation of the movement speed between the terminal and the vehicle; role characteristics characterize the role of the terminal within the vehicle; route constraint rationality characterizes the feasibility of the terminal and the vehicle under preset road network constraints; and group consistency characterizes the similarity of the movement behavior patterns between the terminal and multiple associated terminals of various vehicles.

[0043] In order to efficiently store, update and retrieve historical feature data and association status between terminals and vehicles, in one or more embodiments of this application, this application may maintain a candidate terminal dictionary (terminal_candidates) for vehicles. The candidate terminal dictionary is a data structure used to store information of all terminals that are observed simultaneously with the vehicle. This application does not limit the specific content of the candidate terminal dictionary, and it can be set according to actual needs. For example, the candidate terminal dictionary may also store features reflecting the association status, association confidence, and other information.

[0044] Taking a key-value pair data structure as an example, the candidate terminal dictionary uses the terminal's unique identifier (TID) as the key and a structure containing feature F, association confidence C, and a co-occurrence record list as the value. The co-occurrence record list stores historical feature data of the terminal and the vehicle being detected simultaneously. This application does not limit the specific type of historical feature data; it can be set according to actual needs, such as co-occurrence frequency, mean distance, distance variance, cumulative sum and cross product of speed sequences, total co-occurrence duration, and cross-day counts, to improve the accuracy of subsequent association determination. This application can also perform validity checks on the updated current feature data, such as removing obviously abnormal distance or speed values ​​to prevent erroneous data from contaminating the feature statistics.

[0045] Meanwhile, in order to maintain the cumulative association evidence between each current neighboring terminal and the vehicle, and to support subsequent incremental feature updates and confidence calculations, in one or more embodiments of this application, this application can determine whether each current neighboring terminal exists in a pre-built candidate terminal dictionary; if so, then update the historical feature data according to the observation data of the current observation event to obtain the current feature data; if not, then add the current neighboring terminal to the candidate terminal dictionary, and initialize the current feature data according to the observation data of the current observation event; and determine the current co-occurrence features of the current neighboring terminal according to the current feature data.

[0046] In this embodiment, by maintaining a candidate terminal dictionary, the application can quickly locate the state of each current neighboring terminal when a current observation event arrives, realize incremental updates or initialization of feature data, and update the dictionary content in a timely manner after determining the correlation and other observation events. The candidate terminal dictionary can be stored in memory to support high-concurrency access, or it can be incorporated into the database for historical backtracking.

[0047] S102: For each of the current neighboring terminals, determine the association between the current neighboring terminal and the vehicle based on the current co-occurrence features and the scene features of the scene in which the vehicle is located.

[0048] In one or more embodiments of this application, in order to adapt the association determination results to the complex and ever-changing traffic environment and avoid misjudgment or omission in different scenarios by using fixed determination criteria, in this step, this application needs to combine the current co-occurrence features with the scene features of the scene where the vehicle is located to determine the association relationship between the current neighboring terminal and the vehicle.

[0049] It should be noted that this application does not limit the specific content of scene features, which can be set according to actual needs. For example, scene features may include road type features, traffic time period features, traffic density features, weather condition features, signal quality features, etc. Road type features represent the road condition attributes of the current road segment where the vehicle is located, which can be determined by at least one of the following methods: navigation map data, roadside unit broadcast information, or historical driving trajectory matching results; traffic time period features represent the time interval attributes of the current observed event, which can be determined by the current clock information and preset time period division rules; traffic density features represent the density of terminal distribution within a preset range of the vehicle, which can be calculated based on the number of nearby terminals or the terminal density per unit area determined in the current observed event; weather condition features represent the environmental meteorological conditions at the current moment, which can be obtained through vehicle-mounted sensors, meteorological service platforms, or roadside sensing devices; signal quality features represent the strength and stability of positioning signals or communication signals in the current environment, which can be determined by at least one parameter among the global navigation satellite system signal reception quality, base station signal strength, or signal-to-noise ratio. This application does not limit the specific methods for determining the association relationship based on scene features and current co-occurrence features. For example, the current co-occurrence features and scene features are concatenated into joint features, which are then input into a pre-trained classification model, and the model directly outputs the determination result of whether an association relationship exists. Alternatively, a rule base can be pre-built, and corresponding judgment rule sets can be configured for different scene feature values. Each rule contains logical judgment conditions for one or more dimensions of the co-occurrence features, and then the association relationship can be determined based on the rule set. Or, the dimensions of the current co-occurrence features can be dynamically weighted according to the scene features (such as increasing the weight of speed consistency features in high-speed scenarios), a weighted comprehensive score can be calculated, and then compared with a preset threshold to determine the association relationship.

[0050] In the above-mentioned method for associating vehicles and terminals, this application can respond to the current observation event by determining the terminal that is spatiotemporally adjacent to the vehicle in real time, and extracting the current co-occurrence features between the terminal and the vehicle based on the current observation data. Then, it can combine scene features to make scene-adaptive association determination. Without waiting for the vehicle to complete a full journey or accumulate long-term trajectory data, the association and binding of the terminal and the vehicle can be completed at the current observation time. This achieves low-latency dynamic vehicle association, improves robustness to data discontinuity, and can meet the actual needs of scenarios such as intelligent traffic management and dynamic safety early warning.

[0051] Furthermore, in one or more embodiments of this application, this application can maintain a dynamic vehicle dynamic profile for vehicles, so as to continuously record their observed spatiotemporal information, type labels, and multidimensional features and confidence levels of candidate terminals during the vehicle's operation, thereby supporting real-time, incremental association determination and self-evolutionary updates.

[0052] It should be noted that this application does not limit the specific format of the vehicle dynamic file; it can be set according to actual needs, such as using relational database tables, document-oriented storage structures, or key-value pair caching, etc. Taking a key-value pair structure as an example, the vehicle dynamic file is as follows: Specifically, the key is the unique identifier (VID) of the vehicle, which can be composed of the license plate number and the vehicle's visual features. Visual features are generated by hashing the vehicle type, color, and main appearance features extracted from the vehicle image, used to enhance uniqueness and assist in detecting cloned license plates. The value is a dynamic dictionary, including but not limited to a spatiotemporal log, vehicle type labels, and a candidate terminal dictionary. The spatiotemporal log stores the spatiotemporal information of the observed vehicles in chronological order, such as time, longitude, latitude, speed, and heading. The vehicle type labels include basic types and dynamically inferred subtypes. Basic types include, but are not limited to, private cars, taxis, and buses, which can be obtained from license plate color rules or registration information. Subtypes are inferred online through behavioral patterns; for example, ride-hailing services can be inferred through APP connection modes or high-frequency short-distance shuttle modes. Private cars can be inferred through high-frequency, regular, long-distance passenger-carrying behavior.

[0053] In real-world road environments, due to factors such as signal coverage blind spots, inconsistent probe scanning cycles, or terminal sleep states, there may be instances where a terminal is actually on the vehicle but is not detected. In step S101, to further improve the accuracy of determining the association confidence level, in one or more embodiments of this application, a co-occurrence strength capable of resisting data misses is designed as a co-occurrence feature, as follows: First, the historical feature data in this application includes the total number of observations of the vehicle after the current neighboring terminal and the vehicle were first observed simultaneously, the total number of simultaneous co-occurrences of the current neighboring terminal and the vehicle, the number of interval observations between the current co-occurrence and the previous co-occurrence, and the effective co-occurrence count after compensating for missed detections; the current co-occurrence feature includes the co-occurrence intensity. Second, based on the observation data of the current observation event, the total number of observations, the total number of co-occurrences, and the number of interval observations are updated, and then the effective co-occurrence count is determined based on these updated parameters. Finally, the co-occurrence intensity is determined based on the effective co-occurrence count and the updated total number of observations.

[0054] For example, after terminal T and vehicle V are first observed simultaneously (co-occurrence), the total number of observations of vehicle V is: The total number of occurrences of terminal T and transportation is Initialization command , , Incremental updates are performed on these three historical feature data points, that is, the total number of times the vehicle is observed each time, regardless of whether terminal T is observed. Increase by 1; if terminal T and vehicle V co-occur in the current observed event, then the total number of occurrences is [number missing]. Increase by 1, and record the interval between the current co-occurrence and the previous co-occurrence as the number of observations. (at the time of the first update) ), update as follows :

[0055] In formula (1), This is the compensation coefficient for missed detections, with a range of values. This is used to estimate the probability that the terminal is still on the vehicle during an unobserved gap; The value can be customized according to the actual scenario, such as setting it to 0.3; if it is believed that the terminal is still on the vehicle when the missed detection is highly likely, a larger value of 0.4 can be used, otherwise a smaller value of 0.2 can be used. If the terminal T and the vehicle do not co-occur in this observation, then and The co-occurrence intensity remains unchanged. The range is The stronger the co-occurrence intensity, the longer the relative duration of terminal T on the vehicle.

[0056] Assuming α = 0.3, the vehicle was observed 4 times consecutively. Terminal T was observed to co-occur with the vehicle in the 1st, 2nd, and 4th observations, but there was a missed detection in the 3rd observation. , , The update process is as follows: First co-occurrence, initialization , , , ; Second co-occurrence, Updated , , , No co-occurrence for the third time, update. , , , ; 4th co-occurrence, ,renew , , , .

[0057] In this example, even if there is a missed detection in the third observation, the compensation mechanism of this application still keeps the F1 at a high level, demonstrating high robustness to missing data.

[0058] In real-world road environments, due to factors such as positioning errors and signal attenuation, single spatial distance measurements between the terminal and vehicles may fluctuate, and relying solely on a single proximity measurement cannot ensure the reliability of the association determination. In step S101, to eliminate accidental proximity interference and comprehensively evaluate spatial density and stability, in one or more embodiments of this application, spatial proximity stability is designed as a co-occurrence feature, as follows: First, the historical feature data in this application includes the historical mean distance and historical cumulative squared difference between the current neighboring terminal and the vehicle; the current co-occurrence feature includes spatial proximity stability. Second, this application can determine the current distance between the current neighboring terminal and the vehicle in the current observation event, and then uses an online recursive algorithm to update the historical mean distance based on the distance increment between the current distance and the historical mean distance, obtaining the current mean distance; and updates the historical cumulative squared difference based on the distance increment and the current mean distance, obtaining the current cumulative squared difference. Finally, spatial proximity stability is determined based on the current mean distance and the current cumulative squared difference.

[0059] It should be noted that this application does not limit the specific type of distance, which can be set according to actual needs, such as Euclidean distance, Manhattan distance, Mahalanobis distance or geodetic distance, etc. In practical applications, an appropriate distance metric can be selected according to the coordinate system, accuracy and computational complexity of the positioning data, and this application does not limit it in this regard.

[0060] Continuing with the previous example, whenever terminal T and vehicle V co-occur, record the Euclidean distance between their locations. An online recursive algorithm (Welford) is used to incrementally calculate the distance mean and distance variance of the distance sequence, eliminating the need to store all historical distances. After the first co-occurrence of terminal T and vehicle V (k=1), the distance mean is initialized. Cumulative Squared Difference For the first Secondary co-occurrence ( ), current distance is Perform the following calculations in sequence:

[0061] In equations (2) to (6), The average of the current distance and the historical distance The distance increment; Equation (3) is the update process of the historical distance mean; Let be the historical cumulative squared difference; Equation (4) is the update process for the historical cumulative squared difference; The square root of the current distance variance This represents the current standard deviation of the distance, reflecting the degree of distance fluctuation. This is a preset reference distance used to normalize the mean. The value can be set according to actual needs, such as 50 meters. For spatial proximity stability, the value range is (0,1]. The closer it is to 1, the more spatially close and stable the terminal and the vehicle are.

[0062] In this embodiment, this application does not need to store all historical distances, but only needs to maintain... , and That is, the mean and variance of the distance are updated in real time each time a co-occurrence occurs, and an exponential decay term is used to penalize excessively large average distances. The penalty distance fluctuates too much, and the product of the two means that F2 can only approach 1 if the terminal is consistently close to the vehicle over a long period of time.

[0063] If the distance between three consecutive co-occurrences is 5 meters, then , F2 = 0.905; if the distances are 5 meters, 8 meters, and 6 meters respectively, then , F2 = 0.488, a significant decrease; if the distance is large and fluctuates greatly, for distances of 50 meters, 60 meters, and 56 meters respectively, then... , F2=0.0534, effectively suppressing interference from distant or unstable targets.

[0064] In this example, spatial proximity stability can reflect the association status between the terminal and the vehicle from the dual dimensions of spatial density and stability.

[0065] In real-world road environments, terminal speed estimation can be affected by factors such as positioning jitter and inconsistent sampling intervals, making a single speed comparison insufficient to accurately reflect whether the terminal and the vehicle are moving synchronously. In step S101, to verify the consistency of their movement behavior patterns from a kinematic perspective, in one or more embodiments of this application, speed consistency is designed as a co-occurrence feature, as follows: First, the historical feature data in this application includes the historical speed sequence of the current neighboring terminal and the historical speed sequence of the vehicle; the current co-occurrence feature includes speed consistency. Second, this application can determine the Pearson correlation coefficient between two historical speed sequences and map this coefficient to the [0,1] interval. Finally, based on the normalized Pearson correlation coefficient, speed consistency is determined.

[0066] It should be noted that this application does not limit the specific type of velocity, which can be set according to actual needs. For example, velocity can be characterized by the difference between continuous positioning points or the average velocity between adjacent co-occurring points.

[0067] Continuing with the previous example, whenever terminal T and vehicle V co-occur, their speeds are recorded. Based on their historical speed sequences, the process of determining speed consistency is as follows:

[0068] In equations (7) to (10), This is a historical speed sequence of a vehicle across multiple consecutive observations. This is a historical velocity sequence of the terminal across multiple consecutive observations. This represents the average speed of the vehicle over multiple consecutive observations. This represents the average velocity of the terminal across multiple consecutive observations. The Pearson correlation coefficient has a range of [ 1,1]; To ensure speed consistency, the value range is (0,1]. The closer to 1, the more consistent the speed change trends of the terminal and the vehicle, such as synchronous acceleration, deceleration, or constant speed; the closer to 0, the opposite the speed change trends, such as the terminal decelerating when the vehicle accelerates; 0.5 indicates no significant linear correlation and cannot provide effective kinematic evidence.

[0069] In this embodiment, by calculating the Pearson correlation coefficient of the historical speed sequences of the terminal and the vehicle, this application can quantitatively assess whether the two are moving synchronously from a kinematic perspective, providing physical motion evidence independent of spatial distance for correlation determination, and avoiding misjudgment caused by positioning errors or accidental proximity.

[0070] If the historical speed sequence of the vehicle observed three times is [60, 65, 70], in km / h; and the historical speed sequence of the terminal is [58, 63, 68], in km / h; , Pearson correlation coefficient ; If the historical speed sequence of three observations of the vehicle is [60, 60, 60], in km / h; and the historical speed sequence of the terminal is [58, 63, 68], in km / h; and the Pearson correlation coefficient is... ; If the historical speed sequences of terminals and vehicles show an inverse relationship, then... .

[0071] In this example, speed consistency can reflect the association state between the terminal and the vehicle from a kinematic perspective.

[0072] In step S101, in order to assess the credibility of the association between the terminal and the vehicle from the perspective of social attributes and long-term behavioral patterns, in one or more embodiments of this application, role characteristics are designed as co-occurrence characteristics, as follows: First, the historical feature data in this application includes registration matching information between the current neighboring terminal and the vehicle, cumulative co-occurrence duration, first co-occurrence time, co-occurrence days, etc.; the current co-occurrence features include role features; the role of the terminal in the vehicle includes, but is not limited to, driver, long-term passenger, temporary passenger, etc.

[0073] Secondly, this application can determine three role sub-features: role score R, co-occurrence duration percentage D, and cross-day score Day. The role score is determined based on whether the terminal matches the vehicle registration information or meets the long-term co-occurrence condition. If the terminal ID matches the vehicle registration information, then R=1.0; if the number of days the terminal co-occurs with the vehicle is greater than a preset number of days threshold (e.g., 3 days) and the cumulative co-occurrence duration percentage is greater than a preset percentage threshold (e.g., 70%), then it is judged as a long-term passenger, R=0.8; other cases are considered temporary passengers, R=0.3. The co-occurrence duration percentage is the ratio of the cumulative co-occurrence duration of the terminal and the vehicle to the current time minus the first co-occurrence time. The cross-day score Day=min(1.0, number of different natural days that co-occurred / 7), used to reflect the long-term nature of the association. In this embodiment, a seven-day week is used as the normalization period.

[0074] Finally, the three sub-features of the above roles are weighted and summed to obtain the role features. ,in, , , as well as The sub-feature weights are assigned to each element within the character traits. In this application, these sub-feature weights can be set according to actual needs, such as... , as well as This example setting can emphasize the dominance of character scores.

[0075] It should be noted that the specific values ​​of registration information, day thresholds, duration percentage thresholds, and each weight in this application can be customized according to the actual application scenario, and this application does not impose any restrictions. For example, in the ride-hailing scenario, the weight of the role score can be appropriately reduced, while in the private car scenario, the weight of registration information can be increased.

[0076] Continuing with the previous example, if the terminal and vehicle registration information match (driver), R=1.0, the co-occurrence duration ratio D=0.9, spanning 5 days (Day 5 / 7 ≈ 0.714), then F4=0.5×1.0+0.3×0.9+0.2×0.714=0.913. If it's a long-term passenger, R=0.8, D=0.6, spanning 4 days (Day ≈ 0.571), then F4=0.694. If it's a temporary passenger, R=0.3, D=0.1, spanning 1 day (Day ≈ 0.143), then F4=0.209.

[0077] In this embodiment, role features can effectively distinguish between drivers, long-term passengers and temporary passengers, providing strong prior information for the association confidence model, especially when registration information is available or long-term co-occurrence patterns are obvious, thus improving the accuracy of association discrimination.

[0078] In real-world road environments, the positioning points of terminals and vehicles may be affected by factors such as tall buildings blocking the view and multipath effects. Euclidean distance alone cannot distinguish between situations where roads are inaccessible, such as those separated by rivers or overpasses. In step S101, to verify whether the terminal and vehicle are moving along the same route from a physical path feasibility perspective, in one or more embodiments of this application, the rationality of route constraints is designed as a co-occurrence feature. The determination process is as follows: First, the historical feature data in this application includes the last co-occurrence time of the current neighboring terminal T and the vehicle V. The first co-occurrence position of the current neighboring terminal at the last co-occurrence time. And the second co-occurrence position of the vehicle at the last co-occurrence time. Current co-occurrence characteristics include the rationality of route constraints. This application can obtain the time of co-occurrence of a current neighboring terminal when it and a vehicle are observed simultaneously multiple times. The third co-occurrence position And the fourth co-occurrence position of the means of transportation at this co-occurrence time. Determine the time interval between the current co-occurrence time and the previous co-occurrence time. ).

[0079] Secondly, under the preset route constraints, the shortest path distance to the current nearest terminal is determined based on the first co-occurrence location and the third co-occurrence location. and shortest travel time Furthermore, based on the second and fourth co-occurrence locations, the shortest path distance for the means of transport is determined. Based on the time interval and the shortest travel time of the current nearest terminal, determine the feasibility of passage; based on the difference between the shortest path distance of the current nearest terminal and the shortest path distance of the vehicle, determine the distance matching degree.

[0080] Finally, the rationality of the route constraints is determined based on the feasibility of passage and the degree of distance compliance. ,as follows:

[0081] In equations (11) to (13), For feasibility of passage, the closer its value is to 1, the less the actual passage time is than the theoretical minimum time; The distance matching score is the closest value to 1, indicating that the path lengths of the two paths are similar. To ensure the rationality of route constraints, the value range is (0,1]. The closer it is to 1, the more consistent the movement path of the terminal and the vehicle is under the road network constraints.

[0082] It should be noted that this application does not limit the specific implementation method of the map route planning service, which can be set according to actual needs, such as using third-party map software or building a self-built road network index; this application does not limit the method of determining the shortest path distance and the shortest travel time, and classic path search algorithms such as Dijkstra and A* can be used. When the positioning accuracy of the terminal or vehicle is low, the threshold for determining the feasibility of passage can be appropriately relaxed, or data with excessive positioning errors can be filtered in the preprocessing.

[0083] Continuing with the previous example, the shortest path distance between the two co-occurrence locations of the terminal. Shortest travel time The shortest path distance between two co-occurring locations of a vehicle Shortest travel time The time interval between the current co-occurrence and the previous co-occurrence is 70 seconds. Distance fit. Reasonableness of time F5 ≈ 0.961, indicating that the movement is highly reasonable under the road network constraints. When continuous co-occurrence pairs cannot be obtained, F5 takes a neutral value of 0.5, which does not affect the overall judgment.

[0084] In this example, this application introduces high-precision map road network data to comprehensively determine whether the terminal and the vehicle are moving along the same physical route from two dimensions: path length matching degree and time feasibility. This effectively filters out false associations caused by Euclidean distances that are close but the road network is inaccessible (such as crossing rivers, passing through buildings, going up and down overpasses, etc.), and improves the accuracy of association discrimination in complex road environments.

[0085] Furthermore, in order to improve the accuracy of association discrimination by leveraging swarm intelligence in complex signal environments, in one or more embodiments of this application, swarm consistency may be designed as a co-occurrence feature, as follows: First, the historical feature data in this application includes the observation time of the vehicles and the first behavioral data of the associated terminals of multiple vehicles at the observation time of the vehicles. This first behavioral data includes, but is not limited to, normalized speed and normalized distance from the vehicles. Current co-occurrence features include group consistency. This application can determine a group behavior template based on each first behavioral data point. For example, taking M observation times of the vehicles, for each observation time of the vehicles, the median of the behavioral values ​​of all associated terminals at that observation time is calculated as a representative value of the group behavior. These representative values ​​of the group behavior are then arranged in chronological order to construct a group behavior template. .

[0086] Secondly, obtain the second action data of the current neighboring terminal at the time of vehicle observation. .

[0087] Finally, group consistency is determined based on the similarity between the second behavioral data and the group behavioral template. Continuing with the previous example, the Dynamic Time Warping (DTW) algorithm is used to calculate... and Minimum cumulative distance between The average distance is obtained by dividing by the sequence length M. Then, population consistency is determined through exponential decay. ,in, This is a scaling parameter used to control the rate at which similarity decays with increasing average distance. If there is currently no group behavior template, a neutral value is assigned. .

[0088] It should be noted that this application does not limit the specific content or size of the behavioral data, template sequence length, or scale parameter, and these can be set according to actual needs. For example, the scale parameter can be matched to the observation error level of the behavioral data; let the observation error be... ;exist hour, ,So ,get This application does not limit the specific method for determining similarity; it can be set according to actual needs. In this example, the DTW algorithm can tolerate the asynchrony between the observation times of the terminal and the vehicle, effectively aligning the behavioral sequence offset caused by the difference in observation times.

[0089] Continuing with the previous example, suppose the group behavior template (with speed as the element) for the three most recent observation times of each mode of transportation is as follows: The second line of data from the current neighboring terminal After DTW alignment, and , and , and Alignment, then , Assuming the observation error is 0.03, in hour, ,So ,get ; .

[0090] In this example, group consistency uses the confirmed in-vehicle group behavior as a dynamic reference. By aligning the temporal behavior sequence, it captures the collaborative pattern between the new terminal and the group, improving the ability to determine associations in scenarios with intersecting signals and ambiguous individual characteristics.

[0091] In step S102, in order to achieve accurate and robust association determination under different road environments and traffic load conditions, in one or more embodiments of this application, the confidence threshold is dynamically determined based on the scene characteristics of the vehicle's location, as follows: First, a confidence threshold is determined based on scene features and a preset threshold mapping relationship. Scene features include at least one of road type features, traffic time period features, and traffic density features. Second, for each current neighboring terminal, based on current co-occurrence features, a preset association confidence model is used to obtain the current association confidence between the current neighboring terminal and the vehicle. Finally, for each current neighboring terminal, if the current association confidence is greater than the confidence threshold, then the current neighboring terminal and the vehicle are associated; if the current association confidence is not greater than the confidence threshold, then the current neighboring terminal and the vehicle are not associated.

[0092] In this embodiment, the specific setting and updating method of the confidence threshold is not limited. It can be set according to actual needs, such as adjusting a preset fixed confidence threshold based on scene characteristics, a value calculated based on rules, or a value dynamically generated by the model. The threshold update cycle can be updated in real time with each observation event, or it can be calibrated offline periodically. For example, a fixed threshold, such as 0.7, can be used as the confidence threshold. In scenarios with sparse traffic and little signal interference, such as highways, a lower threshold (such as 0.6) can be dynamically switched to reduce the false alarm rate. In complex scenarios such as urban side roads and dense traffic, a higher threshold (such as 0.75) can be dynamically switched to reduce the false alarm rate. The specific form of the association confidence model is not limited and can be set according to actual needs, such as using support vector machines, decision trees, or lightweight neural networks. Of course, the updating method of model parameters is also not limited, and stochastic gradient descent, batch gradient descent, or adaptive learning rate optimizers can be used. In one or more embodiments of this application, the association confidence scores output multiple times by the association confidence model can be fused to reduce fluctuations caused by single observation anomalies and make the association determination more stable. For example, an exponentially weighted moving average can be used to smooth the association confidence scores output multiple times consecutively. For newly appearing current neighboring terminals that are not stored in the candidate terminal dictionary, the confidence score output by the association confidence model is low. The confidence score can be increased by combining multiple subsequent observations to achieve progressive association confirmation.

[0093] In one or more embodiments of this application, a logistic regression model can be used as the association confidence model. The linear combination of current co-occurring features is mapped to a probability value using a sigmoid function, which serves as the current association confidence. This application does not limit the training method of the association confidence model; it can be set according to actual needs. For example, the association confidence model can be pre-trained using discrete historical data. Of course, the values ​​of each type of feature in different scenarios can be set according to actual needs. Here, the road type feature of highways is set to 0, the road type feature of urban arterial roads is set to 1, the road type feature of secondary arterial roads is set to 2, and the road type feature of branch roads is set to 3; the traffic period feature during morning peak hours is set to 1, the traffic period feature during off-peak hours is set to 2, the traffic period feature during evening peak hours is set to 3, and the traffic period feature during nighttime hours is set to 4; the traffic density feature is a normalized value representing the current traffic density of the road segment.

[0094] For example, scene features are ], For road type characteristics, Characteristics of traffic periods, Traffic flow density is a characteristic feature. A logistic regression function is used as the threshold mapping relationship, as follows:

[0095] In equation (14), For scene features; The weight vector represents the threshold mapping relationship. This is the bias term for the threshold mapping relationship; This is the sigmoid function, used to map linear combination values ​​to the (0,1) interval, making them comparable to the association confidence level.

[0096] In addition, the parameters of the threshold mapping relationship and It can be randomly initialized or obtained based on pre-training with historical data. Of course, this application can also dynamically adjust the confidence threshold using an online learning algorithm based on the accuracy feedback of historical judgments in different contexts, thereby enabling the confidence threshold to adapt to the dynamic association judgments of different scenarios. For example, ], , Then the linear part Thus, the confidence threshold is obtained. .

[0097] Figure 3 This is a schematic diagram illustrating the process of calculating the association confidence level provided in an embodiment of this application. Figure 3 As shown, the process of calculating the association confidence includes steps S300 to S304, as follows: S300: Is the current neighboring terminal T in the candidate terminal dictionary of the vehicle? If yes, execute S301; if no, execute S302.

[0098] S301: Update the current co-occurrence features based on the observation data of the current observed event.

[0099] S302: Add the current neighboring terminal to the candidate terminal dictionary, and initialize the current co-occurrence feature based on the observation data of the current observation event.

[0100] S303: Input the current co-occurrence features into the association confidence model to obtain the current association confidence.

[0101] S304: Determine the association relationship between the vehicle and the current neighboring terminal based on the current association confidence level and confidence level threshold.

[0102] Furthermore, in order to continuously improve the accuracy of the confidence level, this application can use actual association feedback for online updates during operation. In one or more embodiments of this application, this application can obtain the true association confidence level between the current neighboring terminal and the vehicle, thereby determining the prediction error between the current association confidence level and the true association confidence level, and updating the association confidence level model based on the prediction error and the current co-occurrence characteristics.

[0103] For example, using a logistic regression model as the association confidence model, the functional form of this association confidence model is as follows:

[0104] In equation (15), It is the sigmoid function. It is the weight vector of the association confidence model. It is the bias term of the association confidence model. It is the normalized current co-occurrence feature, and C is the current association confidence.

[0105] The confidence level of the true association between the current nearby terminals and the vehicle is determined through subsequent trajectory verification and other methods. 0 represents no association, and 1 represents association. Therefore, an online learning approach is used to continuously optimize the model parameters of the association confidence model. ,as follows:

[0106] In equations (16) to (18), This represents the prediction error between the current association confidence level and the true association confidence level. The preset learning rate is used to control the step size of model parameter updates. It can be a fixed constant or it can decrease dynamically with the number of iterations. This application does not impose any restrictions on this. and These are the updated weight vector and bias term, respectively. By performing the above update after each real feedback is obtained, the model parameters can continuously approach the optimal decision boundary in an incremental manner, achieving adaptation to different scenarios, thereby continuously improving the prediction accuracy of association confidence without relying on complete historical retraining.

[0107] Figure 4 This is a schematic diagram illustrating the process of determining the association relationship and optimizing parameters online, as provided in the embodiments of this application. Figure 4 As shown, the process of determining the association relationship and optimizing the model online includes steps S400 to S405, as follows: S400: Is the current association confidence level greater than the confidence threshold? If yes, proceed to S401; if no, proceed to S402.

[0108] S401: The current nearest terminal is associated with the vehicle, output (VID, TID, C).

[0109] S402: Store the relevant information of the current neighboring terminal in the vehicle's waiting terminal dictionary and maintain the candidate status.

[0110] S403: Determine the true association confidence between the current neighboring terminal and the vehicle, and construct feedback samples based on the true association confidence, the current association confidence, and the current co-occurrence features.

[0111] S404: Based on feedback samples, update the model parameters and confidence threshold of the association confidence model online.

[0112] S405: Based on the updated association confidence model and confidence threshold, perform association determination for the next observed event.

[0113] In practical applications, the features mentioned above need to undergo normalization and missing value imputation after calculation to achieve feature standardization and ensure the stability of the model input. For features with fixed value ranges (such as co-occurrence intensity and role features), they can be directly linearly scaled to the [0,1] interval. For features with non-fixed value ranges (such as the distance mean involved in spatial proximity stability), an online algorithm is used to dynamically maintain their mean and standard deviation and perform Z-score standardization. For features that cannot be calculated due to insufficient data (such as route constraint rationality requiring two consecutive co-occurrences), historical rolling mean or a neutral value of 0.5 is used for imputation, and a binary flag is added to indicate the imputation status.

[0114] For example, if the current co-occurrence features include co-occurrence strength, spatial proximity stability, velocity consistency, role characteristics, route constraint rationality, and group consistency, then this current co-occurrence feature... Current model parameters , Then the linear part This leads to the current association confidence level. If subsequent feedback confirms a genuine correlation, then the confidence level for a genuine correlation is... Then the prediction error learning rate After the update, the model parameters were adjusted.

[0115] Furthermore, to ensure data security and user privacy, this application may also perform data preprocessing such as data anonymization on the acquired feature data to reduce the risk of information leakage. Data preprocessing includes, but is not limited to, processing vehicle identifiers (such as license plate numbers) and terminal identifiers (such as IMEI, IMSI, MAC addresses) using irreversible hash algorithms (such as SHA-256); blurring or gridding data involving location precision; not storing or transmitting raw biometric features (such as facial images), but only retaining abstract feature hash values; and setting data lifecycle management strategies to automatically clean up or anonymize expired data.

[0116] Furthermore, to automatically identify behaviors such as passenger transfers and vehicle platooning in dynamic traffic environments, the structured knowledge output by the system is enriched. In one or more embodiments of this application, the application can monitor and infer the spatiotemporal evolution patterns of associated confidence, as follows: Regarding transfer events: upon detecting the connection between the target terminal T and the first mode of transportation... The association confidence is reduced, and the target terminal and the second means of transportation are related. When the confidence level of the association increases and the first and second modes of transportation have a spatiotemporal intersection, the transfer hypothesis is triggered. Based on the historical co-occurrence characteristics of the target terminal with the first and second modes of transportation, the transfer hypothesis is verified. If the transfer hypothesis is valid, the target terminal, the first mode of transportation, and the second mode of transportation are marked as transfer events, the association between the target terminal and the first mode of transportation is removed, and the association between the target terminal and the second mode of transportation is established.

[0117] For example, T and The association confidence levels at three consecutive vehicle observation times showed a decreasing trend, and were all less than 0.5; T and The association confidence level at three consecutive times of vehicle observation showed an upward trend, and was greater than 0.7 in two consecutive instances. and They appear simultaneously at service areas or intersections, creating a spatiotemporal convergence point that triggers the transfer hypothesis.

[0118] For collaborative vehicle fleet events: when multiple vehicles are detected to be within a distance of a preset distance threshold or when the similarity between the group behavior templates of multiple vehicles is greater than a preset similarity threshold, multiple vehicles are marked as collaborative vehicle fleet events, a unique identifier is assigned to the fleet, and changes in fleet members are continuously monitored, outputting the unique identifier of the fleet and a list of fleet members.

[0119] For example, three freight vehicles, V1, V2, and V3, travel continuously on the same highway segment. At six consecutive checkpoints, the average distance between them remains within 60 meters (the preset distance threshold is 100 meters). The group behavior template for each vehicle is extracted, and the speed pattern similarity between each pair is calculated using Time-Depth Wrap (DTW). The DTW distances are 0.05, 0.06, and 0.04, respectively, all less than the preset threshold of 0.1. V1, V2, and V3 are marked as a cooperative fleet, assigned a fleet ID "C2025," and member changes are continuously monitored. When V2 leaves the fleet to refuel, increasing its distance from V1 to 120 meters and its speed pattern DTW distance to 0.35, V2 is automatically removed from the fleet, the fleet member list is updated to include V1 and V3, and the fleet ID "C2025" and the current members are output.

[0120] For driver identification and abnormal operation: When the role score R in the role characteristics (F4) of the terminal and vehicle is detected to be 1.0 or 0.8, and the terminal's behavior pattern matches the characteristics of a driver (such as frequent connection to in-vehicle Bluetooth, and the operation APP being constantly running), the terminal is identified as a high-probability driver terminal; when the vehicle type tag is detected as a private car, but the vehicle frequently associates with multiple different terminals, travels long distances outside of commuting hours, and the driver terminal and the vehicle's registered owner information do not match, an abnormal operation warning is triggered.

[0121] For detecting vehicles with cloned license plates: When the same license plate number is detected but the visual fingerprint of the vehicle is inconsistent, a cloned license plate warning is triggered, and the driving trajectory of the two vehicles and the associated terminal information are analyzed to output the vehicle identification of the cloned license plate and its related information.

[0122] In this embodiment, through advanced scenario processing, the system can be upgraded from simply associating terminals with vehicles to intelligently understanding personnel flow, vehicle collaboration, driving behavior, compliance, and safety incidents, providing higher-level knowledge support for traffic control, emergency dispatch, and other tasks.

[0123] Based on the aforementioned method for associating a vehicle with a terminal, this application also provides a specific embodiment of a device for associating a vehicle with a terminal.

[0124] like Figure 5 As shown, Figure 5 This is a schematic diagram of the structure of a vehicle-terminal association device provided in an embodiment of this application. The device includes an event triggering module 501, a co-occurrence feature module 502, and an association determination module 503.

[0125] The event triggering module 501 is used to determine one or more current neighboring terminals of the vehicle within a preset spatiotemporal proximity range in response to a current observation event of the vehicle. The co-occurrence feature module 502 is used to determine, for each of the current neighboring terminals, the current co-occurrence feature of the current neighboring terminal and the vehicle being observed simultaneously, based on the observation data of the current observation event; The association determination module 503 is used to determine the association relationship between each current neighboring terminal and the vehicle based on the current co-occurrence features and the scene features of the scene in which the vehicle is located.

[0126] In one feasible implementation, the association determination module is specifically used to determine a confidence threshold based on the scene features and a preset threshold mapping relationship. The scene features include at least one of road type features, traffic time period features, and traffic density features. For each current neighboring terminal, based on the current co-occurrence features, a preset association confidence model is used to obtain the current association confidence between the current neighboring terminal and the vehicle. For each current neighboring terminal, if the current association confidence is greater than the confidence threshold, then the current neighboring terminal and the vehicle are associated; if the current association confidence is not greater than the confidence threshold, then the current neighboring terminal and the vehicle are not associated.

[0127] In one feasible implementation, the current co-occurrence feature includes at least one of co-occurrence intensity, spatial proximity stability, speed consistency, role characteristics, route constraint rationality, and group consistency; the co-occurrence intensity characterizes the frequency at which the terminal and the vehicle are observed simultaneously; the spatial proximity stability characterizes the degree of fluctuation in the spatial distance between the terminal and the vehicle; the speed consistency characterizes the correlation between the movement speeds of the terminal and the vehicle; the role characteristics characterize the role of the terminal within the vehicle; the route constraint rationality characterizes the feasibility of the terminal and the vehicle under preset road network constraints; and the group consistency characterizes the similarity of movement behavior patterns between the terminal and associated terminals of multiple vehicles.

[0128] In one feasible implementation, the co-occurrence feature module is specifically used to determine whether the current neighboring terminal exists in a pre-built candidate terminal dictionary, which stores information about all terminals simultaneously observed with the vehicle; if so, the historical feature data of the current neighboring terminal and the vehicle being simultaneously observed is updated based on the observation data to obtain the current feature data; if not, the current neighboring terminal is added to the candidate terminal dictionary, and the current feature data is initialized based on the observation data of the current observation event; the current co-occurrence feature is determined based on the current feature data.

[0129] In one feasible implementation, the co-occurrence feature module can also be used to: 1) The historical feature data includes the historical mean distance and historical cumulative squared difference between the current neighboring terminal and the vehicle; 2) The current co-occurrence feature includes spatial proximity stability, which characterizes the fluctuation of the spatial distance between the terminal and the vehicle; 3) Determine the current distance between the current neighboring terminal and the vehicle in the current observation event; 4) Use an online recursive algorithm to update the historical mean distance based on the distance increment between the current distance and the historical mean distance to obtain the current mean distance; 5) Update the historical cumulative squared difference based on the distance increment and the current mean distance to obtain the current cumulative squared difference; 6) Determine the spatial proximity stability based on the current mean distance and the current cumulative squared difference.

[0130] In one feasible implementation, the co-occurrence feature module can also be used to include the historical feature data including the last co-occurrence time of the current neighboring terminal and the vehicle, the first co-occurrence position of the current neighboring terminal at the last co-occurrence time, and the second co-occurrence position of the vehicle at the last co-occurrence time; the current co-occurrence feature includes route constraint rationality, which characterizes the feasibility of the terminal and the vehicle under preset road network constraints; and, in the case that the current neighboring terminal and the vehicle are observed multiple times simultaneously, to obtain the third co-occurrence position of the current neighboring terminal at the current co-occurrence time and the fourth co-occurrence position of the vehicle at the current co-occurrence time. The process involves: determining the co-occurrence location; identifying the time interval between the current co-occurrence time and the previous co-occurrence time; determining the shortest path distance and shortest travel time for the current neighboring terminal based on the first and third co-occurrence locations under preset route constraints; determining the shortest path distance for the vehicle based on the second and fourth co-occurrence locations; determining travel feasibility based on the time interval and the shortest travel time for the current neighboring terminal; determining distance matching degree based on the difference between the shortest path distance for the current neighboring terminal and the shortest path distance for the vehicle; and determining the rationality of the route constraints based on the travel feasibility and the distance matching degree.

[0131] In one feasible implementation, the co-occurrence feature module can also be used to: include the historical feature data including the vehicle observation time and the first behavioral data of multiple associated terminals of the vehicle at the vehicle observation time; the current co-occurrence feature includes group consistency, which characterizes the similarity of the movement behavior patterns between the terminal and the multiple associated terminals of the vehicle; determine a group behavior template based on each of the first behavioral data; obtain the second behavioral data of the current neighboring terminal at the vehicle observation time; and determine the group consistency based on the similarity between the second behavioral data and the group behavior template.

[0132] In one feasible implementation, the co-occurrence feature module can also be used to obtain the true association confidence between the current neighboring terminal and the vehicle; determine the prediction error between the current association confidence and the true association confidence; and update the association confidence model based on the prediction error and the current co-occurrence feature.

[0133] In one feasible implementation, the association determination module can also be used to trigger a transfer hypothesis when the association confidence between the target terminal and the first vehicle is reduced, the association confidence between the target terminal and the second vehicle is increased, and the first and second vehicles have a spatiotemporal intersection point; verify whether the transfer hypothesis is valid based on the historical co-occurrence characteristics of the target terminal with the first and second vehicles respectively; if the transfer hypothesis is valid, mark the target terminal, the first vehicle, and the second vehicle as a transfer event; and mark the multiple vehicles as a cooperative vehicle fleet event when the distance between multiple vehicles is less than a preset distance threshold and the similarity between the group behavior templates of multiple vehicles is greater than a preset similarity threshold.

[0134] Figure 6 This illustration shows a hardware structure diagram of an association device between a vehicle and a terminal provided in an embodiment of this application.

[0135] An associated device for a vehicle and a terminal may include a processor 601 and a memory 602 storing computer program instructions.

[0136] Specifically, the processor 601 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0137] Memory 602 may include mass storage for data or instructions. For example, and not limitingly, memory 602 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. In one embodiment, memory 602 may include removable or non-removable (or fixed) media, or memory 602 may be non-volatile solid-state memory. Memory 602 may be internal or external to the integrated gateway disaster recovery device.

[0138] In one embodiment, memory 602 may be read-only memory (ROM). In one instance, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0139] Memory 602 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Therefore, generally, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the vehicle-terminal association method according to one aspect of this application.

[0140] The processor 601 reads and executes computer program instructions stored in the memory 602 to achieve... Figure 1 The method for associating vehicles and terminals in the illustrated embodiment.

[0141] In one example, the associated device between a vehicle and a terminal may further include a communication interface 603 and a bus 604. Wherein, as Figure 6 As shown, the processor 601, memory 602, and communication interface 603 are connected through bus 604 and complete communication with each other.

[0142] The communication interface 603 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0143] Bus 604 includes hardware, software, or both, that couples components of a vehicle to an associated device of a terminal. For example, and not as a limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 604 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0144] Furthermore, in conjunction with the method for associating a vehicle and a terminal in the above embodiments, this application embodiment can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the methods for associating a vehicle and a terminal in the above embodiments.

[0145] This application also provides a computer program product, including a computer program that, when executed by a processor, implements any of the methods for associating a vehicle with a terminal as described in the above embodiments.

[0146] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0147] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, read-only memory (ROM), flash memory, erasable read-only memory (EROM), floppy disks, compact disc read-only memory (CD-ROM), optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0148] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0149] The aspects of this application have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0150] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for associating a means of transportation with a terminal, characterized in that, include: In response to a current observation event of a vehicle, determine one or more current neighboring terminals of the vehicle within a preset spatiotemporal proximity range; For each of the current neighboring terminals, based on the observation data of the current observation event, determine the current co-occurrence feature of the current neighboring terminal and the vehicle being observed simultaneously; For each of the current neighboring terminals, the association between the current neighboring terminal and the vehicle is determined based on the current co-occurrence features and the scene features of the scene in which the vehicle is located.

2. The method according to claim 1, characterized in that, For each of the current neighboring terminals, the association between the current neighboring terminal and the vehicle is determined based on the current co-occurrence features and the scene features of the scene in which the vehicle is located. Specifically, this includes: Based on the scene features and the preset threshold mapping relationship, a confidence threshold is determined. The scene features include at least one of road type features, traffic time period features, and traffic density features. For each of the current neighboring terminals, based on the current co-occurrence features, the current association confidence between the current neighboring terminal and the vehicle is obtained through a preset association confidence model; For each of the current neighboring terminals, if the current association confidence level is greater than the confidence level threshold, then the current neighboring terminal is associated with the vehicle; if the current association confidence level is not greater than the confidence level threshold, then the current neighboring terminal is not associated with the vehicle.

3. The method according to claim 1, characterized in that, The current co-occurrence features include at least one of co-occurrence intensity, spatial proximity stability, speed consistency, role characteristics, route constraint rationality, and group consistency; the co-occurrence intensity characterizes the frequency at which the terminal and the vehicle are observed simultaneously; the spatial proximity stability characterizes the degree of fluctuation in the spatial distance between the terminal and the vehicle; the speed consistency characterizes the correlation between the movement speeds of the terminal and the vehicle; the role characteristics characterize the role of the terminal within the vehicle; the route constraint rationality characterizes the feasibility of the terminal and the vehicle under preset road network constraints; and the group consistency characterizes the similarity of movement behavior patterns between the terminal and associated terminals of multiple vehicles.

4. The method according to claim 1, characterized in that, Based on the observation data of the current observation event, determine the current co-occurrence characteristics of the current neighboring terminal and the vehicle being observed simultaneously, specifically including: Determine whether the current neighboring terminal exists in a pre-built candidate terminal dictionary, which is used to store information on all terminals that are observed simultaneously with the vehicle. If so, then based on the observation data, update the historical feature data of the current neighboring terminal and the vehicle being observed simultaneously to obtain the current feature data; If not, add the current neighboring terminal to the candidate terminal dictionary and initialize the current feature data based on the observation data of the current observation event; Based on the current feature data, determine the current co-occurrence feature.

5. The method according to claim 4, characterized in that, The historical feature data includes the historical mean distance and historical cumulative squared difference between the current neighboring terminal and the vehicle; the current co-occurrence feature includes spatial proximity stability, which characterizes the degree of fluctuation in the spatial distance between the terminal and the vehicle; Based on the observation data of the current observation event, determine the current co-occurrence characteristics of the current neighboring terminal and the vehicle being observed simultaneously, specifically including: Determine the current distance between the current nearest terminal and the vehicle in the current observation event; An online recursive algorithm is used to update the historical average distance based on the distance increment between the current distance and the historical average distance, so as to obtain the current average distance. Based on the distance increment and the current distance mean, update the historical cumulative squared difference to obtain the current cumulative squared difference; The spatial proximity stability is determined based on the current mean distance and the current cumulative squared difference.

6. The method according to claim 4, characterized in that, The historical feature data includes the last co-occurrence time of the current neighboring terminal and the vehicle, the first co-occurrence position of the current neighboring terminal at the last co-occurrence time, and the second co-occurrence position of the vehicle at the last co-occurrence time; the current co-occurrence feature includes route constraint rationality, which characterizes the feasibility of the terminal and the vehicle under preset road network constraints; Based on the observation data of the current observation event, determine the current co-occurrence characteristics of the current neighboring terminal and the vehicle being observed simultaneously, specifically including: In cases where the current neighboring terminal and the vehicle are observed simultaneously multiple times, the third co-occurrence position of the current neighboring terminal at the current co-occurrence time and the fourth co-occurrence position of the vehicle at the current co-occurrence time are obtained. Determine the time interval between the current co-occurrence time and the previous co-occurrence time; Under preset route constraints, the shortest path distance and shortest travel time of the current neighboring terminal are determined based on the first co-occurrence location and the third co-occurrence location; and the shortest path distance of the vehicle is determined based on the second co-occurrence location and the fourth co-occurrence location. The feasibility of passage is determined based on the time interval and the shortest passage time of the current neighboring terminal; The distance matching degree is determined based on the difference between the shortest path distance of the current nearest terminal and the shortest path distance of the vehicle; The rationality of the route constraints is determined based on the described feasibility of passage and the described distance matching degree.

7. The method according to claim 4, characterized in that, The historical feature data includes the vehicle observation time and the first behavior data of multiple associated terminals of the vehicle at the vehicle observation time; the current co-occurrence feature includes group consistency, which characterizes the similarity of the movement behavior patterns between the terminal and the associated terminals of multiple vehicles. Based on the observation data of the current observation event, determine the current co-occurrence characteristics of the current neighboring terminal and the vehicle being observed simultaneously, specifically including: Based on the first behavior data, determine the group behavior template; Obtain the second behavioral data of the currently nearby terminal at the time of observation of the vehicle; The group consistency is determined based on the similarity between the second behavioral data and the group behavior template.

8. The method according to claim 2, characterized in that, After obtaining the current association confidence between the current neighboring terminal and the vehicle based on the current co-occurrence characteristics and a preset association confidence model, the method further includes: Obtain the true association confidence level between the current nearest terminal and the vehicle; Determine the prediction error between the current association confidence level and the true association confidence level; The association confidence model is updated based on the prediction error and the current co-occurrence features.

9. The method according to claim 7, characterized in that, The method further includes: The transfer hypothesis is triggered when the confidence level of the association between the target terminal and the first mode of transportation decreases, the confidence level of the association between the target terminal and the second mode of transportation increases, and the first mode of transportation and the second mode of transportation have a spatiotemporal intersection point. Based on the historical co-occurrence characteristics of the target terminal with the first mode of transportation and the second mode of transportation, verify whether the transfer hypothesis is valid. If the transfer assumption holds true, the target terminal, the first vehicle, and the second vehicle are marked as a transfer event; If multiple vehicles are detected to be within a distance of a preset distance threshold, or if the similarity between the group behavior templates of the multiple vehicles is greater than a preset similarity threshold, the multiple vehicles will be marked as a collaborative vehicle fleet event.

10. A computer-readable storage medium, characterized in that, The computer storage medium stores computer program instructions, which, when executed by a processor, implement the method for associating a vehicle with a terminal as described in any one of claims 1-9.

Images