Expressway abnormal target trajectory prediction method and system based on knowledge graph

Abstract

The invention relates to the technical field of intelligent traffic, and discloses an expressway abnormal target trajectory prediction method and system based on a knowledge graph, and the method comprises the steps: collecting vehicle, environment, road and traffic event data through a multi-source sensor, carrying out the data cleaning and conversion, guaranteeing the data quality, extracting key features from the multi-source data, and carrying out the prediction of the abnormal target trajectory of an expressway. The method comprises the following steps: constructing a highway knowledge graph, integrating multi-dimensional information, calculating a smoothness index, a safety index and a path deviation index based on the knowledge graph, comprehensively generating a track state coefficient, introducing an environment correction coefficient and a road state correction coefficient, dynamically correcting the track state coefficient, and comparing the corrected coefficient with a threshold value. According to the method, multiple factors such as vehicle behaviors, environments and roads are integrated, the prediction accuracy is improved, the model is adjusted through real-time environment and road data, the method adapts to complex scenes, and efficient track anomaly prediction is achieved.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, specifically to a method and system for predicting the trajectory of abnormal targets on highways based on knowledge graphs. Background Technology

[0002] Currently, abnormal target detection on highways largely relies on traditional monitoring equipment and single sensors, which have significant limitations in data processing. Traditional methods struggle to effectively integrate multi-source heterogeneous data, such as vehicle information, surveillance video data, and environmental data, resulting in insufficient comprehensive target perception capabilities. This leads to delays in identifying and predicting abnormal trajectories, hindering early warning and increasing the risk of traffic accidents. Moreover, when encountering special circumstances such as sudden weather or road congestion, traditional methods cannot quickly adjust the prediction model, resulting in large deviations in the prediction results and further reducing the accuracy of the prediction. Therefore, this invention proposes a method and system for predicting abnormal target trajectories on highways based on knowledge graphs. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for predicting the trajectory of abnormal targets on highways based on knowledge graphs, thereby solving the aforementioned technical problems.

[0004] A knowledge graph-based method for predicting the trajectory of abnormal targets on highways, the method comprising the following steps: Step S1: Collect multi-source data of the vehicle to be tested and process the collected data; Step S2: Extract multi-source data feature information of the vehicle to be detected and construct a highway knowledge graph; Step S3: Analyze the trajectory of the vehicle to be detected based on the constructed knowledge graph; Step S4: Based on the analysis results, predict whether the trajectory of the vehicle to be detected is abnormal, and determine whether an early warning needs to be issued based on the prediction results.

[0005] As a further description of the technical solution of the present invention, step S1 of the data processing includes data cleaning and data transformation; The data cleaning employs a hybrid algorithm based on rules and machine learning to remove noisy data, missing data, and outliers. The data conversion process transforms data from different formats into a format suitable for processing.

[0006] As a further description of the technical solution of the present invention, the working process of step S2 includes: The multi-source data feature information includes: vehicle data to be detected, environmental data, and road data; The vehicle data to be detected includes: path feature parameters and behavior feature parameters; The environmental data includes: temperature, humidity, visibility, and wind speed; The road data includes traffic flow and road surface conditions.

[0007] As a further description of the technical solution of the present invention, the working process of step S3 includes: The trajectory curvature and acceleration rate of change of the vehicle under test are periodically acquired during its movement, and a smoothness index coefficient calculation model is constructed, with the expression as follows: ; In the formula, n is the total number of vehicle trajectory data points to be detected, and i belongs to n. It is the trajectory curvature at time i. It is the rate of change of acceleration at time i. and These are the maximum allowable values ​​for curvature and acceleration, respectively. The number of dangerous driving behaviors during the operation of the vehicle under test is obtained, and a safety index coefficient calculation model is constructed, with the following expression: ; In the formula, m represents the total number of dangerous behaviors, and k belongs to m. Let k be the number of dangerous behaviors. The maximum number of events allowed for the k-th type of dangerous behavior is set at C=1 if no dangerous behavior occurs. The perpendicular distance from the vehicle to be detected to the preset path is periodically acquired during its travel, and a path deviation index coefficient calculation model is constructed, with the expression as follows: ; In the formula, The vertical distance from the vehicle to be detected to the preset path at time i. Maximum permissible deviation distance; The trajectory state coefficient is calculated based on the smoothness index coefficient, safety index coefficient, and path deviation index coefficient.

[0008] As a further description of the technical solution of the present invention, the process of calculating the trajectory state coefficient based on the smoothness index coefficient, the safety index coefficient, and the path deviation index coefficient includes: Construct a trajectory state coefficient calculation model, the expression of which is: ; In the formula, , and These represent the weighting coefficients corresponding to the smoothness index coefficient, the safety index coefficient, and the path deviation index coefficient, respectively. It is the trajectory state coefficient.

[0009] As a further description of the technical solution of the present invention, the working process of step S4 includes: The ambient temperature, humidity, visibility, and wind speed during the vehicle's operation are obtained, and an environmental correction coefficient calculation model is constructed, expressed as follows: ; In the formula, , , and These represent the ambient temperature, humidity, visibility, and wind speed during the operation of the vehicle under test. , , and These represent the standard values ​​for ambient temperature, humidity, visibility, and wind speed, respectively. , , and These represent the reference values ​​for the differences in ambient temperature, humidity, visibility, and wind speed, respectively. , , and These represent the weighting coefficients for ambient temperature, humidity, visibility, and wind speed, respectively. This is the environmental correction factor.

[0010] As a further description of the technical solution of the present invention, the working process of step S4 also includes: Traffic flow and road conditions during the vehicle's journey are obtained, and a road condition correction coefficient calculation model is constructed, expressed as follows: ; In the formula, The traffic flow during the operation of the vehicle to be tested. The flatness index is the smoothness index of the vehicle under test during its operation. and These are the weighting coefficients corresponding to traffic flow and road conditions. The standard traffic flow set for the system, This is the road condition correction factor.

[0011] As a further description of the technical solution of the present invention, the working process of step S4 also includes: The trajectory state coefficient is corrected by combining environmental correction coefficients and road condition correction coefficients, and the expression is as follows: ; In the formula, and For conversion factors, For the corrected trajectory state coefficients, Compared with the trajectory state coefficient threshold set by the system, if If the value is less than the corresponding threshold set by the system, the trajectory status of the vehicle to be detected is predicted to be abnormal.

[0012] A knowledge graph-based system for predicting the trajectory of abnormal targets on highways, the system comprising: The data acquisition and processing module is used to collect data on vehicles, environment, roads and traffic events through multi-source sensors, and to clean and transform the data to ensure data quality. The feature extraction and knowledge graph construction module is used to extract key features from multi-source data, construct a highway knowledge graph, and integrate multi-dimensional information. The trajectory analysis module is used to calculate smoothness indicators, safety indicators (and path deviation indicators) based on knowledge graphs, and to generate trajectory state coefficients in combination. The environment and road correction module is used to introduce environmental correction coefficients and road condition correction coefficients to dynamically correct the trajectory condition coefficients. The anomaly prediction and early warning module compares the corrected coefficients with the threshold. If the coefficients are lower than the threshold, the module predicts an abnormal trajectory and triggers an early warning.

[0013] The beneficial effects of this invention are: This invention collects vehicle, environmental, road, and traffic event data from multiple sources of sensors (such as high-definition cameras, thermal imagers, and drones), performs data cleaning and transformation to ensure data quality, extracts key features (such as vehicle behavior, environmental parameters, and road conditions) from the multi-source data, constructs a highway knowledge graph, integrates multi-dimensional information, and calculates smoothness indicators (curvature, rate of change of acceleration), safety indicators (number of dangerous behaviors), and path deviation indicators (vertical distance) based on the knowledge graph. It then comprehensively generates trajectory state coefficients, introduces environmental correction coefficients and road condition correction coefficients, and dynamically corrects the trajectory state coefficients. The corrected coefficients are compared with a threshold; if they fall below the threshold, the trajectory is deemed abnormal and an early warning is triggered. This invention integrates multiple factors such as vehicle behavior, environment, and road conditions to improve prediction accuracy. By adjusting the model with real-time environmental and road data, it adapts to complex scenarios and achieves efficient trajectory anomaly prediction. Attached Figure Description

[0014] The invention will now be further described with reference to the accompanying drawings.

[0015] Figure 1 This is a partial flowchart of the knowledge graph-based method for predicting the trajectory of abnormal targets on highways provided by the present invention. Detailed Implementation

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

[0017] Please see Figure 1 As shown, a knowledge graph-based method for predicting the trajectory of abnormal targets on highways includes the following steps: Step S1: Collect multi-source data of the vehicle to be tested and process the collected data; Step S2: Extract multi-source data feature information of the vehicle to be detected and construct a highway knowledge graph; Step S3: Analyze the trajectory of the vehicle to be detected based on the constructed knowledge graph; Step S4: Based on the analysis results, predict whether the trajectory of the vehicle to be detected is abnormal, and determine whether an early warning needs to be issued based on the prediction results.

[0018] Through the above technical solution, this invention collects vehicle, environmental, road, and traffic event data using multi-source sensors (such as high-definition cameras, thermal imagers, and drones), performs data cleaning and transformation to ensure data quality, extracts key features (such as vehicle behavior, environmental parameters, and road conditions) from the multi-source data, constructs a highway knowledge graph, integrates multi-dimensional information, and calculates smoothness indicators (curvature, rate of change of acceleration), safety indicators (number of dangerous behaviors), and path deviation indicators (vertical distance) based on the knowledge graph. A trajectory state coefficient is then generated, and environmental and road condition correction coefficients are introduced to dynamically correct the trajectory state coefficient. The corrected coefficient is compared with a threshold; if it falls below the threshold, a trajectory anomaly is determined and an early warning is triggered. This invention integrates multiple factors such as vehicle behavior, environment, and road conditions to improve prediction accuracy. By adjusting the model using real-time environmental and road data, it adapts to complex scenarios and achieves efficient trajectory anomaly prediction.

[0019] Implementation method 1: Hardware deployment, installing sensors (such as cameras and weather stations) at key nodes of the highway, and using drones to assist in collecting route data; Data processing employs a hybrid algorithm to clean noisy data and standardize data format; The model calculates smoothness, safety, and deviation indices in real time, and generates initial trajectory state coefficients by combining weights; then it is dynamically corrected by environmental and road factors. The system outputs real-time monitoring of the corrected trajectory status coefficients, and notifies inspection personnel of any abnormalities via sound, light, or platform notifications.

[0020] As a further description of the technical solution of the present invention, step S1 of the data processing includes data cleaning and data transformation; The data cleaning employs a hybrid algorithm based on rules and machine learning to remove noisy data, missing data, and outliers. The data conversion process transforms data from different formats into a format suitable for processing.

[0021] As a further description of the technical solution of the present invention, the working process of step S2 includes: The multi-source data feature information includes: vehicle data to be detected, environmental data, and road data; The vehicle data to be detected includes: path feature parameters and behavior feature parameters; The environmental data includes: temperature, humidity, visibility, and wind speed; The road data includes traffic flow and road surface conditions.

[0022] As a further description of the technical solution of the present invention, the working process of step S3 includes: The trajectory curvature and acceleration rate of change of the vehicle under test are periodically acquired during its movement, and a smoothness index coefficient calculation model is constructed, with the expression as follows: ; In the formula, n is the total number of vehicle trajectory data points to be detected, and i belongs to n. It is the trajectory curvature at time i. It is the rate of change of acceleration at time i. and These are the maximum allowable values ​​for curvature and acceleration, respectively. The number of dangerous driving behaviors during the operation of the vehicle under test is obtained, and a safety index coefficient calculation model is constructed, with the following expression: ; In the formula, m represents the total number of dangerous behaviors, and k belongs to m. Let k be the number of dangerous behaviors. The maximum number of events allowed for the k-th type of dangerous behavior is set at C=1 if no dangerous behavior occurs. The perpendicular distance from the vehicle to be detected to the preset path is periodically acquired during its travel, and a path deviation index coefficient calculation model is constructed, with the expression as follows: ; In the formula, The vertical distance from the vehicle to be detected to the preset path at time i. Maximum permissible deviation distance; The trajectory state coefficient is calculated based on the smoothness index coefficient, safety index coefficient, and path deviation index coefficient.

[0023] As a further description of the technical solution of the present invention, the process of calculating the trajectory state coefficient based on the smoothness index coefficient, the safety index coefficient, and the path deviation index coefficient includes: Construct a trajectory state coefficient calculation model, the expression of which is: ; In the formula, , and These represent the weighting coefficients corresponding to the smoothness index coefficient, the safety index coefficient, and the path deviation index coefficient, respectively. It is the trajectory state coefficient.

[0024] Through the above technical solution, this embodiment provides a method for calculating trajectory state coefficients, using the formula... Calculate the curvature and rate of change of acceleration of the vehicle trajectory to assess the smoothness of the driving trajectory. If the curvature or acceleration changes too much, it indicates that the vehicle may be exhibiting abnormal driving behavior (such as sharp turns, rapid acceleration / deceleration). The larger the value of S, the smoother the trajectory; the closer it is to 0, the higher the probability of abnormality. This is achieved through the formula... The system tracks dangerous driving behaviors of vehicles (such as frequent lane changes and sudden braking) and calculates a safety factor. If no dangerous behavior occurs, C=1; if the number of dangerous behaviors approaches a threshold, C approaches 0. The final value is determined using the formula... The system calculates the vertical distance between the vehicle and the preset path to assess whether it deviates from the normal driving route. The greater the deviation distance, the higher the D value and the greater the risk of anomaly. The system comprehensively considers smoothness, safety, and path deviation indicators to calculate the trajectory state coefficient, which reflects the overall degree of abnormality in the vehicle's driving.

[0025] Implementation method: Data acquisition, which involves collecting data such as vehicle trajectory, acceleration, and steering angle through onboard sensors (such as GPS, IMU), roadside cameras, and other equipment; Real-time calculation of indicators, periodic (e.g., every second or every 100ms) calculation of S, C, D, and updating of trajectory state system P; Dynamic weight adjustment adjusts the weight coefficients based on different road types (such as highways and curves). For example, in road sections with many curves, the weight coefficient is increased. (Smoothness weighting) In sections of road with high traffic volume, increase (Safety behavior weights).

[0026] As a further description of the technical solution of the present invention, the working process of step S4 includes: The ambient temperature, humidity, visibility, and wind speed during the vehicle's operation are obtained, and an environmental correction coefficient calculation model is constructed, expressed as follows: ; In the formula, , , and These represent the ambient temperature, humidity, visibility, and wind speed during the operation of the vehicle under test. , , and These represent the standard values ​​for ambient temperature, humidity, visibility, and wind speed, respectively. , , and These represent the reference values ​​for the differences in ambient temperature, humidity, visibility, and wind speed, respectively. , , and These represent the weighting coefficients for ambient temperature, humidity, visibility, and wind speed, respectively. This is the environmental correction factor.

[0027] As a further description of the technical solution of the present invention, the working process of step S4 also includes: Traffic flow and road conditions during the vehicle's journey are obtained, and a road condition correction coefficient calculation model is constructed, expressed as follows: ; In the formula, The traffic flow during the operation of the vehicle to be tested. The flatness index is the smoothness index of the vehicle under test during its operation. and These are the weighting coefficients corresponding to traffic flow and road conditions. The standard traffic flow set for the system, This is the road condition correction factor.

[0028] As a further description of the technical solution of the present invention, the working process of step S4 also includes: The trajectory state coefficient is corrected by combining environmental correction coefficients and road condition correction coefficients, and the expression is as follows: ; In the formula, and For conversion factors, For the corrected trajectory state coefficients, Compared with the trajectory state coefficient threshold set by the system, if If the value is less than the corresponding threshold set by the system, the trajectory status of the vehicle to be detected is predicted to be abnormal.

[0029] Through the above technical solution, through the formula Environmental correction coefficients are calculated by monitoring environmental parameters such as temperature (T), humidity (H), visibility (V), and wind speed (W) during vehicle operation, and calculating their impact on the vehicle's trajectory. If environmental parameters (such as heavy fog or strong winds) deviate from the standard, then... An increase indicates that the environment has a significant impact on driving behavior, as shown by the formula. The road condition correction coefficient is calculated, and the impact of road conditions on vehicle trajectory is assessed by combining traffic flow and road surface smoothness. Finally, the initial trajectory state coefficient is dynamically adjusted by combining the environmental correction coefficient (E) and the road correction coefficient (R) to obtain the final correction value. If E or R is large (adverse environment or congested road), the initial trajectory state coefficient will be adjusted accordingly to avoid misjudging normal trajectories as abnormal ones.

[0030] Implementation method: Data acquisition: Environmental data is obtained in real time through meteorological sensors, visibility detectors, anemometers, etc.; road data is obtained through cameras, radar or intelligent transportation systems (ITS); and road surface smoothness is measured through road detection equipment. Real-time calculation of correction coefficients and periodic calculation of E and R ensure dynamic adaptation to changes in environment and road conditions; The trajectory status is dynamically adjusted based on E and R. For example, in foggy weather (high E), the criteria for judging trajectory anomalies are relaxed to avoid misjudgment due to low visibility. In congested road sections (high R), the criteria for judging trajectory anomalies are relaxed to reduce false alarms caused by low-speed driving.

[0031] Early warning decision-making, if If the value is less than the corresponding threshold set by the system, the system will trigger an alert.

[0032] A knowledge graph-based system for predicting the trajectory of abnormal targets on highways, the system comprising: The data acquisition and processing module is used to collect data on vehicles, environment, roads and traffic events through multi-source sensors, and to clean and transform the data to ensure data quality. The feature extraction and knowledge graph construction module is used to extract key features from multi-source data, construct a highway knowledge graph, and integrate multi-dimensional information. The trajectory analysis module is used to calculate smoothness indicators, safety indicators (and path deviation indicators) based on knowledge graphs, and to generate trajectory state coefficients in combination. The environment and road correction module is used to introduce environmental correction coefficients and road condition correction coefficients to dynamically correct the trajectory condition coefficients. The anomaly prediction and early warning module compares the corrected coefficients with the threshold. If the coefficients are lower than the threshold, the module predicts an abnormal trajectory and triggers an early warning.

[0033] It should be noted that all weight coefficients, reference values, and thresholds in this invention are empirical values. The weight coefficients can be modified by combining the characteristics of the data type to be evaluated. The feature parameters extracted in this invention have all been normalized, and all parameters are calculated under the same standard.

[0034] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for predicting the trajectory of abnormal targets on highways based on knowledge graphs, characterized in that, The method includes the following steps: Step S1: Collect multi-source data of the vehicle to be tested and process the collected data; Step S2: Extract multi-source data feature information of the vehicle to be detected and construct a highway knowledge graph; Step S3: Analyze the trajectory of the vehicle to be detected based on the constructed knowledge graph; Step S4: Based on the analysis results, predict whether the trajectory of the vehicle to be detected is abnormal, and determine whether an early warning needs to be issued based on the prediction results.

2. The method for predicting the trajectory of abnormal targets on highways based on knowledge graphs according to claim 1, characterized in that, The data processing step S1 includes data cleaning and data transformation. The data cleaning employs a hybrid algorithm based on rules and machine learning to remove noisy data, missing data, and outliers. The data conversion process transforms data from different formats into a format suitable for processing.

3. The method for predicting the trajectory of abnormal targets on highways based on knowledge graphs according to claim 2, characterized in that, The working process of step S2 includes: The multi-source data feature information includes: vehicle data to be detected, environmental data, and road data; The vehicle data to be detected includes: path feature parameters and behavior feature parameters; The environmental data includes: temperature, humidity, visibility, and wind speed; The road data includes traffic flow and road surface conditions.

4. The method for predicting the trajectory of abnormal targets on highways based on knowledge graphs according to claim 2, characterized in that, The working process of step S3 includes: The trajectory curvature and acceleration rate of change of the vehicle under test are periodically acquired during its movement, and a smoothness index coefficient calculation model is constructed, with the expression as follows: ； In the formula, n is the total number of vehicle trajectory data points to be detected, and i belongs to n. It is the trajectory curvature at time i. It is the rate of change of acceleration at time i. and These are the maximum allowable values ​​for curvature and acceleration, respectively. The number of dangerous driving behaviors during the operation of the vehicle under test is obtained, and a safety index coefficient calculation model is constructed, with the following expression: ; In the formula, m represents the total number of dangerous behaviors, and k belongs to m. Let k be the number of dangerous behaviors. The maximum number of events allowed for the k-th type of dangerous behavior is set at C=1 if no dangerous behavior occurs. The perpendicular distance from the vehicle to be detected to the preset path is periodically acquired during its travel, and a path deviation index coefficient calculation model is constructed, with the expression as follows: ； In the formula, The vertical distance from the vehicle to be detected to the preset path at time i. Maximum permissible deviation distance; The trajectory state coefficient is calculated based on the smoothness index coefficient, safety index coefficient, and path deviation index coefficient.

5. The method for predicting the trajectory of abnormal targets on highways based on knowledge graphs according to claim 4, characterized in that, The process of calculating the trajectory state coefficient based on the smoothness index coefficient, safety index coefficient, and path deviation index coefficient includes: Construct a trajectory state coefficient calculation model, the expression of which is: ； In the formula, , and These represent the weighting coefficients corresponding to the smoothness index coefficient, the safety index coefficient, and the path deviation index coefficient, respectively. It is the trajectory state coefficient.

6. The method for predicting the trajectory of abnormal targets on highways based on knowledge graphs according to claim 2, characterized in that, The working process of step S4 includes: The ambient temperature, humidity, visibility, and wind speed during the vehicle's operation are obtained, and an environmental correction coefficient calculation model is constructed, expressed as follows: ； In the formula, , , and These represent the ambient temperature, humidity, visibility, and wind speed during the operation of the vehicle under test. , , and These represent the standard values ​​for ambient temperature, humidity, visibility, and wind speed, respectively. , , and These represent the reference values ​​for the differences in ambient temperature, humidity, visibility, and wind speed, respectively. , , and These represent the weighting coefficients for ambient temperature, humidity, visibility, and wind speed, respectively. This is the environmental correction factor.

7. The method for predicting the trajectory of abnormal targets on highways based on knowledge graphs according to claim 6, characterized in that, The working process of step S4 also includes: Traffic flow and road conditions during the vehicle's journey are obtained, and a road condition correction coefficient calculation model is constructed, expressed as follows: ； In the formula, The traffic flow during the operation of the vehicle to be tested. The flatness index is the smoothness index of the vehicle under test during its operation. and These are the weighting coefficients corresponding to traffic flow and road conditions. The standard traffic flow set for the system, This is the road condition correction factor.

8. The method for predicting the trajectory of abnormal targets on highways based on knowledge graphs according to claim 7, characterized in that, The working process of step S4 also includes: The trajectory state coefficient is corrected by combining environmental correction coefficients and road condition correction coefficients, and the expression is as follows: ； In the formula, and For conversion factors, For the corrected trajectory state coefficients, Compared with the trajectory state coefficient threshold set by the system, if If the value is less than the corresponding threshold set by the system, the trajectory status of the vehicle to be detected is predicted to be abnormal.

9. A knowledge graph-based highway anomaly target trajectory prediction system, used to implement the knowledge graph-based highway anomaly target trajectory prediction method according to any one of claims 1-8, characterized in that, The system includes: The data acquisition and processing module is used to collect data on vehicles, environment, roads and traffic events through multi-source sensors, and to clean and transform the data to ensure data quality. The feature extraction and knowledge graph construction module is used to extract key features from multi-source data, construct a highway knowledge graph, and integrate multi-dimensional information. The trajectory analysis module is used to calculate smoothness indicators, safety indicators (and path deviation indicators) based on knowledge graphs, and to generate trajectory state coefficients in combination. The environment and road correction module is used to introduce environmental correction coefficients and road condition correction coefficients to dynamically correct the trajectory condition coefficients. The anomaly prediction and early warning module compares the corrected coefficients with the threshold. If the coefficients are lower than the threshold, the module predicts an abnormal trajectory and triggers an early warning.

Images