A traffic monitoring method and system for vehicle tracking early warning

Abstract

The present application relates to the technical field of traffic monitoring, in particular to a traffic monitoring method and system for vehicle tracking and early warning, comprising the following steps: acquiring real-time position, speed and acceleration data of all vehicles in a monitored road area, aggregating adjacent vehicles into dynamic traffic micro-clusters according to the spatial distance and time window between vehicles, and establishing a regional motion state benchmark. The present application sets the spatial distance and time window between vehicles, divides the dynamic traffic micro-clusters, realizes the refinement and dynamic of vehicle clustering, improves the sensitivity and response speed to the changes of the road micro-traffic environment; uses the regional motion state benchmark to calculate the deviation degree of the motion vector of a single vehicle and performs weighted summation operation, realizes accurate quantification of individual driving behavior instability, captures abnormal driving state of the vehicle, and improves the accuracy of abnormal behavior identification.

Description

Technical Field

[0001] This invention relates to the field of traffic monitoring technology, and in particular to a traffic monitoring method and system for vehicle tracking and early warning. Background Technology

[0002] The field of traffic monitoring technology involves activities such as real-time or non-real-time monitoring, data collection, status analysis, and event identification of vehicles, pedestrians, and road infrastructure in the road traffic environment. It aims to improve road traffic safety, optimize traffic flow efficiency, and reduce the risk of traffic accidents, and is an important component of intelligent transportation.

[0003] Current technologies only collect and analyze vehicle location, traffic flow, and speed in a simple and general manner. This results in a limited dimension of traffic condition analysis, leading to delayed or inaccurate warnings of abnormal driving behaviors. For example, it is difficult to quickly identify and locate abnormal driving states such as sudden deceleration or abrupt lane changes, often resulting in missed or false alarms, increasing the likelihood of traffic accidents. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology and to propose a traffic monitoring method and system for vehicle tracking and early warning.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a traffic monitoring method for vehicle tracking and early warning, comprising the following steps:

[0006] The system acquires real-time location, speed, and acceleration data of all vehicles within the monitored road area. Based on the spatial distance and time window settings between vehicles, adjacent vehicles are aggregated into dynamic traffic micro-clusters to establish a regional motion state benchmark.

[0007] Based on the regional motion state benchmark, the motion vector of a single vehicle is extracted, the individual driving deviation value is calculated, and then the continuous individual driving deviation values ​​of a single vehicle within a set time period are summed to obtain a vehicle behavior instability measure.

[0008] Based on the vehicle behavior instability measure of each vehicle, the vehicle behavior instability measures of all vehicles on the target road section are summed to obtain the cumulative driving risk of the road section. Then, the cumulative driving risk of the road section is calculated with the real-time traffic flow, average speed and road capacity of the road section to generate the traffic flow conflict risk index of the road section.

[0009] Based on the traffic flow conflict risk index of the road cross section, the traffic flow conflict risk index is compared with the preset multi-level risk thresholds one by one. According to the comparison results, a preset control strategy set is matched and selected from the associated control strategy library. Based on the preset control strategy set, an adaptive traffic control instruction for the traffic lights or variable information signs at the downstream intersection is generated.

[0010] Preferably, the steps for obtaining the regional motion state reference are as follows:

[0011] By deploying video sensors and millimeter-wave radar within the road area, all vehicles currently traveling within the road area are simultaneously detected in real time, acquiring real-time position data, speed data, and acceleration data for each vehicle, and using timestamps to uniformly mark the data of each vehicle, generating a set of real-time motion parameters for the vehicles.

[0012] Based on the set of real-time vehicle motion parameters, according to the preset spatial distance threshold between vehicles and the time window threshold for collecting vehicle motion status, vehicles whose spatial distance is lower than the spatial distance threshold and whose data collection time difference is within the time window threshold range are identified as adjacent vehicles, and the adjacent vehicles are aggregated to form multiple vehicle groups to generate dynamic traffic micro-clusters.

[0013] Based on the dynamic traffic micro-clusters, the speed and acceleration data of all vehicles in each dynamic traffic micro-cluster are extracted to obtain the regional motion state benchmark.

[0014] Preferably, the step of obtaining the individual driving deviation value is as follows:

[0015] Based on the aforementioned regional motion state benchmark, the velocity and acceleration components of a single vehicle at each time step are extracted, and a four-dimensional motion vector of the vehicle at each time step is constructed. The four-dimensional motion vectors corresponding to all time steps are integrated to form a continuous sequence of vehicle motion vectors, thus obtaining a single vehicle motion vector sequence.

[0016] Based on the single vehicle motion vector sequence, the Mahalanobis distance between the vehicle motion vector at each time step and the average motion vector of the dynamic traffic micro-cluster to which it belongs is calculated to form the individual driving deviation value.

[0017] Preferably, the step of obtaining the vehicle behavior instability measure is as follows:

[0018] Based on the individual driving deviation value, a vehicle behavior instability metric is calculated.

[0019] Preferably, the steps for obtaining the cumulative driving risk of the aforementioned road segment are as follows:

[0020] Based on the vehicle behavior instability measure of all vehicles within the target road section, extract the vehicle behavior instability measure values ​​of all vehicles at the current time step, sum the extracted values, and obtain the cumulative amount of road segment driving risk of the current road section at the current time step.

[0021] Preferably, the steps for obtaining the traffic flow conflict risk index are as follows:

[0022] Based on the cumulative driving risk of the road segment, the real-time traffic flow, road capacity and average vehicle speed of the corresponding time step of the road section are retrieved and included in the calculation input set together with the cumulative driving risk of the road segment to obtain the joint variable set constituting the risk index calculation.

[0023] Based on the aforementioned set of joint variables, a traffic flow conflict risk index is calculated.

[0024] Preferably, the step of obtaining the preset control strategy set is as follows:

[0025] Based on the traffic flow conflict risk index of the road section, a predefined multi-level risk threshold set is called, and each risk threshold in the risk threshold set is extracted one by one. The traffic flow conflict risk index is compared with the value of each risk threshold to determine the risk level that matches the current traffic flow conflict risk index and generate the road section risk level judgment result.

[0026] Based on the risk level determination result of the road section, the risk level determination result is used as the query condition to retrieve the pre-defined control strategy set corresponding to the risk level from the associated control strategy library, forming a preset control strategy set applicable to the current road section risk level.

[0027] Preferably, the step of obtaining the adaptive traffic control command is as follows:

[0028] Based on the preset control strategy set, strategy instructions are extracted one by one from the control strategy set and converted into instruction data streams that can be executed by traffic control equipment. Control actions are generated one by one for traffic lights or variable information signs at downstream intersections according to the instruction data streams, forming adaptive traffic control instructions.

[0029] This invention also provides a traffic monitoring system, comprising:

[0030] The data acquisition module acquires real-time location, speed, and acceleration data of all vehicles within the monitored road area. Based on the spatial distance and time window settings between vehicles, it aggregates adjacent vehicles into dynamic traffic micro-clusters and establishes a regional motion state benchmark.

[0031] The behavior assessment module extracts the motion vector of a single vehicle based on the regional motion state benchmark, calculates the individual driving deviation value, and then sums the continuous individual driving deviation values ​​of a single vehicle within a set time period to obtain a vehicle behavior instability measure.

[0032] The risk calculation module sums the vehicle behavior instability measures of all vehicles on the target road section based on the vehicle behavior instability measures of each vehicle to obtain the cumulative driving risk of the road section. Then, the cumulative driving risk of the road section is calculated with the real-time traffic flow, average speed and road capacity of the road section to generate the traffic flow conflict risk index of the road section.

[0033] The control instruction module compares the traffic flow conflict risk index with preset multi-level risk thresholds one by one based on the traffic flow conflict risk index of the road cross section. According to the comparison results, it matches and selects a preset control strategy set from the associated control strategy library. Based on the preset control strategy set, it generates adaptive traffic control instructions for traffic lights or variable information signs at downstream intersections.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0035] This invention achieves refined and dynamic vehicle clustering by setting spatial distances and time windows between vehicles, thereby improving sensitivity and response speed to changes in the micro-traffic environment. It utilizes regional motion state benchmarks to calculate the deviation of individual vehicle motion vectors and perform weighted summation operations, accurately quantifying the instability of individual driving behavior, capturing abnormal driving states, and improving the accuracy of abnormal behavior identification. At the road cross-section level, it integrates vehicle behavior instability with multi-dimensional traffic parameters such as real-time traffic flow, road capacity, and average vehicle speed, generating a traffic flow conflict risk index in real time through collaborative computation. This makes risk assessment more consistent with actual traffic conditions, improving the real-time performance and reliability of traffic risk warnings. Furthermore, it compares the traffic flow conflict risk index level by level using multi-level risk thresholds, matching control strategies, and dynamically generating control commands for downstream traffic lights or forward variable message signs, achieving proactive and adaptive traffic intervention, reducing the probability of traffic accidents, and improving overall road traffic safety and efficiency. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the steps of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0038] Please see Figure 1 This invention provides a technical solution: a traffic monitoring method for vehicle tracking and early warning, comprising the following steps:

[0039] The system acquires real-time location, speed, and acceleration data of all vehicles within the monitored road area. Based on the spatial distance and time window settings between vehicles, adjacent vehicles are aggregated into dynamic traffic micro-clusters to establish a regional motion state benchmark.

[0040] Based on the regional motion state benchmark, the motion vector of a single vehicle is extracted, the individual driving deviation value is calculated, and then the continuous individual driving deviation values ​​of a single vehicle within a set time period are summed to obtain a vehicle behavior instability measure.

[0041] Based on the vehicle behavior instability measurement of each vehicle, the vehicle behavior instability measurement of all vehicles on the target road section is summed to obtain the cumulative amount of road segment driving risk. Then, the cumulative amount of road segment driving risk is calculated with the real-time traffic flow, average speed and road capacity of the road section to generate the traffic flow conflict risk index of the road section.

[0042] Based on the traffic flow conflict risk index of the road cross section, the traffic flow conflict risk index is compared with the preset multi-level risk thresholds one by one. According to the comparison results, the preset control strategy set is matched and selected from the associated control strategy library. Based on the preset control strategy set, adaptive traffic control instructions are generated for the traffic lights or variable information signs at the downstream intersection.

[0043] The steps for obtaining the regional motion state benchmark are as follows:

[0044] By deploying video sensors and millimeter-wave radar within the road area, all vehicles currently traveling within the road area are simultaneously detected in real time, acquiring real-time position data, speed data, and acceleration data for each vehicle, and using timestamps to uniformly mark the data of each vehicle, generating a set of real-time motion parameters for the vehicles.

[0045] Based on the real-time vehicle motion parameter set, according to the preset inter-vehicle spatial distance threshold and the time window threshold for vehicle motion state collection, vehicles with a spatial distance lower than the spatial distance threshold and a data collection time difference within the time window threshold range are identified as adjacent vehicles, and adjacent vehicles are aggregated to form multiple vehicle groups, generating dynamic traffic micro-clusters.

[0046] Based on dynamic traffic micro-clusters, the speed and acceleration data of all vehicles in each dynamic traffic micro-cluster are extracted to obtain the regional motion state benchmark.

[0047] Specifically, at key locations in the monitored road area, such as 100 to 300 meters upstream of major intersections, accident-prone sections, or highway ramp merging areas, multiple sets of video sensors and millimeter-wave radars are deployed along the driving direction. This ensures that the sensor field of view completely covers all lanes and that there is at least 20% overlap between adjacent sensors. All sensors are connected to a unified Network Time Protocol (NTP) server to achieve timestamp synchronization of the collected data, with synchronization accuracy controlled within 10 milliseconds. After the system starts, the video sensors continuously capture high-definition video streams of the road at a frequency of 25 frames per second, while the millimeter-wave radar emits electromagnetic waves at a frequency of 20 Hz and receives target reflection signals. The system processes the two data sources in parallel. For the video stream data, a YOLOv5 object detection model pre-trained on a large-scale traffic image dataset (such as the BDD100K or KITTI dataset) is invoked to identify all vehicles in each frame of the image and generate a target for each vehicle. A bounding box is formed. Then, the DeepSORT tracking algorithm is used to predict the vehicle's position in the next frame using Kalman filtering and combining this with vehicle appearance features to assign each vehicle a unique tracking ID within a short time, achieving stable tracking across frames. Subsequently, using the homography transformation matrix pre-calculated using the checkerboard calibration method, the center pixel coordinates of each vehicle's bounding box are converted to world coordinates (x, y) on the actual road plane. By calculating the change in world coordinates between consecutive frames and dividing by the frame interval (40 milliseconds), the vehicle's velocity components on the x and y axes are obtained. The acceleration components are then differentiated. For millimeter-wave radar data, the system directly parses its output target list, obtaining information such as distance, azimuth, and radial velocity for each target. This information is used to calculate the Cartesian coordinate position and velocity vector of the target vehicle. Then, an extended Kalman filter (EKF) is initiated for data fusion. The state vector of this filter is defined as... The position measurements provided by video data and the position and velocity measurements provided by radar data are used as the update inputs for EKF. Through fusion processing, a more accurate and robust set of motion state estimates is generated for each vehicle that is jointly detected. Finally, the unique tracking ID, high-precision synchronization timestamp, fused position data, velocity data and acceleration data of all vehicles at each time step (e.g. every 50 milliseconds) are integrated to form a set of real-time motion parameters of the vehicles.

[0048] Based on the real-time vehicle motion parameter set generated in the previous process, the system initiates the aggregation process of dynamic traffic micro-clusters. The core of this process is to determine the correlation of all vehicles based on two dynamically adjusted thresholds: a spatial distance threshold between vehicles and a time window threshold for vehicle motion state acquisition. First, the time window threshold for vehicle motion state acquisition is set, taking into account the frequency of data acquisition and processing. For example, if the data update frequency of the real-time vehicle motion parameter set is 20 Hz (i.e., one data point every 50 milliseconds), then the time window threshold is set to 150 milliseconds, or 3 data update cycles, to tolerate potential network transmission latency and computation time, ensuring that the vehicle data used for comparison reflects traffic conditions at similar times. Next, the system sets the spatial distance threshold between vehicles. This threshold is not a fixed value but is related to the instantaneous speed of the vehicles. Specifically, it is calculated by adding the reaction distance to the basic safety distance, i.e., spatial distance threshold = 5 meters + vehicle instantaneous speed × The 1.5-second distance threshold consists of 5 meters (the minimum safe following distance when the vehicle is stationary or moving at low speed) and 1.5 seconds (the average reaction time of the driver during the following process). For example, if a vehicle is traveling at 72 km / h (20 m / s), the corresponding spatial distance threshold is 5 + 20 × 1.5 = At 35 meters, when the system is aggregating, it takes any un-clustered vehicle in the set as the core point, traverses all other vehicles in the set of real-time vehicle motion parameters, and checks whether the absolute value of the difference between the data timestamps of each pair of vehicles is less than 150 milliseconds. If the time window requirement is met, it further calculates the Euclidean distance between the two vehicles in the world coordinate system and compares this distance with the dynamic spatial distance threshold calculated based on the speed of the core point vehicle. All vehicles that meet the threshold conditions in both time and space are identified as adjacent vehicles. All vehicles that are reachable through adjacency (i.e., vehicles A and B are adjacent, and B and C are adjacent, so A, B, and C belong to the same group) are grouped into the same vehicle group. This process is repeated until all vehicles have been visited, and finally a dynamic traffic micro-cluster composed of multiple independent vehicle groups is generated.

[0049] Based on the generated dynamic traffic micro-cluster, the system calculates the overall motion characteristics of each independent vehicle group to construct a regional motion state benchmark. Specifically, during execution, the system iterates through each vehicle group within the dynamic traffic micro-cluster. For a selected group, it first extracts the speed data (in two-dimensional vector form) of these vehicles at the current time step from the real-time vehicle motion parameter set, based on the list of vehicle IDs contained within the group. Formal representation) and acceleration data (in two-dimensional vectors) (Formal representation) Before performing statistical calculations, an outlier removal step is first performed to enhance the robustness of the baseline. For speed data, the system calculates the median and interquartile range (IQR) of the speed modulus (i.e., rate) of all vehicles in the group, removing outliers that are outside the range (first quartile - 1.5 × IQR, third quartile + 1.5 × IQR). Vehicles identified by IQR (Independent Quality Rating) are marked as speed outliers. Similarly, outlier detection is performed on acceleration modulus. These marked outliers are temporarily excluded in the subsequent average calculation. Then, the system calculates the average velocity vector and average acceleration vector of the group. The average velocity vector is calculated by adding the x-components of the velocity vectors of all non-outliers and dividing by the total number of non-outliers. Similarly, the y-components are processed in the same way to obtain the y-component of the average velocity. These two components together constitute the average velocity vector of the group. The calculation process for the average acceleration vector is exactly the same as that for the average velocity vector, except that the data object is replaced by the acceleration vector component. The calculated average velocity vector and average acceleration vector of each group are associated and stored together with their corresponding group ID. This set, which contains the collective motion state descriptors of all vehicle groups, ultimately constitutes the regional motion state benchmark.

[0050] The steps for obtaining individual driving deviation values ​​are as follows:

[0051] Based on the regional motion state benchmark, the velocity and acceleration components of a single vehicle at each time step are extracted, and a four-dimensional motion vector of the vehicle at each time step is constructed. The four-dimensional motion vectors corresponding to all time steps are integrated to form a continuous sequence of vehicle motion vectors, thus obtaining a single vehicle motion vector sequence.

[0052] Based on the sequence of individual vehicle motion vectors, the Mahalanobis distance between the vehicle motion vector at each time step and the average motion vector of the dynamic traffic micro-cluster to which it belongs is calculated to form the individual driving deviation value.

[0053] Specifically, based on the regional motion state benchmark, the system, for each independently tracked vehicle in the real-time motion parameter set, calls its unique tracking ID and continuously extracts its motion data at each time step from the set. The length of a time step is set to 50 milliseconds, consistent with the 20 Hz data update frequency of the front-end sensors. At each time step, the system accurately extracts the longitudinal and lateral components of the vehicle's velocity vector in the road coordinate system, as well as the longitudinal and lateral components of the acceleration vector in the same coordinate system. For example, at time step t, the system obtains that the longitudinal velocity of vehicle ID "A001" is 15.8 m / s, the lateral velocity is -0.3 m / s, the longitudinal acceleration is 1.1 m / s², and the lateral acceleration is -0.2 m / s². Subsequently, the system combines these four values ​​in a predefined order, i.e., [longitudinal velocity, lateral velocity, longitudinal acceleration, lateral acceleration], into a four-dimensional array. This array is the four-dimensional motion vector of the vehicle at time step t, i.e., [15.8, -0.3, 1.1, ...]. [-0.2] The system maintains a first-in-first-out data queue for each vehicle. The length of the queue is determined according to a preset time period, which is set to 5 seconds. This time period is determined based on research on driving behavior analysis. It can capture meaningful driving patterns and react promptly to emergencies. Therefore, the data queue stores the four-dimensional motion vectors of the past 5 seconds, i.e., 100 consecutive time steps. Whenever a new four-dimensional motion vector of a new time step is calculated and added to the front of the queue, the oldest vector at the end of the queue is removed. In this way, the system dynamically maintains an ordered set containing all the motion states of each vehicle in the last 5 seconds, resulting in a single vehicle motion vector sequence.

[0054] Based on the single vehicle motion vector sequence generated in the previous step, the system processes each time step in the sequence, calculating the degree of deviation of each four-dimensional motion vector from the behavioral benchmark of its respective cluster. Specifically, for a specific vehicle's four-dimensional motion vector at time step t, the system first determines the ID of the dynamic traffic micro-cluster to which the vehicle belongs at that moment by querying the dynamic affiliation record between the vehicle and the cluster. Then, using this cluster ID, the system retrieves the overall motion statistics of the cluster from the regional motion state benchmark, including the four-dimensional average motion vector and the covariance matrix of the four-dimensional motion vector. The four-dimensional average motion vector is obtained by averaging the components of the four-dimensional motion vectors of all vehicles in the cluster, while the 4x4 covariance matrix is ​​obtained by calculating the variance of the four-dimensional motion vectors of all vehicles in the cluster in each dimension and the covariance between different dimensions. This covariance matrix describes the overall distribution of vehicle motion states within the cluster and the correlation between variables. For example, it reflects not only the dispersion of speed but also the correlation pattern between speed changes and acceleration changes. Then, the system uses Mahalanobis distance to quantify the difference between the individual vehicle motion vector and the cluster average motion vector. The calculation formula is as follows: ,in, It is the four-dimensional motion vector of a single vehicle at time step t. It is the four-dimensional average motion vector of its cluster. It is the inverse of the cluster covariance matrix. This calculation process first obtains the difference vector between the vehicle vector and the cluster average vector, and then transforms the distance difference in the multidimensional space into a dimensionless scalar value that takes into account the data distribution pattern by multiplying it on the left by its transpose and on the right by the inverse of the covariance matrix. This scalar value is the individual driving deviation value of the vehicle at the current time step.

[0055] The steps for obtaining vehicle behavior instability metrics are as follows:

[0056] Based on individual driving deviation values, a measure of vehicle behavior instability is calculated using the following formula:

[0057] ;

[0058] in, A measure of vehicle behavior instability. This represents the total number of time steps within the time period. For time step The Mahalanobis distance represents the degree of deviation between the vehicle motion vector and the average motion vector of the dynamic traffic micro-cluster under the covariance structure. For time step The velocity vector magnitude represents the instantaneous speed of the vehicle. Let be the average velocity vector magnitude of the dynamic traffic micro-cluster, and represent the average speed of the cluster. It is a constant positive number.

[0059] Specifically, the formula: The advantage of the formula is that it does not use the deviation of multidimensional motion states (Mahaviron distance) in isolation. Instead, it introduces a nonlinear weighting mechanism for velocity differences through a hyperbolic tangent function (tanh). This weighting term This causes the driving deviation value to be significantly higher when the instantaneous speed of a vehicle is significantly higher than the average speed of its cluster. The speed is amplified at high speeds and suppressed at low speeds, capturing the higher risk of anomalous behaviors (such as rapid acceleration and lane changes) compared to similar behaviors at low speeds. The formula design uses the relative magnitude of speed as a regulating factor, making the final metric more sensitive to potentially dangerous high-speed instability. Furthermore, the use of a root mean square structure to aggregate weighted deviations over a time period T smooths out instantaneous and accidental disturbances, resulting in a more time-representative and robust instability score, rather than an overreaction to single-point events. The final vehicle behavior instability metric is thus obtained. It can more accurately reflect the driver's consistent driving style and potential risk level.

[0060] The total number of time steps within a time period defines the size of the time window for calculating vehicle behavior instability metrics. Its value requires a trade-off between real-time assessment and result stability. A shorter time period responds more quickly to changes in driving behavior but may introduce noise due to accidental driving maneuvers; a longer time period provides a more stable assessment but may lag in response to sudden risks. In this method, analysis of a large amount of real traffic data streams containing normal driving and pre-accident events reveals that a 5-second time window achieves the best balance in most scenarios. Therefore, under the condition that the system's data acquisition frequency is 20 Hz (i.e., each time step is 0.05 seconds), The value is set to 100 (i.e., 5 seconds / 0.05 seconds / step), which means that the vehicle behavior instability metric is based on a comprehensive evaluation of the vehicle’s past 100 consecutive motion states.

[0061] For time step The Mahalanobis distance, calculated using the preceding steps, quantifies the distance of a single vehicle at a given time step. The Mahalanobis distance is a measure of the difference between a vehicle's four-dimensional motion vector (containing velocity and acceleration components) and the average motion state of its dynamic traffic micro-cluster. It is a dimensionless numerical value representing the statistical distance of a vehicle's motion state from its cluster center, taking into account the variance and correlation of each motion variable within the cluster. For example, in the calculation of the previous step, at time step t=1, the difference between a vehicle's four-dimensional motion vector and its cluster center, after considering the covariance structure of the overall motion distribution of the cluster, yields a Mahalanobis distance of 2.1.

[0062] For time step The velocity vector magnitude represents the vehicle's instantaneous velocity at that moment. This value is directly obtained from the vehicle's real-time motion parameter set, obtained by analyzing the vehicle's velocity at time steps. The two components of the velocity vector (longitudinal velocity) and lateral speed The calculation is performed using the following formula: For example, at time step t=1, the sensor fusion system measures the longitudinal velocity of a vehicle as 19.9 m / s and the lateral velocity as 2.0 m / s. Then its instantaneous speed... The calculation result is meters per second.

[0063] The average velocity vector magnitude of the dynamic traffic micro-cluster represents the overall average speed of the cluster to which the vehicle belongs. This value is derived from the regional motion state benchmark. First, the average velocity vector of the cluster needs to be obtained from the benchmark. This vector is obtained by averaging the x and y components of the velocity vectors of all vehicles in the cluster. Then, the magnitude of this average velocity vector is calculated using the following formula: For example, a cluster contains three vehicles with speed vectors of (15.8 m / s, 0.1 m / s), (16.0 m / s, -0.1 m / s), and (16.2 m / s, 0.0 m / s), respectively. The x-component of the average speed vector is (15.8 + 16.0 + 16.2) / 3 = 16.0 m / s, and the y-component is (0.1 - 0.1 + 0.0) / 3 = 0.0 m / s. Therefore, the average speed vector is (16.0, 0.0), and its magnitude is... for meters per second.

[0064] It is a constant positive number, its function is to prevent the denominator from being zero in the calculation. Under specific traffic conditions, such as when severe traffic congestion causes the vehicle group to be completely stationary, it represents the average velocity vector magnitude of the dynamic traffic micro-cluster. It could be zero.

[0065] Calculation process:

[0066] To calculate the time of a single vehicle in a time interval containing 3 time steps (i.e. Vehicle behavior instability measurement over a period of time For example, the detailed calculation process is shown below:

[0067] Based on the aforementioned steps, the average rate of the cluster to which the vehicle belongs has been obtained. meters per second, and a constant .

[0068] The instantaneous speed of the vehicle is monitored and calculated over three consecutive time steps. and Mahal distance as follows:

[0069] Time step t=1: The vehicle accelerates rapidly to overtake. meters per second .

[0070] Time step t=2: Vehicles return to the main road. meters per second .

[0071] Time step t=3: Vehicle resumes cruise control. meters per second .

[0072] The calculation process is as follows:

[0073] Calculate the weighted deviation term for each time step .

[0074] For t=1:

[0075] Speed ​​difference term: ;

[0076] Weighting factor: ;

[0077] Weighted deviation term: ;

[0078] For t=2:

[0079] Speed ​​difference term: ;

[0080] Weighting factor: ;

[0081] Weighted deviation term: ;

[0082] For t=3:

[0083] Speed ​​difference term: ;

[0084] Weighting factor: ;

[0085] Weighted deviation term: ;

[0086] Calculate the root mean square of the weighted deviation term.

[0087] Sum of Mean Squares: ;

[0088] ;

[0089] Prescription: ;

[0090] The results indicate that the vehicle's behavioral instability measures over these three time steps... The calculated value is 2.05. This value comprehensively reflects the degree to which a vehicle's driving behavior deviates from its traffic flow baseline during the observation period, with a focus on the additional risks posed by speeding. The system compares this value with a preset risk level threshold, which is obtained through statistical analysis of massive amounts of historical traffic data. For example, the analysis results show that for normally driving vehicles... Values ​​below 1.5 account for 95% of the time; therefore, 1.5 is set as the dividing line between stable and moderately unstable driving, including driving behaviors that indicate an impending accident or dangerous event. Since the values ​​are generally higher than 3.0, 3.0 is set as the dividing line between moderately unstable and high-risk driving. The current calculated value of 2.05 is between 1.5 and 3.0, indicating that the vehicle exhibits a moderate degree of behavioral instability and requires continuous monitoring by the system.

[0091] The steps to obtain the cumulative driving risk of a road segment are as follows:

[0092] Based on the vehicle behavior instability measurement of all vehicles within the target road section, the vehicle behavior instability measurement values ​​of all vehicles at the current time step are extracted, and the extracted values ​​are summed to obtain the cumulative amount of road segment driving risk at the current time step.

[0093] Specifically, based on the vehicle behavior instability measurement of all vehicles within the target road section, the system first defines the spatial range of the target road section. For example, for an important intersection, the section is defined as a 500-meter area extending along the road centerline at its upstream entrance. At each time step, i.e., every 50 milliseconds, the system initiates a risk accumulation calculation. At this time step, the system first iterates through the set of real-time vehicle motion parameters maintained within the section. By comparing the real-time location data of each vehicle with the preset geofence coordinates of the section, it filters out all vehicles currently traveling within the 500-meter section and obtains a unique identification ID list for these vehicles. Subsequently, based on this ID list, the system queries and extracts the latest calculated vehicle behavior instability measurement value for each vehicle from a dynamically updated data table. This data table uses the vehicle ID and timestamp as a combined primary key and stores the historical and current measurement values ​​for all vehicles. The extraction operation returns a list of values. For example, at the current time step, there are 35 vehicles within the section, and the system extracts these 35 measurement values, such as [2.05, 0.8, 1.7, ..., [1.1] Next, the system performs a summation on this list of 35 values, that is, adds up all the values. For example, if the sum of these 35 vehicle behavior instability measurement values ​​is 48.6, then this value is confirmed as the cumulative amount of road segment driving risk at the current time step.

[0094] The steps to obtain the traffic flow conflict risk index are as follows:

[0095] Based on the cumulative amount of driving risk in a road segment, the real-time traffic flow, road capacity, and average vehicle speed of the corresponding time step of the road section are retrieved and included in the calculation input set along with the cumulative amount of driving risk in the road segment, thus obtaining the joint variable set that constitutes the risk index calculation.

[0096] Based on the joint variable set, the traffic flow conflict risk index is calculated using the following formula:

[0097] ;

[0098] in, Traffic flow conflict risk index This represents the cumulative driving risk on the road segment. To provide the real-time traffic flow at a specified road section at the current moment, To specify the traffic capacity of a road section, To specify the average vehicle speed across a road section, The optimal vehicle speed set for the cross-section. As a congestion-sensitive factor, This is a velocity offset sensitivity factor.

[0099] Specifically, based on the accumulated driving risk of the current road section at the current time step obtained in the previous step, the system simultaneously retrieves three macroscopic traffic flow parameters corresponding to that section and time step: real-time traffic flow, road capacity, and average vehicle speed. Real-time traffic flow is counted by setting a virtual detection line at the end of the section. The system counts the total number of vehicles that passed this detection line in the past 5 minutes and then extrapolates this count to the hourly traffic flow. For example, if 150 vehicles passed through in the past 5 minutes, the real-time traffic flow is calculated as 150 vehicles / 5 minutes × 150%. 60 minutes / hour = 1800 vehicles / hour. Road capacity is a static parameter pre-set in the system's road network database. It is calculated based on the physical properties of the cross section, such as the number of lanes, lane width, lateral clearance, and road gradient, according to the standard method in the Highway Capacity Manual (HCM). For example, the theoretical capacity of a two-lane urban arterial road with a design speed of 80 km / h is set to 3600 vehicles / hour. Finally, the average vehicle speed is obtained by calculating the average of the instantaneous speeds of all vehicles within the cross section at the current time step. For example, if the average instantaneous speed of 35 vehicles within the cross section is 62 km / h, then the average vehicle speed is 62 km / h. The system packages these three macro parameters together with the previously obtained cumulative road segment driving risk (e.g., 48.6) to form a data structure containing four key variables, resulting in a set of joint variables constituting the risk index calculation.

[0100] formula: The advantage of the formula lies in the exponential amplification term. This introduces two key macro-level traffic flow regulation factors: traffic congestion level (based on real-time vehicle flow). With road capacity The ratio of the two values ​​(represented by the average vehicle speed) and traffic flow speed stability (represented by the average vehicle speed) With optimal vehicle speed The degree of deviation is represented by the exponential function, which makes the risk index more... The increase in congestion and speed deviation exhibits a non-linear, exponential amplification effect, which coincides with the phenomenon that the risk of accidents rises sharply when traffic flow approaches its break point in reality. This can be verified through sensitive factors. and Through adjustments, the model can be calibrated for the characteristics of different road types, ultimately outputting a traffic flow conflict risk index. It not only quantifies the current risk level, but also determines whether the risk is driven by individual driving behavior or by the deterioration of the overall traffic flow.

[0101] The cumulative driving risk for a road segment is directly derived from the calculation results of the preceding steps. It is obtained by summing the vehicle behavior instability measures of all vehicles within the target road section at a specific time step. It quantifies the basic risk level of the road segment at that moment, which is composed of the unstable behaviors of individual drivers. The higher the value, the more or more severe unstable driving behaviors exist within the road segment. For example, in the preceding steps, the system identified 35 vehicles within the target section and extracted their respective vehicle behavior instability measures. By summing these 35 values, the cumulative driving risk for the road segment at the current time step is obtained. It is 48.6.

[0102] The real-time traffic flow at a specified road cross-section, expressed in vehicles per hour (vph), is calculated by collecting traffic data from sensors (or virtual loops) deployed at the end of the road cross-section within a fixed time window. The data collection time window is typically set to 5 minutes to balance real-time performance with smoothness. For example, if the system detects that 150 vehicles passed the exit monitoring point of the cross-section in the past 5 minutes, the current real-time traffic flow is calculated as follows: Vehicles per hour.

[0103] The capacity of a road cross-section is specified in "vehicles per hour" (vph). This is a static parameter pre-calibrated according to road design standards and physical conditions, stored in the system's road network configuration database. Its calibration process follows traffic engineering standards and comprehensively considers factors such as the number of lanes, lane width, shoulder width, road longitudinal slope, proportion of large vehicles, and lateral disturbances. For example, for a two-lane, one-way urban arterial road cross-section with a design speed of 80 km / h, no significant gradient, and no disturbances such as bus stops, its basic single-lane capacity is 1900 vehicles per hour. Therefore, the total capacity of this two-lane cross-section is calibrated as [value missing]. Vehicles per hour.

[0104] The average vehicle speed for a specified road cross-section, expressed in kilometers per hour (km / h), is calculated by averaging the instantaneous speeds of all vehicles within that cross-section at the current time step. It reflects the immediate macroscopic speed of the traffic flow. The system first obtains the set of instantaneous speeds of all vehicles within the cross-section and then calculates their average. For example, if the system detects 35 vehicles within the cross-section at the current time step, the average speed of the cross-section is obtained by averaging the instantaneous speeds of these 35 vehicles. km / h.

[0105] The optimal speed set for a road section, measured in kilometers per hour (km / h), is the speed at which the traffic flow at the road section reaches its maximum (i.e., its capacity). It is not the road's speed limit. It is obtained through statistical analysis of long-term historical traffic data (speed-flow relationship) for that section. Specifically, a speed-flow scatter plot is plotted, and a curve describing the relationship is fitted (such as the Greenberg or Underwood model). The speed corresponding to the peak flow point on the curve is the optimal speed. For example, analysis of three consecutive months of historical data for a target section reveals that the hourly flow is highest when the average speed is around 65 km / h. Therefore, the optimal speed for this section is set to [missing value]. km / h.

[0106] The congestion sensitivity factor is a dimensionless weighting coefficient used to adjust the sensitivity of the traffic flow conflict risk index to the degree of traffic congestion (i.e., the ratio of flow to capacity). The value of this factor is determined by regression analysis of historical data, which includes historical records of traffic congestion levels and the frequency of traffic accidents or serious traffic conflict events that occurred during the same period. For road sections (such as ramp merging areas) where congestion is likely to trigger chain reactions and accidents once it occurs. The value will be higher. The calibration process involves constructing a logistic regression model with congestion level as the independent variable and accident rate as the dependent variable. The coefficient of the congestion level term in the model is... Based on the reference value and analysis of historical accident data for the target road section, it was found that for every 0.1 increase in congestion, the logarithmic probability of accident risk increases by 0.15. Therefore, a congestion sensitivity factor was established. .

[0107] The speed deviation sensitivity factor is a dimensionless weighting coefficient used to adjust the sensitivity of the traffic flow conflict risk index to the degree to which the average vehicle speed deviates from the optimal speed. (Square term) This ensures that both driving below the optimal speed (which usually means increased congestion) and driving above the optimal speed (which means reduced safety margin) increase risk. The calibration process and Similarly, by analyzing the relationship between historical speed data and accident data, it can be determined that for road sections such as highways where high speed uniformity is required, The value will be higher. After analyzing historical data of the target road section, a stronger correlation was found between speed deviation and accident risk. Therefore, a speed deviation sensitivity factor was set. .

[0108] Calculation process:

[0109] The example values ​​of the aforementioned parameters are substituted into the formula for calculation, where... , vehicles / hour vehicles / hour km / h km / h , .

[0110] Congestion items within the calculation index:

[0111] ;

[0112] Calculate the velocity offset term within the exponential part:

[0113] ;

[0114] Calculate the complete exponent value:

[0115] ;

[0116] Calculate the exponential amplification factor:

[0117] ;

[0118] Calculate the final traffic flow conflict risk index :

[0119] ;

[0120] The result indicates that the current traffic flow conflict risk index of the road section is 99.33. This value is a comprehensive risk score that combines individual driving instability at the micro level (basic risk value of 48.6) with traffic flow status at the macro level (congestion and speed deviation work together to amplify the basic risk by about 2.04 times). The system compares this result with preset multi-level risk thresholds, such as: low risk (0-50), medium risk (50-100), high risk (100-200), and very high risk (>200). The currently calculated value of 99.33 is at the upper limit of the medium risk range and close to the high risk level.

[0121] The steps for obtaining the preset control strategy set are as follows:

[0122] Based on the traffic flow conflict risk index of road cross-section, a predefined multi-level risk threshold set is called, each risk threshold in the risk threshold set is extracted one by one, the traffic flow conflict risk index is compared with the value of each risk threshold, the risk level matching the current traffic flow conflict risk index is determined, and the road cross-section risk level judgment result is generated.

[0123] Based on the risk level determination results of the road section, the risk level determination results are used as query conditions to retrieve the pre-defined control strategy set corresponding to the risk level from the associated control strategy library, forming a preset control strategy set applicable to the current road section risk level.

[0124] Specifically, based on the traffic flow conflict risk index of road cross-sections, the system calls a predefined multi-level risk threshold set. This set is constructed based on a comprehensive analysis of historical traffic data, accident data, and traffic simulation results for a specific road cross-section. Specifically, it first collects the calculated traffic flow conflict risk index values ​​for all time points within the past year for that cross-section, forming a large historical index dataset. Simultaneously, it collects traffic accident reports from the same period, marking the risk index value 5 minutes before each accident as a high-risk sample. Then, it uses the percentile method to divide the historical index dataset. For example, the 70th percentile of the dataset is used as the boundary between "low risk" and "medium risk," the 90th percentile as the boundary between "medium risk" and "high risk," and the 98th percentile as the boundary between "high risk" and "extremely high risk." These thresholds are then fine-tuned based on the distribution of accident samples to ensure that the vast majority of accident samples fall into the high-risk or extremely high-risk range, ultimately forming a clear threshold set, for example: ["Low risk": (0, 50], "Medium risk": (50,

[100] , “High Risk”: (100, 200], “Extremely High Risk”: (>200)], The system compares the currently calculated traffic flow conflict risk index, for example, 99.33, with the thresholds in this set one by one. The comparison process starts from the lowest risk level. First, it checks whether 99.33 is greater than 0 and less than or equal to 50. The result is no. Then it checks whether it is greater than 50 and less than or equal to 100. The result is yes. Once a matching interval is found, the comparison process ends, and the system determines the current risk level as “medium risk” and generates the road section risk level determination result.

[0125] Based on the road section risk level assessment result generated in the previous step, such as "medium risk," the system uses this assessment result as the unique query keyword to access a built-in associated control strategy library. This library is a structured database that pre-stores control strategy sets for different risk levels, road types, and time periods (such as peak, off-peak, and nighttime). These strategy sets were jointly developed by a team of traffic engineers and traffic management experts based on traffic engineering theory, simulation optimization results, and management practice experience. Each strategy has undergone rigorous verification and testing. For example, the library stores a strategy set corresponding to the "medium risk" level, which might include the following: Content: Strategy 1, for downstream traffic lights, the instruction is "extend cycle by 10%", and the parameter is "green light ratio tilted towards main road direction by 5%"; Strategy 2, for upstream variable message signs, the instruction is "display text", and the parameter is "traffic congestion ahead, please slow down"; Strategy 3, for traffic broadcast system interface, the instruction is "broadcast traffic conditions", and the parameter is "moderate congestion risk on XX section, detour recommended". When the system performs a query operation, it uses "moderate risk" as an index to accurately match and retrieve this set of multiple specific response measures from the database. This set is a preset control strategy set consisting of multiple sub-strategies customized for the current situation.

[0126] The steps for obtaining adaptive traffic control commands are as follows:

[0127] Based on a preset control strategy set, strategy instructions are extracted one by one from the control strategy set and converted into instruction data streams that can be executed by traffic control equipment. Based on the instruction data streams, control actions are generated for traffic lights at downstream intersections or variable message signs ahead, forming adaptive traffic control instructions.

[0128] Specifically, based on the pre-set control strategy set applicable to the current road section risk level, the system begins to parse and execute the strategy instructions one by one, transforming the high-level strategy descriptions into specific instructions that the underlying traffic control equipment can recognize and execute. For the first strategy in the pre-set control strategy set, "For downstream traffic lights, the instruction is 'cycle extended by 10%', and the parameter is 'green light ratio tilted towards the main road direction by 5%'", the system first connects to the traffic signal controller at the downstream intersection to query its current signal timing scheme. For example, if the current scheme is a cycle of 120 seconds and a green light time of 50 seconds for the main road, the system calculates the new timing parameters according to the instruction: the new cycle is 120 seconds × (1 + 10%) = 132 seconds, and the new green light time for the main road is 50 seconds × (1 + 5%) = After 52.5 seconds (rounded to 53 seconds), these new parameters are encapsulated into command data packets conforming to the communication protocol of this type of traffic signal (such as the NTCIP protocol) and sent to the traffic signal via the network. For the second strategy, "For the upstream variable information sign, the instruction is 'display text', and the parameter is 'Traffic congestion ahead, please slow down'", the system connects to the variable information sign (VMS) controller at the designated location, encodes the text parameter "Traffic congestion ahead, please slow down" into a format supported by the VMS controller (e.g., a specific encoding containing information such as font, color, and display mode), generates a control data stream and sends it. After receiving the data, the VMS will update the display content. The system processes all strategy instructions in sequence according to the order of the strategy set, generates and sends corresponding control actions for each target device. These actions together constitute a real-time, closed-loop adaptive traffic control instruction for the current traffic situation.

[0129] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A traffic monitoring method for vehicle tracking and early warning, characterized in that, Includes the following steps: The system acquires real-time location, speed, and acceleration data of all vehicles within the monitored road area. Based on the spatial distance and time window settings between vehicles, adjacent vehicles are aggregated into dynamic traffic micro-clusters to establish a regional motion state benchmark. Based on the regional motion state benchmark, the motion vector of a single vehicle is extracted, the individual driving deviation value is calculated, and then the continuous individual driving deviation values ​​of a single vehicle within a set time period are summed to obtain a vehicle behavior instability measure. Based on the vehicle behavior instability measure of each vehicle, the vehicle behavior instability measures of all vehicles on the target road section are summed to obtain the cumulative driving risk of the road section. Then, the cumulative driving risk of the road section is calculated with the real-time traffic flow, average speed and road capacity of the road section to generate the traffic flow conflict risk index of the road section. Based on the traffic flow conflict risk index of the road cross section, the traffic flow conflict risk index is compared with the preset multi-level risk thresholds one by one. According to the comparison results, a preset control strategy set is matched and selected from the associated control strategy library. Based on the preset control strategy set, an adaptive traffic control instruction for the traffic lights or variable information signs at the downstream intersection is generated. The formula for calculating the vehicle behavior instability metric is as follows: ； in, A measure of vehicle behavior instability. This represents the total number of time steps within the time period. For time steps The Mahalanobis distance represents the degree of deviation between the vehicle motion vector and the average motion vector of the dynamic traffic micro-cluster under the covariance structure. For time steps The velocity vector magnitude represents the instantaneous speed of the vehicle. Let be the average velocity vector magnitude of the dynamic traffic micro-cluster, and represent the average speed of the cluster. It is a constant positive number; The formula for calculating the traffic flow conflict risk index is as follows: ； in, Traffic flow conflict risk index This represents the cumulative driving risk on the road segment. To provide the real-time traffic flow at a specified road section at the current moment, To specify the traffic capacity of a road section, To specify the average vehicle speed across a road section, The optimal vehicle speed set for the cross-section. As a congestion-sensitive factor, This is a velocity offset sensitivity factor.

2. The traffic monitoring method for vehicle tracking and early warning according to claim 1, characterized in that, The steps for obtaining the regional motion state reference are as follows: By deploying video sensors and millimeter-wave radar within the road area, all vehicles currently traveling within the road area are simultaneously detected in real time, acquiring real-time position data, speed data, and acceleration data for each vehicle, and using timestamps to uniformly mark the data of each vehicle, generating a set of real-time motion parameters for the vehicles. Based on the set of real-time vehicle motion parameters, according to the preset spatial distance threshold between vehicles and the time window threshold for collecting vehicle motion status, vehicles whose spatial distance is lower than the spatial distance threshold and whose data collection time difference is within the time window threshold range are identified as adjacent vehicles, and the adjacent vehicles are aggregated to form multiple vehicle groups to generate dynamic traffic micro-clusters. Based on the dynamic traffic micro-clusters, the speed and acceleration data of all vehicles in each dynamic traffic micro-cluster are extracted to obtain the regional motion state benchmark.

3. The traffic monitoring method for vehicle tracking and early warning according to claim 1, characterized in that, The steps for obtaining the individual driving deviation value are as follows: Based on the aforementioned regional motion state benchmark, the velocity and acceleration components of a single vehicle at each time step are extracted, and a four-dimensional motion vector of the vehicle at each time step is constructed. The four-dimensional motion vectors corresponding to all time steps are integrated to form a continuous sequence of vehicle motion vectors, thus obtaining a single vehicle motion vector sequence. Based on the single vehicle motion vector sequence, the Mahalanobis distance between the vehicle motion vector at each time step and the average motion vector of the dynamic traffic micro-cluster to which it belongs is calculated to form the individual driving deviation value.

4. The traffic monitoring method for vehicle tracking and early warning according to claim 1, characterized in that, The steps for obtaining the vehicle behavior instability metric are as follows: Based on the individual driving deviation value, a vehicle behavior instability metric is calculated.

5. The traffic monitoring method for vehicle tracking and early warning according to claim 1, characterized in that, The steps for obtaining the cumulative driving risk of the aforementioned road segment are as follows: Based on the vehicle behavior instability measure of all vehicles within the target road section, extract the vehicle behavior instability measure values ​​of all vehicles at the current time step, sum the extracted values, and obtain the cumulative amount of road segment driving risk of the current road section at the current time step.

6. The traffic monitoring method for vehicle tracking and early warning according to claim 1, characterized in that, The steps for obtaining the traffic flow conflict risk index are as follows: Based on the cumulative driving risk of the road segment, the real-time traffic flow, road capacity and average vehicle speed of the corresponding time step of the road section are retrieved and included in the calculation input set together with the cumulative driving risk of the road segment to obtain the joint variable set constituting the risk index calculation. Based on the aforementioned set of joint variables, a traffic flow conflict risk index is calculated.

7. The traffic monitoring method for vehicle tracking and early warning according to claim 1, characterized in that, The steps for obtaining the preset control strategy set are as follows: Based on the traffic flow conflict risk index of the road section, a predefined multi-level risk threshold set is called, and each risk threshold in the risk threshold set is extracted one by one. The traffic flow conflict risk index is compared with the value of each risk threshold to determine the risk level that matches the current traffic flow conflict risk index and generate the road section risk level judgment result. Based on the risk level determination result of the road section, the risk level determination result is used as the query condition to retrieve the pre-defined control strategy set corresponding to the risk level from the associated control strategy library, forming a preset control strategy set applicable to the current road section risk level.

8. The traffic monitoring method for vehicle tracking and early warning according to claim 1, characterized in that, The steps for obtaining the adaptive traffic control command are as follows: Based on the preset control strategy set, strategy instructions are extracted one by one from the control strategy set and converted into instruction data streams that can be executed by traffic control equipment. Control actions are generated one by one for traffic lights or variable information signs at downstream intersections according to the instruction data streams, forming adaptive traffic control instructions.

9. A traffic monitoring system for a traffic monitoring method for vehicle tracking and early warning according to any one of claims 1-8, characterized in that, include: The data acquisition module acquires real-time location, speed, and acceleration data of all vehicles within the monitored road area. Based on the spatial distance and time window settings between vehicles, it aggregates adjacent vehicles into dynamic traffic micro-clusters and establishes a regional motion state benchmark. The behavior assessment module extracts the motion vector of a single vehicle based on the regional motion state benchmark, calculates the individual driving deviation value, and then sums the continuous individual driving deviation values ​​of a single vehicle within a set time period to obtain a vehicle behavior instability measure. The risk calculation module sums the vehicle behavior instability measures of all vehicles on the target road section based on the vehicle behavior instability measures of each vehicle to obtain the cumulative driving risk of the road section. Then, the cumulative driving risk of the road section is calculated with the real-time traffic flow, average speed and road capacity of the road section to generate the traffic flow conflict risk index of the road section. The control instruction module compares the traffic flow conflict risk index with preset multi-level risk thresholds one by one based on the traffic flow conflict risk index of the road cross section. According to the comparison results, it matches and selects a preset control strategy set from the associated control strategy library. Based on the preset control strategy set, it generates adaptive traffic control instructions for traffic lights or variable information signs at downstream intersections.

Images