Terminal area prevailing traffic flow identification method based on trajectory spectral clustering

Abstract

The invention discloses a terminal area prevailing traffic flow recognition method based on trajectory spectral clustering, and the method comprises the steps: data collection, data preprocessing, trajectory similarity calculation, trajectory clustering, traffic flow mode recognition and traffic flow optimization application. The data preprocessing stage is introduced, denoising, missing data filling, time interval unification and other processing are carried out on noise, missing values and abnormal values in traffic data, the consistency and accuracy of trajectory data can be improved through the preprocessing operation, and the reliability of subsequent analysis is ensured; according to the method, a useful traffic mode can be extracted from a large amount of imperfect original data, recognition errors and low-efficiency calculation caused by low data quality are avoided, meanwhile, a spectral clustering algorithm is introduced, a trajectory similarity matrix is processed, data dimensionality reduction is performed in combination with a spectral embedding technology, clustering efficiency is improved, and high-dimensional data processing precision is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic flow identification and analysis, and in particular to a method for identifying prevailing traffic flow in a terminal area based on trajectory spectrum clustering. Background Art

[0002] As urban traffic becomes increasingly complex, real-time identification and prediction of traffic flow has become one of the core requirements of intelligent transportation systems. Existing technologies mostly rely on traditional traffic flow models or traffic-based statistical analysis methods, but these methods are usually difficult to accurately identify and predict the prevailing traffic flow in complex road environments. Although existing trajectory analysis methods can identify single vehicle trajectories, they lack effective clustering and analysis of large-scale, multi-dimensional trajectory data, especially in terminal area applications, where efficiency and accuracy need to be improved urgently.

[0003] The existing technology often has noise, missing values ​​and inconsistency in the collection of traffic data in terminal area traffic flow identification, resulting in poor data quality. The calculation of trajectory similarity may not accurately reflect the differences between different transportation tools and traffic patterns. Traditional clustering methods are prone to inefficiency and low accuracy under high-dimensional data. Summary of the invention

[0004] The main purpose of the present invention is to provide a terminal area prevailing traffic flow identification method based on trajectory spectrum clustering, which can effectively solve the problems of noise, missing values ​​and inconsistency in traffic data collection, resulting in poor data quality. The calculation of trajectory similarity may not accurately reflect the differences between different transportation tools and traffic patterns. Traditional clustering methods are prone to low efficiency and low precision under high-dimensional data.

[0005] To achieve the above object, the technical solution adopted by the present invention is: a method for identifying prevailing traffic flow in a terminal area based on trajectory spectrum clustering, the method comprising:

[0006] S1. Data collection: Based on traffic monitoring equipment, GPS system and vehicle-mounted sensors, the trajectory data of vehicles in the terminal area is obtained, and the trajectory data includes but is not limited to the spatial position, speed, time and acceleration of the vehicle;

[0007] S2, data preprocessing: performing denoising, spatiotemporal alignment and interpolation processing on the trajectory data to improve the consistency and accuracy of the trajectory data;

[0008] S3, trajectory similarity calculation: using the trajectory similarity measurement method, calculate the trajectory similarity of each vehicle in the terminal area and generate a trajectory similarity matrix;

[0009] S4, trajectory clustering: based on the trajectory similarity matrix, a spectral clustering algorithm is used to cluster the trajectory similarity matrix to obtain different traffic flow types;

[0010] S5, traffic flow pattern recognition: based on the trajectory clustering results, identify and analyze the prevailing traffic flow pattern in the terminal area, and identify the characteristics of different traffic flows, including but not limited to high-density flow, low-density flow, uniform flow and congested flow;

[0011] S6. Traffic flow optimization application: Apply the identified traffic flow patterns to the traffic management system.

[0012] Preferably, the data preprocessing includes but is not limited to removing outliers, filling missing data, unifying time intervals and vehicle type labeling, and the vehicle type labeling performs different traffic flow predictions and traffic flow optimization based on different types of vehicles.

[0013] Preferably, the spectral clustering algorithm includes calculating a trajectory similarity matrix and constructing a similarity graph, performing Laplace matrix decomposition on the similarity graph to obtain a eigenvector, and clustering the eigenvector by a clustering algorithm to obtain a traffic flow pattern.

[0014] Preferably, the traffic flow pattern recognition includes analyzing clustering results to identify the time variation characteristics, spatial distribution characteristics and flow type characteristics of the traffic flow.

[0015] Preferably, the trajectory data is collected from vehicles passing through the terminal area in different time periods, and the trajectory data of each vehicle is processed into a time series format, and the trajectory data is collected once per second.

[0016] Preferably, the distance measurement method used in the trajectory similarity calculation is Euclidean distance, dynamic time warping distance or cosine similarity.

[0017] Preferably, the traffic management system includes but is not limited to traffic signal optimization, lane scheduling and flow prediction.

[0018] Preferably, the trajectory clustering uses spectral embedding technology to perform dimensionality reduction processing on trajectory data to improve clustering efficiency and accuracy.

[0019] Compared with the prior art, the present invention has the following beneficial effects:

[0020] 1. In the present invention, the method introduces a data preprocessing stage to perform denoising, filling missing data, unifying time intervals, etc. on the noise, missing values ​​and outliers in the traffic data. These preprocessing operations can significantly improve the consistency and accuracy of trajectory data, thereby ensuring the reliability of subsequent analysis. In this way, the system can extract useful traffic patterns from a large amount of imperfect raw data, avoiding recognition errors and inefficient calculations caused by low data quality.

[0021] 2. In the present invention, the method introduces a spectral clustering algorithm, processes the trajectory similarity matrix, and combines the spectral embedding technology to reduce the data dimension, which not only improves the clustering efficiency, but also optimizes the accuracy of high-dimensional data processing. It can more accurately identify the traffic flow pattern in the terminal area, covering a variety of different flow characteristics such as high-density flow, low-density flow, uniform flow and congested flow, and can provide more accurate traffic flow classification in complex traffic environments, providing more effective data support for traffic management.

[0022] 3. In the present invention, the method accurately identifies and analyzes the temporal changes, spatial distribution and flow characteristics of traffic flow, and can provide efficient decision support for the traffic management system. The identified traffic flow patterns can be applied in real time to key application scenarios such as traffic signal optimization, lane scheduling and flow prediction, helping to achieve dynamic adjustment and optimization of traffic flow. Compared with traditional methods, it can process more real-time dynamic data, thereby improving the accuracy and response speed of traffic flow prediction, reducing congestion and improving traffic efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0024] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0025] A method for identifying prevailing traffic flow in a terminal area based on trajectory spectrum clustering, the method comprising:

[0026] S1. Data collection: Based on traffic monitoring equipment, GPS system and vehicle-mounted sensors, the trajectory data of vehicles in the terminal area is obtained, and the trajectory data includes but is not limited to the spatial position, speed, time and acceleration of the vehicle;

[0027] S2, data preprocessing: performing denoising, spatiotemporal alignment and interpolation processing on the trajectory data to improve the consistency and accuracy of the trajectory data;

[0028] S3, trajectory similarity calculation: using the trajectory similarity measurement method, calculate the trajectory similarity of each vehicle in the terminal area and generate a trajectory similarity matrix;

[0029] S4, trajectory clustering: based on the trajectory similarity matrix, a spectral clustering algorithm is used to cluster the trajectory similarity matrix to obtain different traffic flow types;

[0030] S5, traffic flow pattern recognition: based on the trajectory clustering results, identify and analyze the prevailing traffic flow pattern in the terminal area, and identify the characteristics of different traffic flows, including but not limited to high-density flow, low-density flow, uniform flow and congested flow;

[0031] S6. Traffic flow optimization application: Apply the identified traffic flow patterns to the traffic management system.

[0032] First, the trajectory data of vehicles in the terminal area are collected through traffic monitoring equipment, GPS system and vehicle-mounted sensors, including the spatial position, speed, time and acceleration of the vehicle. The collected trajectory data are denoised, time-space aligned and interpolated to improve the consistency and accuracy of the data. Then, the trajectory similarity measurement method is used to calculate the trajectory similarity of each vehicle and generate a similarity matrix. Based on the similarity matrix, the spectral clustering algorithm is used to cluster the trajectories to obtain different types of traffic flow patterns. Next, the clustering results are analyzed to identify the prevailing traffic flow patterns in the terminal area and extract flow characteristics, such as high-density flow, low-density flow, uniform flow and congested flow. Finally, the identified traffic flow patterns are applied to the traffic management system to achieve optimal scheduling and management of traffic flow.

[0033] The data preprocessing includes but is not limited to removing outliers, filling in missing data, unifying time intervals and vehicle type labeling, and the vehicle type labeling performs different traffic flow predictions and traffic flow optimization based on different types of vehicles.

[0034] Data preprocessing helps improve the quality and consistency of data by removing outliers, filling missing data, unifying time intervals and labeling vehicle types, making subsequent analysis and model predictions more accurate and reliable. In particular, the labeling of vehicle types enables accurate flow prediction and traffic flow optimization for different types of vehicles based on their characteristics, thereby improving the efficiency of traffic management, reducing congestion, optimizing resource allocation, and promoting the improvement of road use efficiency.

[0035] The spectral clustering algorithm includes calculating a trajectory similarity matrix and constructing a similarity graph, performing Laplace matrix decomposition on the similarity graph to obtain a feature vector, and clustering the feature vector using a clustering algorithm to obtain a traffic flow pattern.

[0036] The spectral clustering algorithm can effectively capture the intrinsic similarities between different trajectories by calculating the trajectory similarity matrix and constructing a similarity graph, combined with Laplace matrix decomposition. By clustering the feature vectors, the traffic flow pattern can be accurately identified and divided. This method can extract the potential traffic flow structure from complex traffic data, improve the accuracy and robustness of traffic flow pattern recognition, provide a more detailed analysis basis for traffic management and optimization, and thus achieve accurate traffic flow prediction and control.

[0037] The traffic flow pattern recognition includes analyzing the clustering results to identify the time variation characteristics, spatial distribution characteristics and flow type characteristics of the traffic flow.

[0038] Traffic flow pattern recognition can deeply explore the temporal variation characteristics, spatial distribution characteristics and flow type characteristics of traffic flow through the analysis of clustering results. This helps to fully understand the dynamic changes of the transportation system, accurately identify the peak periods, bottleneck areas and flow change patterns of traffic flow, and provide a scientific basis for traffic management departments to optimize signal control, improve road design and improve traffic efficiency, thereby effectively alleviating traffic congestion and improving travel experience.

[0039] The trajectory data is collected from vehicles passing through the terminal area in different time periods, and the trajectory data of each vehicle is processed into a time series format. The trajectory data is collected once per second.

[0040] Processing trajectory data into a time series format and collecting it at a high frequency of once per second can ensure that the detailed dynamic changes of vehicles in different time periods are captured. This high-frequency and time-series processing method improves the timeliness and accuracy of data, helps to more accurately analyze the instantaneous changes in traffic flow, identify traffic patterns, and predict traffic trends, thereby providing more powerful data support for real-time traffic management, congestion warning, and scheduling optimization.

[0041] The distance measurement method used in the trajectory similarity calculation is Euclidean distance, dynamic time warping distance or cosine similarity.

[0042] It can effectively measure the similarity between different trajectories. The Euclidean distance is suitable for simple spatial distance comparison, the dynamic time warping distance can handle possible time shifts or speed changes in time series, and the cosine similarity is suitable for measuring the similarity of trajectory directions and trends. By combining these different measurement methods, it can more comprehensively capture the similarities between trajectories, improve the accuracy of traffic flow pattern recognition, and provide more accurate support for traffic optimization decisions.

[0043] The traffic management system includes but is not limited to traffic signal optimization, lane scheduling and flow prediction. The trajectory clustering uses spectral embedding technology to reduce the dimension of trajectory data to improve clustering efficiency and accuracy.

[0044] Trajectory clustering uses spectral embedding technology for dimensionality reduction, which can effectively reduce the dimension of data, improve clustering efficiency and accuracy, retain the key features in trajectory data, reduce computational complexity, and avoid the "curse of dimensionality" in high-dimensional data, making clustering results more accurate and stable. In the traffic management system, combined with functions such as traffic signal optimization, lane scheduling and flow prediction, more efficient traffic flow analysis and optimization can be achieved, thereby improving the allocation efficiency of traffic resources and road capacity, reducing congestion and improving travel efficiency.

[0045] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for identifying prevailing traffic flow in terminal area based on trajectory spectrum clustering, characterized in that: The method includes: S1. Data collection: Based on traffic monitoring equipment, GPS system and vehicle-mounted sensors, obtain the vehicle trajectory data in the terminal area, the trajectory data including but not limited to the spatial position, speed, time and acceleration of the vehicle; S2, data preprocessing: performing denoising, spatiotemporal alignment and interpolation processing on the trajectory data to improve the consistency and accuracy of the trajectory data; S3, trajectory similarity calculation: using the trajectory similarity measurement method, calculate the trajectory similarity of each vehicle in the terminal area and generate a trajectory similarity matrix; S4, trajectory clustering: based on the trajectory similarity matrix, a spectral clustering algorithm is used to cluster the trajectory similarity matrix to obtain different traffic flow types; S5, traffic flow pattern recognition: based on the trajectory clustering results, identify and analyze the prevailing traffic flow pattern in the terminal area, and identify the characteristics of different traffic flows, including but not limited to high-density flow, low-density flow, uniform flow and congested flow; S6. Traffic flow optimization application: Apply the identified traffic flow patterns to the traffic management system.

2. According to claim 1, a method for identifying prevailing traffic flow in a terminal area based on trajectory spectrum clustering is characterized by: The data preprocessing includes but is not limited to removing outliers, filling in missing data, unifying time intervals and vehicle type labeling, and the vehicle type labeling performs different traffic flow predictions and traffic flow optimization based on different types of vehicles.

3. The method for identifying prevailing traffic flow in terminal area based on trajectory spectrum clustering according to claim 1, characterized in that: The spectral clustering algorithm includes calculating a trajectory similarity matrix and constructing a similarity graph, performing Laplace matrix decomposition on the similarity graph to obtain a feature vector, and clustering the feature vector using a clustering algorithm to obtain a traffic flow pattern.

4. The method for identifying prevailing traffic flow in terminal area based on trajectory spectrum clustering according to claim 1, characterized in that: The traffic flow pattern recognition includes analyzing the clustering results to identify the time variation characteristics, spatial distribution characteristics and flow type characteristics of the traffic flow.

5. The method for identifying prevailing traffic flow in terminal area based on trajectory spectrum clustering according to claim 1, characterized in that: The trajectory data is collected from vehicles passing through the terminal area in different time periods, and the trajectory data of each vehicle is processed into a time series format. The trajectory data is collected once per second.

6. The method for identifying prevailing traffic flow in terminal area based on trajectory spectrum clustering according to claim 1, characterized in that: The distance measurement method used in the trajectory similarity calculation is Euclidean distance, dynamic time warping distance or cosine similarity.

7. The method for identifying prevailing traffic flow in terminal area based on trajectory spectrum clustering according to claim 1, characterized in that: The traffic management system includes but is not limited to traffic signal optimization, lane scheduling and flow prediction.

8. The method for identifying prevailing traffic flow in terminal area based on trajectory spectrum clustering according to claim 1, characterized in that: The trajectory clustering uses spectral embedding technology to perform dimensionality reduction processing on trajectory data to improve clustering efficiency and accuracy.

Images