A method for representing highway vehicle driving identity information

Abstract

The application discloses a highway vehicle driving identity information representation method, constructs a driving identity index system containing vehicle relative characteristic indexes and vehicle dynamic characteristic indexes, combines multi-dimensional data acquisition and an improved matter-element extension model, and realizes accurate representation of vehicle identity. Specifically, the method comprises the following steps: collecting highway vehicle driving data, constructing a highway global historical data set and a vehicle historical data set; designing double-layer section matter elements and classical domain matter elements, and optimizing characteristic evaluation; calculating a vehicle driving characteristic comprehensive evaluation index value through a correlation coefficient, quantifying the correlation degree of a target vehicle and historical data, and combining a preset threshold to judge the identity identity of the vehicle. The application has the advantages of strong interpretability and low complexity by fusing the matter-element extension theory and multi-dimensional data analysis, can effectively identify escape behaviors such as fake license plates and abnormal trajectories, and provides technical support for intelligent monitoring and auditing of highways.

Description

Technical Field

[0001] This invention relates to the technical field of intelligent transportation, and in particular to a method for representing the identity information of vehicles traveling on highways. Background Technology

[0002] With the continuous growth of highway traffic flow, the accurate representation and identification of vehicle identity information has become an important issue in intelligent traffic management. Traditional vehicle identification methods often rely on single-dimensional information, such as license plate recognition or vehicle model characteristics, which are difficult to cope with complex and ever-changing real-world application scenarios. Especially when facing toll evasion behaviors such as license plate replacement and cloned plates, traditional methods have high false positive and false negative rates. To address this, this study proposes a highway vehicle driving identity information representation method based on an improved matter-element extension model. The aim is to construct a vehicle identification system that reflects the unique driving characteristics of different vehicles under different traffic environments and weather conditions by comprehensively analyzing multi-dimensional characteristic indicators of vehicle driving trajectories. Specifically, this study first constructs a driving identity index system that describes vehicle driving identity information from two perspectives: relative features and dynamic features. All-weather highway vehicle driving data is collected by sensors installed on highways, and the values ​​of each index in the driving identity index system are calculated, thereby constructing a global historical dataset of highways and a historical dataset of vehicles. Based on this, a two-layer matter-element model is constructed to evaluate and analyze the feature indicators in the dataset. A comprehensive evaluation index of vehicle driving characteristics is constructed to evaluate and quantify the correlation between the target vehicle and the dataset, representing the driving identity information of the target vehicle. This process not only helps managers quickly identify the unique attributes of vehicles, but also discovers abnormal behavior patterns by comparing historical data, such as license plate changes and abnormal driving trajectories, providing strong technical support for combating highway toll evasion. This method combines the flexibility of matter-element extension theory with the advantages of multidimensional data analysis, possessing strong universality and scalability, and providing new research ideas and technical means for the fields of intelligent traffic management and vehicle identification. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings and deficiencies of the prior art and provide a method for representing the driving identity information of highway vehicles. This method can represent the unique driving identity information of different vehicles based on their driving data. Compared with machine learning methods, the method of this invention has stronger interpretability, lower complexity, and lower training difficulty.

[0004] To achieve the above objectives, the technical solution provided by this invention is: a method for representing the identity information of vehicles traveling on highways, comprising the following steps:

[0005] S1: Considering vehicle performance, driver style, weather influence, surrounding traffic conditions and the spatiotemporal distribution of nearby vehicles, a driving identity index system that can describe vehicle driving identity information is constructed from the perspectives of relative characteristics and dynamic characteristics.

[0006] S2: Collect highway vehicle driving data, calculate the values ​​of each indicator in the driving identity indicator system, construct a global historical dataset of highways based on the values ​​of each indicator calculated from all-weather data of different years, months and times, and construct a vehicle historical dataset by filtering the same vehicle data in the global historical dataset of highways.

[0007] S3: Construct a highway vehicle identity information representation model. This model addresses the issue that traditional matter-element extension models do not consider differences in historical feature data when solving incompatible problems through quantitative evaluation. It improves the section-domain and classical-domain matter-element models in the matter-element extension model, forming a two-layer matter-element model. This includes highway section-domain matter-element and classical-domain matter-element models constructed using a global highway historical dataset, and vehicle section-domain matter-element and classical-domain matter-element models constructed using a vehicle historical dataset. This two-layer matter-element model enhances the model's ability to capture, identify, and represent vehicle identity information by considering both the global highway historical dataset and the vehicle historical dataset. A comprehensive evaluation index for vehicle driving characteristics is constructed using segment-domain objects and classical domain objects to characterize the driving identity information of the target vehicle. The comprehensive evaluation index for vehicle driving characteristics is a weighted value of the correlation function between the object to be evaluated and the bilayer segment-domain objects and classical domain objects. The object to be evaluated is constructed from the index values ​​of the driving identity index system calculated using the vehicle driving data of the target vehicle. The weighted value is the optimal weight of each index in the driving identity index system trained by the vehicle historical dataset. The comprehensive evaluation index for vehicle driving characteristics characterizes the driving identity information of the target vehicle by quantifying the correlation between the driving characteristic data of the target vehicle and the global historical dataset of the highway and the historical dataset of the vehicle.

[0008] S4: Calculate the values ​​of each indicator in the driving identity indicator system using the target vehicle driving data, input the highway vehicle identity information representation model, and output the comprehensive evaluation index value of vehicle driving characteristics to represent the highway vehicle driving identity information. By comparing the absolute value of the difference between the comprehensive evaluation index values ​​of different vehicles with the preset threshold, the probability of different vehicles belonging to the same vehicle can be determined based on their driving data.

[0009] Furthermore, the specific steps of step S1 are as follows:

[0010] S11: From the perspective of vehicle driving trajectory, describe the heterogeneity of vehicle driving characteristics under different traffic environments, and construct vehicle relative characteristic indicators, including speed consistency indicators, acceleration consistency indicators, and gap distance consistency indicators.

[0011] S12: Construct a highway vehicle distribution network to characterize vehicle clustering on the highway network. Use the degree centrality of vehicle nodes in the highway vehicle distribution network as the node weight to reflect the importance of vehicles in road segment traffic. Define the sensor detection range as the research interval, where the sensor detection range is defined as the range within which vehicle driving and appearance information can be obtained before and after the sensor. Using vehicles in the research interval as nodes, establish edges to connect vehicles traveling in the same direction, constructing an undirected network D. veh_network = {N, L}; Let the set of nodes be N = {n1, n2, ..., n}. i ,...,n M}, where node i in the set is denoted as n. i The number of nodes is M; L = {(n i ,n l The set of edges (n) is defined as follows: |i,l=1,2,...,M,i≠l}. i ,n l () represents an undirected edge connecting node i and node l, and is used to connect nodes in the highway vehicle distribution network whose absolute distance is less than a preset distance threshold d. pre Nodes:

[0012]

[0013] In the formula, x i x l Let y represent the longitudinal velocities of nodes i and l along the road travel direction, respectively. i y l These represent the lateral velocities of node i and node l perpendicular to the road travel direction, respectively.

[0014] S13: Based on complex network theory, calculate the degree centrality of the highway vehicle distribution network as vehicle weights to describe the importance of clustered vehicle groups in traffic flow and the low influence of discrete vehicles in traffic flow. The degree centrality ω of node i is also considered. cen,i Defined as the ratio of node degree to the number of nodes, i.e.:

[0015]

[0016] In the formula, k i Let be the degree of node i, and let represent the number of edges in the network that connect to node i.

[0017] S14: Building Speed ​​Consistency Metrics To characterize the degree of deviation of vehicle speed from the overall traffic flow speed, reflecting the vehicle's speed tendency in the traffic flow environment, that is:

[0018]

[0019] In the formula, v i (t) represents the velocity of node i at time t. The speed is the weighted average speed of the traffic flow, with the weight being the degree centrality ω of the node. cen,i With vehicle attribute ω type,i The weight of vehicle attributes depends on the vehicle type, describing the degree of influence of different types of vehicles on traffic flow speed;

[0020] S15: Constructing Acceleration Consistency Metrics To quantify the difference in acceleration and deceleration frequencies between a vehicle and its surrounding traffic flow, the method characterizes the degree of deviation between the acceleration fluctuations of a vehicle during its driving process and the acceleration fluctuations of the overall traffic flow.

[0021]

[0022] In the formula, a i (t) represents the acceleration of node i at time t. Let t be the weighted average acceleration of traffic flow at node i at time t. s and t e These represent the times when the vehicle enters and leaves the sensor's observation and research range, respectively.

[0023] S16: Construct a gap distance consistency index To characterize the difference between vehicle following distance and the overall spatial distribution pattern of traffic flow, reflecting the vehicle's preference for maintaining following distance in different traffic density environments, i.e.:

[0024]

[0025] In the formula, g i (t) represents the gap distance of node i at time t. This is the weighted average gap distance for traffic flow;

[0026] S17: Construct vehicle dynamic characteristic indices to describe vehicle dynamic characteristics from the perspective of vehicle performance and operating characteristics. Incorporate the influence of weather and highway road environment to describe the heterogeneity of vehicle dynamic driving characteristics, including speed characteristic indices. Gap distance characteristic index Vehicle maximum acceleration index Lane change frequency index and conflict frequency indicators

[0027] S18: Constructing speed characteristic indicators This characterizes the speed change of vehicles within the observation and study range, as well as the difference between the maximum and minimum speeds, taking into account the influence of weather conditions and road friction on the dynamic speed of vehicles.

[0028]

[0029] In the formula, θ represents the intensity of weather conditions, f(θ) is the influence function of weather conditions on driving behavior, σ is the road friction coefficient, and max(v) is the maximum friction coefficient. i (t)) and min(v i (t) represents the maximum and minimum velocities of node i within the observation range, respectively;

[0030] S19: Constructing gap distance characteristic indicators This characterizes the changes in following distance and the difference between the maximum and minimum following distance as the vehicle enters the observation and research range, taking into account the influence of weather conditions and road friction on the following distance maintained by the vehicle during normal driving.

[0031]

[0032] In the formula, max(g) i (t)) and min(g) i (t) represents the maximum and minimum gap distances of node i within the observation range, respectively;

[0033] S110: Constructing the maximum acceleration index for vehicles Describe the heterogeneity of vehicles in terms of performance and driving style, namely:

[0034]

[0035] In the formula, This represents the absolute value of the maximum acceleration of node i within the observation range;

[0036] S111: Constructing a lane change frequency indicator Conflict frequency index This indicates the number of lane changes and traffic conflict incidents during vehicle operation, quantifying driving style and driver psychology. A cautious driving style tends to stay in the same lane and is less likely to cause traffic conflicts; an impulsive driving style maintains a small following distance and frequently changes lanes to increase speed.

[0037]

[0038] In the formula, Num(·) represents the indicator function. This represents the time of the j-th lane change at node i. The observation time [t] of node i was statistically analyzed. s ,t e The number of lane changes within ];

[0039]

[0040] In the formula, TTC i Let Num(·) represent the collision time of node i, and let Num(·) represent the indicator function. This function counts the number of collisions of node i during the observation and research time [t]. s ,t e The number of events with an internal collision time of less than 2.6 seconds;

[0041] The indicator function Num(·) is used to determine whether an internal condition is true, that is:

[0042]

[0043] Time-to-Collision (TTC) is a classic indicator in traffic safety research, characterizing the time between collisions when vehicles maintain their speeds, thus quantifying the risk of a rear-end collision.

[0044]

[0045] In the formula, x front and v front Let x represent the position and speed of the car ahead of node i, respectively. front -x i This represents the relative distance between node i and the vehicle in front of it;

[0046] S112: Standardize the eight characteristic indicators, including the relative and dynamic characteristic indicators of the vehicle, to eliminate the influence of dimensional differences:

[0047]

[0048] In the formula, I jk and I' jk Let μ represent the j-th index value and the standardized value in sample k, respectively. j and σ j are the mean and standard deviation of the j-th feature index, respectively;

[0049] S113: Based on the standardized indicators, construct a driving identity indicator system I to describe vehicle driving identity information, denoted as:

[0050]

[0051] In the formula, I j This represents the j-th indicator in the driving identity indicator system.

[0052] Furthermore, the specific steps of step S2 are as follows:

[0053] S21: Based on the highway network system and geographic information system, and sensors spaced at intervals along the basic highway sections, ramp entrances and exits, and merging zones, collect all-weather data for different years, months, and times, including vehicle trajectory information, weather conditions, vehicle snapshots, and highway vehicle driving data from gantry transaction logs within the sensor coverage area. Record this as U segments of highway vehicle driving data. Based on the collected data, calculate all indicators in the driving identity indicator system, denoted as the highway global historical dataset R. highway ={r1,r2,...,r a ,...,r U}, where r a =(r1,r2,...,r j (r, ..., r8) represents the calculated value of the driving identity index system corresponding to the vehicle driving data of segment a of the highway in the global historical dataset of highways, where r... j This represents the value of the j-th indicator;

[0054] S22: Identify license plate numbers, vehicle models, body colors, and gantry transaction data in captured vehicle images; filter for data on the same vehicle within the overall historical highway dataset, denoted as vehicle history dataset R. vehicle ={r1,r2,...,r P},R vehicle ∈R highway P records containing the same vehicle, r P This represents the calculated value of the driving identity index system corresponding to the highway vehicle driving data in the Pth record of the vehicle history dataset.

[0055] Furthermore, the specific steps of step S3 are as follows:

[0056] S31: Taking the index value of the driving identity index system corresponding to any segment of highway driving data in the global historical dataset of highways as the target vehicle research object, denoted as the object element to be evaluated r in the highway vehicle identity information representation model. x The driving information of the target vehicle is represented by r. x =(r x1 ,r x2 ,...,r xj ,...,r x8 ), where r xj The value of the j-th index in the object to be evaluated;

[0057] S32: Based on the global historical dataset of highways and the historical dataset of vehicles, construct highway section-domain objects and vehicle section-domain objects respectively. Define the value range of the j-th index in the highway section-domain object as follows: Recorded as and These represent the corresponding lower and upper limits of values; the value range of the j-th index in the vehicle segment element is defined as follows: Recorded as and These are the corresponding lower and upper limits of the value range;

[0058] S33: Divide the highway section-domain material element and the vehicle section-domain material element into s classical domain material elements at equal intervals. Define the value range of the j-th index in the q-th highway classical domain material element as follows: Recorded as and These represent the corresponding lower and upper limits of values; the value range of the j-th index in the q-th vehicle classical domain matter element is defined as follows: Recorded as and These are the corresponding lower and upper limits of values:

[0059]

[0060] in, and These represent the distances between classical domain objects of each highway and the distances between classical domain objects of each vehicle, respectively.

[0061]

[0062] In the formula, and Let be the minimum and maximum values ​​of the j-th index in the classical domain matter element of the highway, respectively. and These are the minimum and maximum values ​​of the j-th index in the classic domain object element of the vehicle, respectively;

[0063] S34: Using the vehicle history dataset to train the model, the optimal weights of each indicator in the driving identity indicator system are obtained. First, a target function loss is defined to search for the optimal weight allocation. Based on the comprehensive evaluation index of vehicle driving characteristics, a target function is constructed to reflect the differences between the same vehicle driving characteristics calculated from road traffic information obtained at different observation times or locations. The smaller the target function, the better it demonstrates the performance of the highway vehicle identity information representation model.

[0064]

[0065] In the formula, G yc and G yvLet represent the comprehensive evaluation index of vehicle driving characteristics calculated from the c-th and v-th records in the vehicle history dataset, respectively, where pi represents the pi-th record in the vehicle history dataset, and the same vehicle contains P records;

[0066] S35: Calculate the optimal weights of each indicator in the driving identity indicator system that satisfy the minimization of the objective function. Right now:

[0067]

[0068] In the formula, ω j Let be the weight of the j-th indicator in the driving identity indicator system;

[0069] S36: Calculate the object element r to be evaluated x Correlation coefficient between highway section domain and classical domain matter elements Right now:

[0070]

[0071] In the formula, R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the q-th classical domain matter element of the highway The correlation coefficient between them R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the highway section element The correlation coefficient between them is calculated as follows:

[0072]

[0073] S37: Calculate the object element r to be evaluated x Correlation coefficient between vehicle domain and classical domain matter elements Right now:

[0074]

[0075] In the formula, R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the q-th vehicle classical domain object element The correlation coefficient between them R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the highway section element The correlation coefficient between them is calculated as follows:

[0076]

[0077] S38: Based on the correlation between the object to be evaluated and the two-layer object, the optimal weights obtained from training on the vehicle historical dataset are used as the weights of the correlation coefficient to calculate the comprehensive evaluation index S of vehicle driving characteristics, so as to quantify the heterogeneity of vehicle driving characteristics and thus characterize the driving identity information of the target vehicle, that is:

[0078]

[0079] In the formula, This represents the optimal weight of the j-th indicator in the driving identity indicator system.

[0080] Furthermore, the specific steps of step S4 are as follows:

[0081] S41: Calculate the values ​​of each indicator in the driving identity indicator system based on the driving information of the target vehicle, input the indicator values ​​into the highway vehicle identity information representation model, and output the comprehensive evaluation index value of the vehicle driving characteristics of the target vehicle.

[0082] S42: Using the comprehensive evaluation index value of vehicle driving characteristics as a representation method for highway vehicle identity information, the probability P of different vehicles belonging to the same vehicle is determined by the ratio between the absolute value of the difference between the comprehensive evaluation index values ​​of different vehicles and a preset threshold, based on the driving data of different vehicles. id ,Right now:

[0083]

[0084] In the formula, S z1 and S z2 S represents the comprehensive evaluation index value of vehicle driving characteristics calculated using the driving information of target vehicle z1 and vehicle z2, respectively. pre This is a preset threshold.

[0085] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0086] 1. This invention constructs a driving identity index system based on vehicle relative characteristic index and dynamic characteristic index, which can describe the unique driving characteristics of different vehicles from the perspectives of traffic flow environment and vehicle's own operating characteristics.

[0087] 2. This invention designs a two-layer object element, which constructs highway, vehicle section-domain object elements and classic domain object elements respectively through the global historical dataset of highways and the historical dataset of vehicles, thereby improving the model's ability to capture, identify and represent vehicle driving identity information.

[0088] 3. The comprehensive evaluation index of vehicle driving characteristics constructed by this invention can effectively characterize the driving identity information of the target vehicle, and further determine the probability that they belong to the same vehicle based on the driving data of different vehicles.

[0089] In summary, this invention, based on the theory of matter-element extensions, innovatively constructs a method for representing the driving identity information of vehicles on highways. It constructs highway and vehicle section-domain matter elements and classical domain matter elements respectively using a global historical dataset of highways and a historical dataset of vehicles, achieving the fusion of various indicators in the driving identity indicator system and outputting a comprehensive evaluation index value of vehicle driving characteristics to represent different vehicle driving features. This invention helps to compensate for the insufficient accuracy of highway license plate recognition, representing vehicle identity based on data such as driving trajectories, thereby improving the detection rate of toll evasion behaviors such as license plate changing, and providing an effective means for highway monitoring and management work such as toll evasion auditing and vehicle abnormality monitoring. Attached Figure Description

[0090] Figure 1 This is a framework diagram of the method of the present invention.

[0091] Figure 2 This is a schematic diagram of the highway vehicle distribution network of the present invention. Detailed Implementation

[0092] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0093] like Figure 1 As shown in the figure, this embodiment discloses a method for representing the identity information of vehicles traveling on highways, the specific details of which are as follows:

[0094] S1: Considering factors such as vehicle performance, driver style, weather conditions, surrounding traffic conditions, and the spatiotemporal distribution of nearby vehicles, a driving identity index system that can describe vehicle driving identity information is constructed from both relative and dynamic perspectives. The specific operation steps are as follows:

[0095] S11: From the perspective of vehicle driving trajectory, describe the heterogeneity of vehicle driving characteristics under different traffic environments, and construct vehicle relative characteristic indicators, including speed consistency indicators, acceleration consistency indicators, and gap distance consistency indicators.

[0096] S12: Construct a highway vehicle distribution network, such as Figure 2 As shown, this paper characterizes the vehicle clustering phenomenon on the highway network. The degree centrality of vehicle nodes in the highway vehicle distribution network is used as the node weight to reflect the importance of vehicles in the road segment traffic. The sensor detection range is defined as the research interval, encompassing 500 meters before and after the sensor, allowing the acquisition of vehicle driving and appearance information. Vehicles within the research interval are treated as nodes, and edges are set up for vehicles traveling in the same direction, constructing an undirected network D. veh_network = {N, L}; Let the set of nodes be N = {n1, n2, ..., n}.i ,...,n M}, where node i in the set is denoted as n. i The number of nodes is M; L = {(n i ,n l The set of edges (n) is defined as follows: |i,l=1,2,...,M,i≠l}. i ,n l () represents an undirected edge connecting node i and node l, and is used to connect nodes in the highway vehicle distribution network whose absolute distance is less than a preset distance threshold d. pre Nodes:

[0097]

[0098] In the formula, x i x l Let y represent the longitudinal velocities of nodes i and l along the road travel direction, respectively. i y l These represent the lateral velocities of node i and node l perpendicular to the road travel direction, respectively.

[0099] S13: Based on complex network theory, calculate the degree centrality of the highway vehicle distribution network as vehicle weights to describe the importance of clustered vehicle groups in traffic flow and the low influence of discrete vehicles in traffic flow. The degree centrality ω of node i is also considered. cen,i Defined as the ratio of node degree to the number of nodes, i.e.:

[0100]

[0101] In the formula, k i Let be the degree of node i, and let represent the number of edges in the network that connect to node i.

[0102] S14: Building Speed ​​Consistency Metrics To characterize the degree of deviation of vehicle speed from the overall traffic flow speed, reflecting the vehicle's speed tendency in the traffic flow environment, that is:

[0103]

[0104] In the formula, v i (t) represents the velocity of node i at time t. The speed is the weighted average speed of the traffic flow, with the weight being the degree centrality ω of the node. cen,i With vehicle attribute ω type,i The weight of vehicle attributes depends on the vehicle type, describing the degree of influence of different types of vehicles on traffic flow speed. For example, heavy vehicles are weighted at 1.5, and cars at 1.0.

[0105] S15: Constructing Acceleration Consistency Metrics To quantify the difference in acceleration and deceleration frequencies between a vehicle and its surrounding traffic flow, the method characterizes the degree of deviation between the acceleration fluctuations of a vehicle during its driving process and the acceleration fluctuations of the overall traffic flow.

[0106]

[0107] In the formula, a i (t) represents the acceleration of node i at time t. Let t be the weighted average acceleration of traffic flow at node i at time t. s and t e These represent the times when the vehicle enters and leaves the sensor's observation and research range, respectively.

[0108] S16: Construct a gap distance consistency index To characterize the difference between vehicle following distance and the overall spatial distribution pattern of traffic flow, reflecting the vehicle's preference for maintaining following distance in different traffic density environments, i.e.:

[0109]

[0110] In the formula, g i (t) represents the gap distance of node i at time t. This is the weighted average gap distance for traffic flow;

[0111] S17: Construct vehicle dynamic characteristic indices to describe vehicle dynamic characteristics from the perspective of vehicle performance and operating characteristics. Incorporate the influence of weather and highway road environment to describe the heterogeneity of vehicle dynamic driving characteristics, including speed characteristic indices. Gap distance characteristic index Vehicle maximum acceleration index Lane change frequency index and conflict frequency indicators

[0112] S18: Constructing speed characteristic indicators This characterizes the speed change of vehicles within the observation and study range, as well as the difference between the maximum and minimum speeds, taking into account the influence of weather conditions and road friction on the dynamic speed of vehicles.

[0113]

[0114] In the formula, θ represents the intensity of weather conditions, such as rainfall and visibility; f(θ) is the influence function of weather conditions on driving behavior; σ is the road friction coefficient, with a recommended value of 1.0 for ordinary asphalt roads; max(v i (t)) and min(v i (t) represents the maximum and minimum velocities of node i within the observation range, respectively;

[0115] S19: Constructing gap distance characteristic indicators This characterizes the changes in following distance and the difference between the maximum and minimum following distance as the vehicle enters the observation and research range, taking into account the influence of weather conditions and road friction on the following distance maintained by the vehicle during normal driving.

[0116]

[0117] In the formula, max(g) i (t)) and min(g) i (t) represents the maximum and minimum gap distances of node i within the observation range, respectively;

[0118] S110: Constructing the maximum acceleration index for vehicles Describe the heterogeneity of vehicles in terms of performance and driving style, namely:

[0119]

[0120] In the formula, This represents the absolute value of the maximum acceleration of node i within the observation range;

[0121] S111: Constructing a lane change frequency indicator Conflict frequency index This indicates the number of lane changes and traffic conflict incidents during vehicle operation, quantifying driving style and driver psychology. A cautious driving style tends to stay in the same lane and is less likely to cause traffic conflicts; an impulsive driving style maintains a small following distance and frequently changes lanes to increase speed.

[0122]

[0123] In the formula, Num(·) represents the indicator function. This represents the time of the j-th lane change at node i. The observation time [t] of node i was statistically analyzed. s ,t e The number of lane changes within ];

[0124]

[0125] In the formula, TTC i Let Num(·) represent the collision time of node i, and let Num(·) represent the indicator function. This function counts the number of collisions of node i during the observation and research time [t]. s ,t e The number of events with an internal collision time of less than 2.6 seconds;

[0126] The indicator function Num(·) is used to determine whether an internal condition is true, that is:

[0127]

[0128] Time-to-Collision (TTC) is a classic indicator in traffic safety research, characterizing the time between collisions when vehicles maintain their speeds, thus quantifying the risk of a rear-end collision.

[0129]

[0130] In the formula x front and v front Let x represent the position and speed of the car ahead of node i, respectively. front -x i This represents the relative distance between node i and the vehicle in front of it;

[0131] S112: Standardize the eight characteristic indicators, including the relative and dynamic characteristic indicators of the vehicle, to eliminate the influence of dimensional differences:

[0132]

[0133] In the formula, I jk and I' jk Let μ represent the j-th index value and the standardized value in sample k, respectively. j and σ j are the mean and standard deviation of the j-th feature index, respectively;

[0134] S113: Based on the standardized indicators, construct a driving identity indicator system I to describe vehicle driving identity information, denoted as:

[0135]

[0136] In the formula, I j This represents the j-th indicator in the driving identity indicator system.

[0137] Table 1. Driving Identity Index System

[0138]

[0139]

[0140] S2: Collect highway vehicle driving data, calculate the values ​​of each indicator in the driving identity indicator system, construct a global historical dataset of highways based on the indicator values ​​calculated from all-weather data of different years, months, and times, and filter the data of the same vehicle in the global historical dataset of highways to construct a vehicle historical dataset. The specific operation steps are as follows:

[0141] S21: Based on the highway network system and geographic information system, and sensors spaced at intervals along the basic highway sections, ramp entrances and exits, and merging zones, collect all-weather data for different years, months, and times, including vehicle trajectory information, weather conditions, vehicle snapshots, and highway vehicle driving data from gantry transaction logs within the sensor coverage area. Record this as U segments of highway vehicle driving data. Based on the collected data, calculate all indicators in the driving identity indicator system, denoted as the highway global historical dataset R. highway ={r1,r2,...,r a ,...,r U}, where r a =(r1,r2,...,r j (r, ..., r8) represents the calculated value of the driving identity index system corresponding to the vehicle driving data of segment a of the highway in the global historical dataset of highways, where r... j This represents the value of the j-th indicator;

[0142] S22: Identify license plate numbers, vehicle models, body colors, and gantry transaction data in captured vehicle images; filter for data on the same vehicle within the overall historical highway dataset, denoted as vehicle history dataset R. vehicle ={r1,r2,...,r P},R vehicle ∈R highway P records containing the same vehicle, r P This represents the calculated value of the driving identity index system corresponding to the highway vehicle driving data in the Pth record of the vehicle history dataset.

[0143] S3: Construct a highway vehicle identity information representation model. This model addresses the issue that traditional matter-element extension models do not consider differences in historical feature data when solving incompatible problems through quantitative evaluation. It improves the section-domain matter-element and classical-domain matter-element models by forming a two-layer matter-element model. This includes highway section-domain matter-element and classical-domain matter-element models constructed using a global highway historical dataset, and vehicle section-domain matter-element and classical-domain matter-element models constructed using a vehicle historical dataset. This two-layer matter-element model enhances the model's ability to capture, identify, and represent vehicle identity information by considering both the global highway historical dataset and the vehicle historical dataset. Based on this two-layer section-domain matter-element model... A comprehensive evaluation index for vehicle driving characteristics is constructed using classical domain objects to characterize the driving identity information of the target vehicle. This index is a weighted sum of the correlation functions between the object to be evaluated and the bilayer node domain objects and classical domain objects. The object to be evaluated is constructed from the index values ​​of a driving identity index system calculated using the target vehicle's driving data. The weighting values ​​are the optimal weights of each index in the driving identity index system, trained using a vehicle historical dataset. The comprehensive evaluation index for vehicle driving characteristics characterizes the driving identity information of the target vehicle by quantifying the correlation between the target vehicle's driving characteristic data and the highway global historical dataset and the vehicle historical dataset. The specific operation steps are as follows:

[0144] S31: Taking the index value of the driving identity index system corresponding to any segment of highway driving data in the global historical dataset of highways as the target vehicle research object, denoted as the object element to be evaluated r in the highway vehicle identity information representation model. x The driving information of the target vehicle is represented by r. x =(r x1 ,r x2 ,...,r xj ,...,r x8 ), where r xj The value of the j-th index in the object to be evaluated;

[0145] S32: Based on the global historical dataset of highways and the historical dataset of vehicles, construct highway section-domain objects and vehicle section-domain objects respectively. Define the value range of the j-th index in the highway section-domain object as follows: Recorded as and These represent the corresponding lower and upper limits of values; the value range of the j-th index in the vehicle segment element is defined as follows: Recorded as and These are the corresponding lower and upper limits of the value range;

[0146] S33: Divide the highway section-domain material element and the vehicle section-domain material element into s classical domain material elements at equal intervals. Define the value range of the j-th index in the q-th highway classical domain material element as follows: Recorded as and These represent the corresponding lower and upper limits of values; the value range of the j-th index in the q-th vehicle classical domain matter element is defined as follows: Recorded as and These are the corresponding lower and upper limits of values:

[0147]

[0148] in, and These represent the distances between classical domain objects of each highway and the distances between classical domain objects of each vehicle, respectively.

[0149]

[0150] In the formula, and Let be the minimum and maximum values ​​of the j-th index in the classical domain matter element of the highway, respectively. and These are the minimum and maximum values ​​of the j-th index in the classic domain object element of the vehicle, respectively;

[0151] S34: Using the vehicle history dataset to train the model, the optimal weights of each indicator in the driving identity indicator system are obtained. First, a target function loss is defined to search for the optimal weight allocation. Based on the comprehensive evaluation index of vehicle driving characteristics, a target function is constructed to reflect the differences between the same vehicle driving characteristics calculated from road traffic information obtained at different observation times or locations. The smaller the target function, the better it demonstrates the performance of the highway vehicle identity information representation model.

[0152]

[0153] In the formula, G yc and G yv Let represent the comprehensive evaluation index of vehicle driving characteristics calculated from the c-th and v-th records in the vehicle history dataset, respectively, where pi represents the pi-th record in the vehicle history dataset, and the same vehicle contains P records;

[0154] S35: Calculate the optimal weights of each indicator in the driving identity indicator system that satisfy the minimization of the objective function. Right now:

[0155]

[0156] In the formula, ω jLet be the weight of the j-th indicator in the driving identity indicator system.

[0157] S36: Calculate the object element r to be evaluated x Correlation coefficient between highway section domain and classical domain matter elements Right now:

[0158]

[0159] In the formula, R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the q-th classical domain matter element of the highway The correlation coefficient between them R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the highway section element The correlation coefficient between them is calculated as follows:

[0160]

[0161] S37: Calculate the object element r to be evaluated x Correlation coefficient between vehicle domain and classical domain matter elements Right now:

[0162]

[0163] In the formula, R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the q-th vehicle classical domain object element The correlation coefficient between them R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the highway section element The correlation coefficient between them is calculated as follows:

[0164]

[0165] S38: Based on the correlation between the object to be evaluated and the two-layer object, the optimal weights obtained from training on the vehicle historical dataset are used as the weights of the correlation coefficient to calculate the comprehensive evaluation index S of vehicle driving characteristics, so as to quantify the heterogeneity of vehicle driving characteristics and thus characterize the driving identity information of the target vehicle, that is:

[0166]

[0167] In the formula, This represents the optimal weight of the j-th indicator in the driving identity indicator system.

[0168] S4: Calculate the values ​​of each indicator in the driving identity indicator system using the target vehicle's driving data. Input this value into the highway vehicle identity information representation model and output the comprehensive evaluation index value of vehicle driving characteristics to represent the highway vehicle driving identity information. The probability of different vehicles belonging to the same vehicle is determined by the ratio between the absolute value of the difference between the comprehensive evaluation index values ​​of different vehicles and a preset threshold. The specific operation steps are as follows:

[0169] S41: Calculate the values ​​of each indicator in the driving identity indicator system based on the driving information of the target vehicle, input the indicator values ​​into the highway vehicle identity information representation model, and output the comprehensive evaluation index value of the vehicle driving characteristics of the target vehicle.

[0170] S42: Using the comprehensive evaluation index value of vehicle driving characteristics as a representation method for highway vehicle identity information, the probability P of different vehicles belonging to the same vehicle is determined by the ratio between the absolute value of the difference between the comprehensive evaluation index values ​​of different vehicles and a preset threshold, based on the driving data of different vehicles. id ,Right now:

[0171]

[0172] In the formula, S z1 and S z2 S represents the comprehensive evaluation index value of vehicle driving characteristics calculated using the driving information of target vehicle z1 and vehicle z2, respectively. pre This is a preset threshold.

[0173] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A method for representing the identity information of vehicles traveling on highways, characterized in that, Includes the following steps: S1: Considering vehicle performance, driver style, weather influence, surrounding traffic conditions and the spatiotemporal distribution of nearby vehicles, a driving identity index system that can describe vehicle driving identity information is constructed from the perspectives of relative characteristics and dynamic characteristics. S2: Collect highway vehicle driving data, calculate the values ​​of each indicator in the driving identity indicator system, construct a global historical dataset of highways based on the values ​​of each indicator calculated from all-weather data of different years, months and times, and construct a vehicle historical dataset by filtering the same vehicle data in the global historical dataset of highways. S3: Construct a highway vehicle identity information representation model. This model addresses the issue that traditional matter-element extension models do not consider differences in historical feature data when solving incompatible problems through quantitative evaluation. It improves the section-domain and classical-domain matter-element models in the matter-element extension model, forming a two-layer matter-element model. This includes highway section-domain matter-element and classical-domain matter-element models constructed using a global highway historical dataset, and vehicle section-domain matter-element and classical-domain matter-element models constructed using a vehicle historical dataset. This two-layer matter-element model enhances the model's ability to capture, identify, and represent vehicle identity information by considering both the global highway historical dataset and the vehicle historical dataset. A comprehensive evaluation index for vehicle driving characteristics is constructed using segment-domain objects and classical domain objects to characterize the driving identity information of the target vehicle. The comprehensive evaluation index for vehicle driving characteristics is a weighted value of the correlation function between the object to be evaluated and the bilayer segment-domain objects and classical domain objects. The object to be evaluated is constructed from the index values ​​of the driving identity index system calculated using the vehicle driving data of the target vehicle. The weighted value is the optimal weight of each index in the driving identity index system trained by the vehicle historical dataset. The comprehensive evaluation index for vehicle driving characteristics characterizes the driving identity information of the target vehicle by quantifying the correlation between the driving characteristic data of the target vehicle and the global historical dataset of the highway and the historical dataset of the vehicle. S4: Calculate the values ​​of each indicator in the driving identity indicator system using the target vehicle driving data, input the highway vehicle identity information representation model, and output the comprehensive evaluation index value of vehicle driving characteristics to represent the highway vehicle driving identity information. By comparing the absolute value of the difference between the comprehensive evaluation index values ​​of different vehicles with the preset threshold, the probability of different vehicles belonging to the same vehicle can be determined based on their driving data.

2. The method for representing the identity information of a vehicle traveling on a highway according to claim 1, characterized in that, The specific steps of step S1 are as follows: S11: From the perspective of vehicle driving trajectory, describe the heterogeneity of vehicle driving characteristics under different traffic environments, and construct vehicle relative characteristic indicators, including speed consistency indicators, acceleration consistency indicators, and gap distance consistency indicators. S12: Construct a highway vehicle distribution network to characterize vehicle clustering on the highway network. Use the degree centrality of vehicle nodes in the highway vehicle distribution network as the node weight to reflect the importance of vehicles in road segment traffic. Define the sensor detection range as the research interval, where the sensor detection range is defined as the range within which vehicle driving and appearance information can be obtained before and after the sensor. Using vehicles in the research interval as nodes, establish edges to connect vehicles traveling in the same direction, constructing an undirected network D. veh_network = {N, L}; Let the set of nodes be N = {n1, n2, ..., n}. i ,...,n M }, where node i in the set is denoted as n. i The number of nodes is M; L = {(n i ,n l The set of edges (n) is defined as follows: |i,l=1,2,...,M,i≠l}. i ,n l () represents an undirected edge connecting node i and node l, and is used to connect nodes in the highway vehicle distribution network whose absolute distance is less than a preset distance threshold d. pre Nodes: In the formula, x i x l Let y represent the longitudinal velocities of nodes i and l along the road travel direction, respectively. i y l These represent the lateral velocities of node i and node l perpendicular to the road travel direction, respectively. S13: Based on complex network theory, calculate the degree centrality of the highway vehicle distribution network as vehicle weights to describe the importance of clustered vehicle groups in traffic flow and the low influence of discrete vehicles in traffic flow. The degree centrality ω of node i is also considered. cen,i Defined as the ratio of node degree to the number of nodes, i.e.: In the formula, k i Let be the degree of node i, and let represent the number of edges in the network that connect to node i. S14: Building Speed ​​Consistency Metrics To characterize the degree of deviation of vehicle speed from the overall traffic flow speed, reflecting the vehicle's speed tendency in the traffic flow environment, that is: In the formula, v i (t) represents the velocity of node i at time t. The speed is the weighted average speed of the traffic flow, with the weight being the degree centrality ω of the node. cen,i With vehicle attribute ω type,i The weight of vehicle attributes depends on the vehicle type, describing the degree of influence of different types of vehicles on traffic flow speed; S15: Constructing Acceleration Consistency Metrics To quantify the difference in acceleration and deceleration frequencies between a vehicle and its surrounding traffic flow, the method characterizes the degree of deviation between the acceleration fluctuations of a vehicle during its driving process and the acceleration fluctuations of the overall traffic flow. In the formula, a i (t) represents the acceleration of node i at time t. Let t be the weighted average acceleration of traffic flow at node i at time t. s and t e These represent the times when the vehicle enters and leaves the sensor's observation and research range, respectively. S16: Construct a gap distance consistency index To characterize the difference between vehicle following distance and the overall spatial distribution pattern of traffic flow, reflecting the vehicle's preference for maintaining following distance in different traffic density environments, i.e.: In the formula, g i (t) represents the gap distance of node i at time t. This is the weighted average gap distance for traffic flow; S17: Construct vehicle dynamic characteristic indices to describe vehicle dynamic characteristics from the perspective of vehicle performance and operating characteristics. Incorporate the influence of weather and highway road environment to describe the heterogeneity of vehicle dynamic driving characteristics, including speed characteristic indices. Gap distance characteristic index Vehicle maximum acceleration index Lane change frequency index and conflict frequency indicators S18: Constructing speed characteristic indicators This characterizes the speed change of vehicles within the observation and study range, as well as the difference between the maximum and minimum speeds, taking into account the influence of weather conditions and road friction on the dynamic speed of vehicles. In the formula, θ represents the intensity of weather conditions, f(θ) is the influence function of weather conditions on driving behavior, σ is the road friction coefficient, and max(v) is the maximum friction coefficient. i (t)) and min(v i (t) represents the maximum and minimum velocities of node i within the observation range, respectively; S19: Constructing gap distance characteristic indicators This characterizes the changes in following distance and the difference between the maximum and minimum following distance as the vehicle enters the observation and research range, taking into account the influence of weather conditions and road friction on the following distance maintained by the vehicle during normal driving. In the formula, max(g) i (t)) and min(g) i (t) represents the maximum and minimum gap distances of node i within the observation range, respectively; S110: Constructing the maximum acceleration index for vehicles Describe the heterogeneity of vehicles in terms of performance and driving style, namely: In the formula, This represents the absolute value of the maximum acceleration of node i within the observation range; S111: Constructing a lane change frequency indicator Conflict frequency index This indicates the number of lane changes and traffic conflict incidents during vehicle operation, quantifying driving style and driver psychology. A cautious driving style tends to stay in the same lane and is less likely to cause traffic conflicts; an impulsive driving style maintains a small following distance and frequently changes lanes to increase speed. In the formula, Num(·) represents the indicator function. This represents the time of the j-th lane change at node i. The observation time [t] of node i was statistically analyzed. s ,t e The number of lane changes within ]; In the formula, TTC i Let Num(·) represent the collision time of node i, and let Num(·) represent the indicator function. This function counts the number of collisions of node i during the observation and research time [t]. s ,t e The number of events with an internal collision time of less than 2.6 seconds; The indicator function Num(·) is used to determine whether an internal condition is true, that is: Time-to-Collision (TTC) is a classic indicator in traffic safety research, characterizing the time between collisions when vehicles maintain their speeds, thus quantifying the risk of a rear-end collision. In the formula, x front and v front Let x represent the position and speed of the car ahead of node i, respectively. front -x i This represents the relative distance between node i and the vehicle in front of it; S112: Standardize the eight characteristic indicators, including the relative and dynamic characteristic indicators of the vehicle, to eliminate the influence of dimensional differences: In the formula, I jk and I' jk Let μ represent the j-th index value and the standardized value in sample k, respectively. j and σ j are the mean and standard deviation of the j-th feature index, respectively; S113: Based on the standardized indicators, construct a driving identity indicator system I to describe vehicle driving identity information, denoted as: In the formula, I j This represents the j-th indicator in the driving identity indicator system.

3. The method for representing the identity information of a vehicle traveling on a highway according to claim 1, characterized in that, The specific steps of step S2 are as follows: S21: Based on the highway network system and geographic information system, and sensors spaced at intervals along the basic highway sections, ramp entrances and exits, and merging zones, collect all-weather data for different years, months, and times, including vehicle trajectory information, weather conditions, vehicle snapshots, and highway vehicle driving data from gantry transaction logs within the sensor coverage area. Record this as U segments of highway vehicle driving data. Based on the collected data, calculate all indicators in the driving identity indicator system, denoted as the highway global historical dataset R. highway ={r1,r2,...,r a ,...,r U }, where r a =(r1,r2,...,r j (r, ..., r8) represents the calculated value of the driving identity index system corresponding to the vehicle driving data of segment a of the highway in the global historical dataset of highways, where r... j This represents the value of the j-th indicator; S22: Identify license plate numbers, vehicle models, body colors, and gantry transaction data in captured vehicle images; filter for data on the same vehicle within the overall historical highway dataset, denoted as vehicle history dataset R. vehicle ={r1,r2,...,r P },R vehicle ∈R highway P records containing the same vehicle, r P This represents the calculated value of the driving identity index system corresponding to the highway vehicle driving data in the Pth record of the vehicle history dataset.

4. The method for representing the identity information of a vehicle traveling on a highway according to claim 1, characterized in that, The specific steps of step S3 are as follows: S31: Taking the index value of the driving identity index system corresponding to any segment of highway driving data in the global historical dataset of highways as the target vehicle research object, denoted as the object element to be evaluated r in the highway vehicle identity information representation model. x The driving information of the target vehicle is represented by r. x =(r x1 ,r x2 ,...,r xj ,...,r x8 ), where r xj The value of the j-th index in the object to be evaluated; S32: Based on the global historical dataset of highways and the historical dataset of vehicles, construct highway section-domain objects and vehicle section-domain objects respectively. Define the value range of the j-th index in the highway section-domain object as follows: Recorded as and These are the corresponding lower and upper limits of the value range; The range of values ​​for the j-th index in the vehicle segment object element is defined as follows: Recorded as and These are the corresponding lower and upper limits of the value range; S33: Divide the highway section-domain material element and the vehicle section-domain material element into s classical domain material elements at equal intervals. Define the value range of the j-th index in the q-th highway classical domain material element as follows: Recorded as and These are the corresponding lower and upper limits of the value range; Define the range of values ​​for the j-th index in the q-th vehicle classical domain matter element as follows: Recorded as and These are the corresponding lower and upper limits of values: in, and These represent the distances between classical domain objects of each highway and between classical domain objects of each vehicle, respectively. In the formula, and Let be the minimum and maximum values ​​of the j-th index in the classical domain matter element of the highway, respectively. and These are the minimum and maximum values ​​of the j-th index in the classic domain object element of the vehicle, respectively; S34: Using the vehicle history dataset to train the model, the optimal weights of each indicator in the driving identity indicator system are obtained. First, a target function loss is defined to search for the optimal weight allocation. Based on the comprehensive evaluation index of vehicle driving characteristics, a target function is constructed to reflect the differences between the same vehicle driving characteristics calculated from road traffic information obtained at different observation times or locations. The smaller the target function, the better it demonstrates the performance of the highway vehicle identity information representation model. In the formula, G yc and G yv Let represent the comprehensive evaluation index of vehicle driving characteristics calculated from the c-th and v-th records in the vehicle history dataset, respectively, where pi represents the pi-th record in the vehicle history dataset, and the same vehicle contains P records; S35: Calculate the optimal weights of each indicator in the driving identity indicator system that satisfy the minimization of the objective function. Right now: In the formula, ω j Let be the weight of the j-th indicator in the driving identity indicator system; S36: Calculate the object element r to be evaluated x Correlation coefficient between highway section domain and classical domain matter elements Right now: In the formula, R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the q-th classical domain matter element of the highway The correlation coefficient between them R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the highway section element The correlation coefficient between them is calculated as follows: S37: Calculate the object element r to be evaluated x Correlation coefficient between vehicle domain and classical domain matter elements Right now: In the formula, R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the q-th vehicle classical domain matter element The correlation coefficient between them R represents the value of the j-th index of the object to be evaluated. xj The range of values ​​for the j-th index in the highway section element The correlation coefficient between them is calculated as follows: S38: Based on the correlation between the object to be evaluated and the two-layer object, the optimal weights obtained from training on the vehicle historical dataset are used as the weights of the correlation coefficient to calculate the comprehensive evaluation index S of vehicle driving characteristics, so as to quantify the heterogeneity of vehicle driving characteristics and thus characterize the driving identity information of the target vehicle, that is: In the formula, This represents the optimal weight of the j-th indicator in the driving identity indicator system.

5. The method for representing the identity information of a vehicle traveling on a highway according to claim 1, characterized in that, The specific steps for step S4 are as follows: S41: Calculate the values ​​of each indicator in the driving identity indicator system based on the driving information of the target vehicle, input the indicator values ​​into the highway vehicle identity information representation model, and output the comprehensive evaluation index value of the vehicle driving characteristics of the target vehicle. S42: Using the comprehensive evaluation index value of vehicle driving characteristics as a representation method for highway vehicle identity information, the probability P of different vehicles belonging to the same vehicle is determined by the ratio between the absolute value of the difference between the comprehensive evaluation index values ​​of different vehicles and a preset threshold, based on the driving data of different vehicles. id ,Right now: In the formula, S z1 and S z2 S represents the comprehensive evaluation index value of vehicle driving characteristics calculated using the driving information of target vehicle z1 and vehicle z2, respectively. pre This is a preset threshold.

Images