Urban traffic situation assessment system and method fusing multi-source heterogeneous data

Abstract

The invention belongs to the technical field of intelligent traffic, and discloses an urban traffic situation assessment system and method fusing multi-source heterogeneous data. Comprising the steps of collecting multi-source traffic data in real time; performing anomaly detection on the multi-source traffic data by adopting an anomaly detection algorithm; constructing a space-time coordinate system, mapping the multi-source traffic data into the space-time coordinate system, and obtaining space-time traffic data; carrying out data fusion on the space-time traffic data by adopting a heterogeneous data fusion algorithm, and extracting traffic situation data; constructing a situation assessment model, and analyzing the traffic situation data by using the situation assessment model to obtain a situation index value; based on a multi-index decision theory, analyzing the situation index values, and evaluating the urban traffic situation; according to the method, the correlation among the data sources can be fully mined, the accuracy of data fusion is improved, and comprehensive and accurate evaluation of the urban traffic situation is realized, so that the traffic management efficiency is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and more specifically, to an urban traffic situation assessment system and method that integrates multi-source heterogeneous data. Background Art

[0002] In the process of rapid urbanization, urban traffic is facing increasingly severe challenges, including traffic congestion, frequent traffic accidents and environmental pollution. Traditional traffic management methods often rely on a single data source, which makes it difficult to fully reflect traffic conditions, resulting in insufficient decision support. With the rapid development of information technology, various types of traffic data (such as road monitoring videos, GPS positioning data, traffic sensor data, social media information, etc.) continue to emerge. These data are usually multi-source and heterogeneous. How to effectively integrate and analyze these data has become the key to improving the level of urban traffic management. Therefore, there is an urgent need for an intelligent system that can effectively integrate and analyze multiple data sources to provide a comprehensive and accurate assessment of traffic situation.

[0003] The patent with publication number CN119445846A discloses a traffic situation warning method, system and equipment based on V2X; including: obtaining traffic data collected by multiple data sources, and preprocessing the received traffic data; determining corresponding traffic indicators based on the preprocessed traffic data, wherein the traffic indicators at least include vehicle and pedestrian flow, vehicle average speed and lane traffic situation; wherein the lane traffic situation at least includes a corresponding level representing the road congestion situation within a preset time period, and the corresponding level includes at least one of unblocked, light congestion, moderate congestion and severe congestion; evaluating the corresponding traffic situation status according to the traffic indicators, and outputting corresponding risk warning information if the traffic situation status is abnormal; this invention improves the efficiency of urban traffic management, alleviates traffic congestion problems, and ensures the comprehensiveness and accuracy of information through real-time data monitoring technology.

[0004] However, although the above-mentioned technology has achieved the evaluation of urban traffic situation by fusing multi-source heterogeneous data, the data from different sources cannot be effectively aligned, and there is a lack of targeted fusion algorithms, which fails to fully explore the correlation between various data sources, affecting the accuracy of data fusion; that is, the above-mentioned technology is insufficient in the depth and accuracy of multi-source heterogeneous data fusion, which limits the comprehensive evaluation of urban traffic situation.

[0005] In view of this, the present invention proposes an urban traffic situation assessment system and method that integrates multi-source heterogeneous data to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for evaluating urban traffic situation by integrating multi-source heterogeneous data, comprising:

[0007] Real-time collection of multi-source traffic data, including urban traffic data collected from different data sources;

[0008] Anomaly detection algorithm is used to detect anomalies in multi-source traffic data and delete the anomaly data from the multi-source traffic data;

[0009] Construct a spatiotemporal coordinate system, map multi-source traffic data into the spatiotemporal coordinate system, and obtain spatiotemporal traffic data;

[0010] Adopt heterogeneous data fusion algorithm to fuse spatiotemporal traffic data and extract traffic situation data;

[0011] Construct a situation assessment model, use the situation assessment model to analyze the traffic situation data, obtain the quantitative value of each situation indicator, and mark it as the situation indicator value; based on the multi-index decision-making theory, analyze the situation indicator value and evaluate the urban traffic situation.

[0012] Furthermore, the method for detecting anomalies in multi-source traffic data includes:

[0013] Acquire historical traffic data, where the historical traffic data is multi-source traffic data collected at historical moments and within a normal range; mark the multi-source traffic data collected in real time as real-time traffic data, use the real-time traffic data and the historical traffic data as analytical traffic data, and mark each data in the real-time traffic data as real-time data; and construct corresponding a isolation sets for each data in the multi-source traffic data according to the analytical traffic data, where the a isolation sets are all different, and a is an integer greater than 1;

[0014] According to the isolation set, a group of analysis values ​​corresponding to each real-time data is calculated, and the analysis values ​​include actual values ​​and expected values; the actual value is calculated by counting the number of times each real-time data is isolated in the corresponding isolation set, and marking it as the number of isolations, and taking the number of isolations as the actual value of the corresponding real-time data; the expected value is calculated by counting the number of each type of data in the traffic data, and marking it as the number of data; the correlation coefficient is preset, and the expected value corresponding to each real-time data is calculated according to the number of data and the correlation coefficient; the expression of the expected value is: c = 2 (log 2 (d-1)+θ); where c is the expected value, d is the number of data, and θ is the correlation coefficient;

[0015] According to the analysis value, calculate a anomaly coefficient corresponding to each real-time data; the expression of the anomaly coefficient is: Where e is the anomaly coefficient and f is the actual value. The anomaly coefficients corresponding to each real-time data are added in sequence and then divided by a to obtain the anomaly score corresponding to each real-time data. A scoring threshold is preset and the anomaly score of each real-time data is compared with the scoring threshold. If the anomaly score is less than the scoring threshold, the corresponding real-time data is marked as anomaly data. If the anomaly score is greater than or equal to the scoring threshold, the corresponding real-time data is not marked.

[0016] Furthermore, the step of constructing a corresponding isolation set for each data in the multi-source traffic data includes:

[0017] Step S101: acquiring a data range value according to the analyzed traffic data; the data range value includes a range value corresponding to each type of data in the multi-source traffic data, wherein the maximum value of the range value corresponding to each type of data is the maximum value of the corresponding data in the analyzed traffic data, and the minimum value of the range value corresponding to each type of data is the minimum value of the corresponding data in the analyzed traffic data;

[0018] Step S102: randomly selecting a piece of data not marked as selected data from the multi-source traffic data, and marking it as selected data;

[0019] Step S103: randomly selecting a value in the range value corresponding to the selected data as an isolation point, isolating the selected data according to the isolation point, and obtaining two isolation subsets; wherein, if there is a value in the corresponding value of the selected data that is the same as the isolation point, the corresponding value is marked as the same value, and the same value is isolated into an isolation subset greater than the isolation point;

[0020] Step S104: Obtain the subset range of each isolated subset, the maximum value of the subset range is the maximum value of the corresponding isolated subset, and the minimum value of the subset range is the minimum value of the corresponding isolated subset; randomly select a value as an isolation point in the subset range corresponding to each isolated subset, and obtain b isolation points, where b is the number of isolated subsets containing at least two values; isolate the corresponding isolated subsets according to the b isolation points, and obtain 2b isolated subsets;

[0021] Step S105: looping step S104 until there is only one value in each isolated subset, the loop ends, the isolated subset corresponding to the selected data is taken as the isolated set, and the process proceeds to step S106;

[0022] Step S106: looping steps S102 to S105 until all data in the multi-source traffic data are marked as selected data, the loop ends, and the isolation set corresponding to each data in the multi-source traffic data is obtained.

[0023] Furthermore, the method for constructing a space-time coordinate system includes:

[0024] Set different digital labels for different roads in the city and mark them as road labels; divide the city into grids according to different road labels to obtain g spatial grids, where g is an integer greater than 1, and the spatial grids correspond to the road labels one by one; set the time unit, divide the acquisition interval into grids according to the time unit, and obtain h time grids, where h is an integer greater than 1; set different digital labels for the h time grids in increasing order, and mark the time labels, and the range of the time labels is [1, h]; construct a spatiotemporal coordinate system according to the road labels corresponding to the g spatial grids and the time labels corresponding to the h time grids; wherein the horizontal coordinate of the spatiotemporal coordinate in the spatiotemporal coordinate system is the road label, and the vertical coordinate is the time label;

[0025] The method for obtaining spatiotemporal traffic data comprises:

[0026] According to the time of the road corresponding to each data in the multi-source traffic data, the corresponding road label and time label are obtained; the multi-source traffic data are mapped to the space-time coordinate system according to the corresponding road label and time label, and the space-time coordinates corresponding to each data in the multi-source traffic data are obtained; the data with the same space-time coordinates in the multi-source traffic data are regarded as a group of multi-source space-time data, the multi-source space-time data correspond to the space-time coordinates one by one, and all multi-source space-time data are merged into space-time traffic data.

[0027] Furthermore, the method for extracting traffic situation data includes:

[0028] Each data in the spatiotemporal traffic data is taken as a tensor x(i), x(i)∈R I×J( i )×T ; Where R is a real number, I is the spatial dimension, J(i) is the data feature dimension of the i-th tensor, T is the time dimension, k is the number of data in the spatiotemporal traffic data, i∈[1,k]; the spatial dimension, data feature dimension and time dimension are all dimensions corresponding to each tensor, where the spatial dimension is the road label and the time dimension is the time label; a corresponding rank is set for each dimension corresponding to each tensor and marked as the dimension rank; according to the dimension rank, each tensor is low-rank decomposed to obtain the core tensor and factor matrix corresponding to each tensor, and the factor matrix includes the spatial factor matrix, the data feature factor matrix and the time factor matrix;

[0029] According to the core tensor and factor matrix corresponding to each tensor, a coupling decomposition objective function is constructed; the factor matrix corresponding to each tensor is optimized using the alternating least squares method to minimize the coupling decomposition objective function; the core tensor and factor matrix corresponding to each tensor are obtained when the optimization is completed, and are marked as optimized tensors and optimized matrices respectively; the optimized tensors and optimized matrices corresponding to each data in each group of multi-source spatiotemporal data are fused to obtain the traffic situation data corresponding to each group of multi-source spatiotemporal data.

[0030] Furthermore, the expression of low-rank decomposition is: x(i)≈D(i)×A + ×B 2 (i)×C 3 ;

[0031] Where D(i) is the core tensor corresponding to the ith tensor, A is the spatial factor matrix, B(i) is the data feature factor matrix of the ith tensor, C is the time factor matrix, and D(i)∈R r1×r2×r3 , A∈R r1×I , B(i)∈R r2×J(i) , C∈R r3×T , r1 is the dimension rank corresponding to the spatial dimension, r2 is the dimension rank corresponding to the data feature dimension, and r3 is the dimension rank corresponding to the time dimension; A + represents the multiplication of the spatial factor matrix with the first dimension in the core tensor, B 2 Represents the multiplication of the data feature factor matrix with the second dimension in the core tensor, C 3 Represents the multiplication of the time factor matrix with the third dimension in the core tensor.

[0032] Furthermore, the expression of the coupling decomposition objective function is:

[0033]

[0034] Where F is the coupling decomposition objective function, |||| F is the Frobenius norm, λ is the regularization parameter, λ≥0;

[0035] The expression of traffic situation data is:

[0036] In the formula, Y is the traffic situation data, D(p) is the core tensor corresponding to the p-th data of multi-source spatiotemporal data, is the road label corresponding to the multi-source spatiotemporal data, B(p) is the data feature factor matrix corresponding to the p-th data of the multi-source spatiotemporal data, is the time label corresponding to the multi-source spatiotemporal data, q is the number of data in the multi-source spatiotemporal data, p∈[1,q].

[0037] Furthermore, the situation assessment model includes m indicator assessment models, where m is the number of situation indicators; wherein each indicator assessment model is a deep neural network model, the construction process of each indicator assessment model is consistent, and the indicator assessment model corresponds to the situation indicator one by one;

[0038] Methods for building a situation assessment model include:

[0039] Collect u groups of traffic situation data in advance, set corresponding situation index values ​​for all the u groups of traffic situation data, where u is an integer greater than 1, and convert the traffic situation data and the corresponding situation index values ​​into a corresponding set of feature vectors; use each set of feature vectors as input to a situation assessment model, the situation assessment model uses a set of predicted situation index values ​​corresponding to each group of traffic situation data as output, and uses the actual situation index value corresponding to each group of traffic situation data as a prediction target, where the actual situation index value is the pre-set situation index value corresponding to the traffic situation data; use minimizing the sum of prediction errors of all traffic situation data as a training target; train the situation assessment model until the sum of prediction errors reaches convergence and stops training;

[0040] The extracted traffic situation data are input into each indicator evaluation model respectively to obtain the situation indicator value corresponding to each set of traffic situation data, that is, to obtain the situation indicator value corresponding to each spatiotemporal coordinate in the city.

[0041] Furthermore, the method for evaluating urban traffic situation includes:

[0042] Obtain historical indicator values, which are situation indicator values ​​obtained at historical moments; normalize each historical indicator value to obtain a standardized indicator value; treat standardized indicator values ​​of the same type as an indicator set, and the indicator set corresponds to the situation indicator one by one; count the number of standardized indicator values ​​with different values ​​in each indicator set, and mark them as the number of values; count the number of all standardized indicator values ​​in each indicator set, and mark them as the total number of values; divide the number of values ​​corresponding to each indicator set by the corresponding total number of values, and obtain the probability of occurrence of each standardized indicator value with different values ​​in each indicator set; calculate the randomness of each indicator set based on the probability of occurrence; add the randomness of each indicator set in turn to obtain the total randomness; divide the randomness of each indicator set by the total randomness to obtain the random probability of each indicator set; use the inverse of the random probability corresponding to each indicator set as the weight coefficient of the corresponding situation indicator;

[0043] A related set is preset, and the related set includes the related direction between each situation indicator and the total traffic situation indicator. The total traffic situation indicator is the quantitative value of the urban traffic situation at different time and space coordinates in the city, and the related direction includes positive correlation and negative correlation. According to the related set, the related direction between each situation indicator and the total traffic situation indicator is obtained; each situation indicator value is normalized to obtain a situation reference value; each situation reference value is multiplied by a corresponding weight coefficient to obtain a situation weight value; the situation indicator that is positively correlated with the total traffic situation indicator is marked as a positive indicator, and the situation indicator that is negatively correlated with the total traffic situation indicator is marked as a negative indicator; the situation weight values ​​corresponding to all positive indicators are added in sequence to obtain a positive situation value, and the situation weight values ​​corresponding to all negative indicators are added in sequence to obtain a negative situation value; the situation positive value is subtracted from the situation negative value to obtain the total traffic situation indicator;

[0044] The expression of randomness is:

[0045] In the formula, H is the correlation, P v is the probability of occurrence of the vth standardized indicator in the indicator set, n is the number of standardized indicators with different values ​​in the indicator set, v∈[1,n].

[0046] A system for evaluating urban traffic situation by integrating multi-source heterogeneous data, implementing a method for evaluating urban traffic situation by integrating multi-source heterogeneous data, comprising:

[0047] A data collection module is used to collect multi-source traffic data in real time, where the multi-source traffic data includes urban traffic data collected from different data sources;

[0048] A data detection module, used to use an anomaly detection algorithm to perform anomaly detection on multi-source traffic data and delete the anomaly data from the multi-source traffic data;

[0049] The data mapping module is used to construct a spatiotemporal coordinate system, map multi-source traffic data into the spatiotemporal coordinate system, and obtain spatiotemporal traffic data;

[0050] The data fusion module is used to fuse the spatiotemporal traffic data using a heterogeneous data fusion algorithm and extract traffic situation data;

[0051] The situation assessment module is used to build a situation assessment model, use the situation assessment model to analyze the traffic situation data, obtain the quantitative value of each situation indicator, and mark it as the situation indicator value; based on the multi-indicator decision-making theory, analyze the situation indicator value and evaluate the urban traffic situation.

[0052] The technical effects and advantages of the urban traffic situation assessment system and method integrating multi-source heterogeneous data of the present invention are as follows:

[0053] By collecting multi-source traffic data in real time and using anomaly detection algorithms to pre-process multi-source traffic data, abnormal data can be effectively eliminated and data quality can be improved; a spatiotemporal coordinate system is constructed for data alignment to achieve spatiotemporal alignment and unification of different data sources, laying the foundation for subsequent data fusion; a heterogeneous data fusion algorithm is used to effectively fuse spatiotemporal traffic data, extract core features rich in traffic situation information, and provide reliable data support for subsequent traffic situation assessment; a situation assessment model based on a deep neural network is constructed, which can accurately quantify various traffic situation indicators, and based on multi-indicator decision-making theory, a comprehensive and accurate assessment of urban traffic situation can be achieved; the correlation between various data sources can be fully explored and the accuracy of data fusion can be improved, thereby effectively improving the efficiency of traffic management and alleviating traffic congestion problems. It has important application value in alleviating urban traffic congestion, reducing the risk of traffic accidents, and optimizing resource allocation, and is conducive to improving the overall traffic management level of the city and the public travel experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a schematic diagram of an urban traffic situation assessment system integrating multi-source heterogeneous data according to Embodiment 1 of the present invention;

[0055] Figure 2 This is a flow chart of a method for evaluating urban traffic situation by integrating multi-source heterogeneous data according to Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] Example 1

[0058] See also Figure 1 As shown, the urban traffic situation assessment system that integrates multi-source heterogeneous data described in this embodiment includes a data acquisition module, a data detection module, a data mapping module, a data fusion module and a situation assessment module; each module is connected by wired and / or wireless means to achieve data transmission between modules.

[0059] The data collection module is used to collect multi-source traffic data in real time. The multi-source traffic data includes urban traffic data collected from different data sources.

[0060] Urban traffic data refers to data related to urban traffic, which is used to describe the status and dynamic changes of various aspects such as vehicles, traffic mobility, and traffic safety in the urban traffic network; urban traffic data includes traffic flow data, traffic speed data, traffic accident data, etc.; among them, traffic flow data refers to the number of vehicles passing through different roads per unit time in the collection interval, which is obtained through multiple data sources such as traffic monitoring cameras and road monitoring radars, and is used to measure road utilization and congestion; multi-source traffic data is collected according to a preset collection interval, and the collection interval is pre-set by technical personnel in this field according to actual conditions; traffic speed data refers to the average driving speed of vehicles on different roads per unit time in the collection interval, which is obtained through multiple data sources such as traffic monitoring cameras and speed radars, and is used to measure the road's capacity and traffic smoothness; traffic accident data includes the time and road where the traffic accident occurred, which is obtained through multiple data sources such as traffic monitoring cameras and traffic accident reports issued by traffic management departments, and is used to analyze road safety and accident-prone areas.

[0061] The data detection module is used to use an anomaly detection algorithm to perform anomaly detection on multi-source traffic data and delete the anomaly data from the multi-source traffic data.

[0062] Methods for anomaly detection on multi-source traffic data include:

[0063] Acquire historical traffic data, which is multi-source traffic data within a normal range collected at historical moments, and is obtained through a built-in database (such as a relational database, a NoSQL database, a time series database, etc.) of an urban traffic situation assessment system; mark the multi-source traffic data collected in real time as real-time traffic data, use the real-time traffic data and historical traffic data as analytical traffic data, and mark each data in the real-time traffic data as real-time data; construct corresponding a isolation sets for each data in the multi-source traffic data according to the analytical traffic data, and the a isolation sets are all different, and a is an integer greater than 1;

[0064] According to the isolated set, a group of analysis values ​​corresponding to each real-time data is calculated, and the analysis values ​​include actual values ​​and expected values; the actual value is calculated by counting the number of times each real-time data is isolated in the corresponding isolated set, and marking it as the number of isolations, and taking the number of isolations as the actual value of the corresponding real-time data; the expected value is calculated by counting and analyzing the number of each type of data in the traffic data, and marking it as the number of data; the correlation coefficient is preset, and the correlation coefficient is preset by technical personnel in this field according to the actual situation; the expected value corresponding to each real-time data is calculated according to the number of data and the correlation coefficient; the expression of the expected value is: c=2(log 2 (d-1)+θ); where c is the expected value, d is the number of data, θ is the correlation coefficient, and θ>0;

[0065] According to the analysis value, calculate a anomaly coefficient corresponding to each real-time data; the expression of the anomaly coefficient is: Wherein, e is the anomaly coefficient, and f is the actual value; the anomaly coefficients corresponding to each real-time data are added in sequence, and then divided by a to obtain the anomaly score corresponding to each real-time data; a scoring threshold is preset, and the scoring threshold is preset by technical personnel in this field according to actual conditions; the anomaly score of each real-time data is compared with the scoring threshold respectively; if the anomaly score is less than the scoring threshold, the corresponding real-time data is marked as anomaly data, indicating that the corresponding real-time data is abnormal and cannot be used as a basis for subsequent evaluation of the urban traffic situation; if the anomaly score is greater than or equal to the scoring threshold, the corresponding real-time data is not marked.

[0066] The steps of constructing a corresponding isolation set for each data in the multi-source traffic data include:

[0067] Step S101: acquiring a data range value according to the analyzed traffic data; the data range value includes a range value corresponding to each type of data in the multi-source traffic data, wherein the maximum value of the range value corresponding to each type of data is the maximum value of the corresponding data in the analyzed traffic data, and the minimum value of the range value corresponding to each type of data is the minimum value of the corresponding data in the analyzed traffic data;

[0068] Step S102: randomly selecting a piece of data not marked as selected data from the multi-source traffic data, and marking it as selected data;

[0069] Step S103: randomly selecting a value in the range value corresponding to the selected data as an isolation point, isolating the selected data according to the isolation point, and obtaining two isolation subsets; wherein, if there is a value in the corresponding value of the selected data that is the same as the isolation point, the corresponding value is marked as the same value, and the same value is isolated into an isolation subset greater than the isolation point;

[0070] Step S104: Obtain the subset range of each isolated subset, the maximum value of the subset range is the maximum value of the corresponding isolated subset, and the minimum value of the subset range is the minimum value of the corresponding isolated subset; randomly select a value as an isolation point in the subset range corresponding to each isolated subset, and obtain b isolation points, where b is the number of isolated subsets containing at least two values; isolate the corresponding isolated subsets according to the b isolation points, and obtain 2b isolated subsets;

[0071] Step S105: looping step S104 until there is only one value in each isolated subset, the loop ends, the isolated subset corresponding to the selected data is taken as the isolated set, and the process proceeds to step S106;

[0072] Step S106: looping steps S102 to S105 until all data in the multi-source traffic data are marked as selected data, the loop ends, and the isolation set corresponding to each data in the multi-source traffic data is obtained.

[0073] Exemplarily, the data includes four values ​​3, 5, 7, and 12, and the isolation point is 5. At this time, one isolation subset is 3, and the other isolation subset is 5, 7, and 12; the isolation point is 8, and at this time, one isolation subset is 3, one isolation subset is 5 and 7, and one isolation subset is 12; the isolation point is 6, and at this time, one isolation subset is 3, one isolation subset is 5, one isolation subset is 7, and one isolation subset is 12; since there is only one value in each isolation subset, the isolation set is constructed; among them, the number of isolations corresponding to 3 is 1, the number of isolations corresponding to 5 and 7 is 3, and the number of isolations corresponding to 12 is 2.

[0074] The data mapping module is used to construct a spatiotemporal coordinate system, map multi-source traffic data into the spatiotemporal coordinate system, and obtain spatiotemporal traffic data.

[0075] Methods for constructing a space-time coordinate system include:

[0076] Different digital labels are set for different roads in the city and marked as road labels; the city is divided into grids according to different road labels to obtain g spatial grids, where g is an integer greater than 1, and the spatial grids correspond to the road labels one by one; the time unit is set, and the time unit is pre-set by technical personnel in this field according to actual conditions, such as minutes, hours, days, etc.; according to the time unit, the collection interval is divided into grids to obtain h time grids, where h is an integer greater than 1; different digital labels are set for the h time grids in increasing order, and the time labels are marked, and the range of the time labels is [1, h]; a space-time coordinate system is constructed according to the road labels corresponding to the g spatial grids and the time labels corresponding to the h time grids; wherein the horizontal coordinate of the space-time coordinate in the space-time coordinate system is the road label, and the vertical coordinate is the time label.

[0077] Methods for obtaining spatiotemporal traffic data include:

[0078] According to the time of the road corresponding to each data in the multi-source traffic data, the corresponding road label and time label are obtained; the multi-source traffic data are mapped to the space-time coordinate system according to the corresponding road label and time label, and the space-time coordinates corresponding to each data in the multi-source traffic data are obtained; the data with the same space-time coordinates in the multi-source traffic data are regarded as a group of multi-source space-time data, the multi-source space-time data correspond to the space-time coordinates one by one, and all multi-source space-time data are merged into space-time traffic data.

[0079] The data fusion module is used to adopt heterogeneous data fusion algorithm to fuse spatiotemporal traffic data and extract traffic situation data.

[0080] Methods for extracting traffic situation data include:

[0081] Each data in the spatiotemporal traffic data is taken as a tensor x(i), x(i)∈R I×J( i )×T ; Wherein, R is a real number, I is the spatial dimension, J(i) is the data feature dimension of the i-th tensor, T is the time dimension, k is the number of data in the spatiotemporal traffic data, i∈[1,k]; the spatial dimension, data feature dimension and time dimension are all dimensions corresponding to each tensor, where the spatial dimension is the road label, the time dimension is the time label, and the data feature dimension, for example, the data feature dimension corresponding to the traffic flow data is the number of vehicles, and the data feature dimension corresponding to the traffic speed data is the average driving speed, etc.; a corresponding rank is set for each dimension corresponding to each tensor and marked as the dimension rank; according to the dimension rank, each tensor is low-rank decomposed to obtain the core tensor and factor matrix corresponding to each tensor, and the factor matrix includes the spatial factor matrix, the data feature factor matrix and the time factor matrix.

[0082] The expression of low-rank decomposition is: x(i)≈D(i)×A + ×B 2 (i)×C 3 ;

[0083] Where D(i) is the core tensor corresponding to the ith tensor, A is the spatial factor matrix, B(i) is the data feature factor matrix of the ith tensor, C is the time factor matrix, and D(i)∈R r1×r2×r3 , A∈R r1×I , B(i)∈R r2×J(i) , C∈R r3×T , r1 is the dimension rank corresponding to the spatial dimension, r2 is the dimension rank corresponding to the data feature dimension, and r3 is the dimension rank corresponding to the time dimension; A + represents the multiplication of the spatial factor matrix with the first dimension (i.e., the spatial dimension) in the core tensor, B 2 Indicates that the data feature factor matrix is ​​multiplied by the second dimension (i.e., data feature dimension) in the core tensor, C 3 It represents the multiplication of the time factor matrix and the third dimension (i.e., the time dimension) in the core tensor. It should be noted that the reason why the approximate equality sign is used in the expression of low-rank decomposition is that there is an error between the tensor and the corresponding low-rank decomposition result, so only the approximate equality sign can be used, and the equality sign cannot be used.

[0084] According to the core tensor and factor matrix corresponding to each tensor, a coupling decomposition objective function is constructed; the factor matrix corresponding to each tensor is optimized by the alternating least squares method to minimize the coupling decomposition objective function. The alternating least squares method is an existing technology and will not be described in detail here; the core tensor and factor matrix corresponding to each tensor are obtained when the optimization is completed, and are marked as optimized tensors and optimized matrices respectively; the optimized tensors and optimized matrices corresponding to each data in each group of multi-source spatiotemporal data are fused to obtain the traffic situation data corresponding to each group of multi-source spatiotemporal data.

[0085] The expression of the coupling decomposition objective function is:

[0086]

[0087] Where F is the coupling decomposition objective function, |||| F is the Frobenius norm, which is used to measure the error, λ is the regularization parameter, λ≥0.

[0088] The expression of traffic situation data is:

[0089] In the formula, Y is the traffic situation data, D(p) is the core tensor corresponding to the p-th data of multi-source spatiotemporal data, is the road label corresponding to the multi-source spatiotemporal data, B(p) is the data feature factor matrix corresponding to the p-th data of the multi-source spatiotemporal data, is the time label corresponding to the multi-source spatiotemporal data, q is the number of data in the multi-source spatiotemporal data, p∈[1,q].

[0090] The situation assessment module is used to build a situation assessment model, use the situation assessment model to analyze the traffic situation data, obtain the quantitative value of each situation indicator, and mark it as the situation indicator value; based on the multi-indicator decision-making theory, analyze the situation indicator value and evaluate the urban traffic situation.

[0091] The situation assessment model includes m indicator assessment models, where m is the number of situation indicators, such as congestion level, travel efficiency, travel safety, etc.; among them, each indicator assessment model is a deep neural network model, and the construction process of each indicator assessment model is consistent. The indicator assessment model corresponds to the situation indicator one by one, that is, one indicator assessment model outputs one situation indicator value.

[0092] Methods for building a situation assessment model include:

[0093] Collect u groups of traffic situation data in advance, set corresponding situation index values ​​for the u groups of traffic situation data, where u is an integer greater than 1, and convert the traffic situation data and the corresponding situation index values ​​into a corresponding set of feature vectors; the situation index values ​​corresponding to the traffic situation data are collected by a person skilled in the art in the process of historically quantifying the situation index values, and each group of traffic situation data is analyzed in combination with actual experience, and the corresponding situation index values ​​are set, and the corresponding situation index values ​​are set for the u groups of traffic situation data in sequence;

[0094] Each group of feature vectors is used as the input of the situation assessment model. The situation assessment model uses a group of predicted situation index values ​​corresponding to each group of traffic situation data as output, and uses the actual situation index value corresponding to each group of traffic situation data as the prediction target. The actual situation index value is the pre-set situation index value corresponding to the traffic situation data; the training target is to minimize the sum of the prediction errors of all traffic situation data; wherein the calculation formula of the prediction error is η w =(φ w -ε w ) 2 , where η w is the prediction error, w is the group number of the feature vector corresponding to the traffic situation data, θ w is the predicted situation index value corresponding to the wth group of traffic situation data, ε w is the actual situation index value corresponding to the wth group of traffic situation data; the situation assessment model is trained until the sum of the prediction errors reaches convergence and the training is stopped.

[0095] The extracted traffic situation data are input into each indicator evaluation model respectively to obtain the situation indicator value corresponding to each set of traffic situation data, that is, to obtain the situation indicator value corresponding to each spatiotemporal coordinate in the city.

[0096] Methods for assessing urban traffic situation include:

[0097] Obtain historical indicator values, which are situation indicator values ​​obtained at historical moments, and are obtained through the built-in database of the urban traffic situation assessment system; perform normalization processing on each historical indicator value (such as maximum-minimum standardization, Z-score standardization, etc.) to obtain standardized indicator values; regard standardized indicator values ​​of the same type as an indicator set, and the indicator set corresponds to the situation indicator one by one; count the number of standardized indicator values ​​with different values ​​in each indicator set, and mark them as the number of values; count the number of all standardized indicator values ​​in each indicator set, and mark them as the total number of values; divide the number of values ​​corresponding to each indicator set by the corresponding total number of values, and obtain the probability of occurrence of each standardized indicator value with different values ​​in each indicator set; calculate the randomness of each indicator set based on the probability of occurrence; add the randomness of each indicator set in turn to obtain the total randomness; divide the randomness of each indicator set by the total randomness to obtain the random probability of each indicator set; use the inverse of the random probability corresponding to each indicator set as the weight coefficient of the corresponding situation indicator;

[0098] A preset correlation set includes the correlation direction between each situation indicator and the total traffic situation indicator. The total traffic situation indicator is a quantitative value of the urban traffic situation at different time and space coordinates in the city, and the correlation direction includes positive correlation and negative correlation. The correlation set is preset by technical personnel in this field based on actual experience and literature review. According to the correlation set, the correlation direction between each situation indicator and the total traffic situation indicator is obtained. Each situation indicator value is normalized to obtain a situation reference value. Each situation reference value is multiplied by a corresponding weight coefficient to obtain a situation weight value. The situation indicator that is positively correlated with the total traffic situation indicator is marked as a positive indicator, such as travel efficiency, travel safety, etc. The situation indicator that is negatively correlated with the total traffic situation indicator is marked as a negative indicator, such as congestion level, etc. The situation weight values ​​corresponding to all positive indicators are added in sequence to obtain a positive situation value, and the situation weight values ​​corresponding to all negative indicators are added in sequence to obtain a negative situation value. The situation positive value is subtracted from the situation negative value to obtain the total traffic situation indicator.

[0099] The expression of randomness is:

[0100] In the formula, H is the correlation, P v is the probability of occurrence of the vth standardized indicator in the indicator set, n is the number of standardized indicators with different values ​​in the indicator set, v∈[1,n].

[0101] This embodiment collects multi-source traffic data in real time and uses an anomaly detection algorithm to pre-process the multi-source traffic data, which can effectively eliminate abnormal data and improve data quality; constructs a spatiotemporal coordinate system for data alignment to achieve spatiotemporal alignment and unification of different data sources, laying a foundation for subsequent data fusion; adopts a heterogeneous data fusion algorithm to effectively fuse spatiotemporal traffic data, extract core features rich in traffic situation information, and provide reliable data support for subsequent traffic situation evaluation; constructs a situation evaluation model based on a deep neural network, which can accurately quantify various traffic situation indicators, and based on the multi-index decision-making theory, realizes a comprehensive and accurate evaluation of the urban traffic situation; can fully explore the correlation between various data sources, improve the accuracy of data fusion, thereby effectively improving the efficiency of traffic management and alleviating traffic congestion problems, which has important application value in alleviating urban traffic congestion, reducing the risk of traffic accidents, optimizing resource allocation, etc., and is conducive to improving the overall traffic management level of the city and the public travel experience.

[0102] Example 2

[0103] See also Figure 2 As shown, the part not described in detail in this embodiment is described in Example 1, and a method for evaluating urban traffic situation by integrating multi-source heterogeneous data is provided, the method comprising:

[0104] Real-time collection of multi-source traffic data, including urban traffic data collected from different data sources;

[0105] Anomaly detection algorithm is used to detect anomalies in multi-source traffic data and delete the anomaly data from the multi-source traffic data;

[0106] Construct a spatiotemporal coordinate system, map multi-source traffic data into the spatiotemporal coordinate system, and obtain spatiotemporal traffic data;

[0107] Adopt heterogeneous data fusion algorithm to fuse spatiotemporal traffic data and extract traffic situation data;

[0108] Construct a situation assessment model, use the situation assessment model to analyze the traffic situation data, obtain the quantitative value of each situation indicator, and mark it as the situation indicator value; based on the multi-index decision-making theory, analyze the situation indicator value and evaluate the urban traffic situation.

[0109] Example 3

[0110] The present application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable codes, and when the computer-readable codes are executed by the one or more processors, the urban traffic situation assessment method for integrating multi-source heterogeneous data as described above may be executed.

[0111] The method or system according to the implementation mode of the present application can also be implemented with the help of the architecture of the electronic device shown in the present application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. A storage device in the electronic device, such as a ROM or a hard disk, can store a method for evaluating urban traffic situation by integrating multi-source heterogeneous data provided in the present application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in the present application is only exemplary. When implementing different devices, one or more components in the electronic device shown in the present application may be omitted according to actual needs.

[0112] Example 4

[0113] One embodiment of the present application discloses a computer-readable storage medium. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by a processor, a method for evaluating urban traffic situation by fusing multi-source heterogeneous data according to an embodiment of the present application described with reference to the above figures can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0114] In addition, according to the implementation of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided by the present application, for example: a method for evaluating urban traffic situation by integrating multi-source heterogeneous data. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0115] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0116] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for evaluating urban traffic situation by integrating multi-source heterogeneous data, characterized in that: include: Real-time collection of multi-source traffic data, including urban traffic data collected from different data sources; Anomaly detection algorithm is used to detect anomalies in multi-source traffic data and delete the anomaly data from the multi-source traffic data; Construct a spatiotemporal coordinate system, map multi-source traffic data into the spatiotemporal coordinate system, and obtain spatiotemporal traffic data; Adopt heterogeneous data fusion algorithm to fuse spatiotemporal traffic data and extract traffic situation data; Construct a situation assessment model, use the situation assessment model to analyze the traffic situation data, obtain the quantitative value of each situation indicator, and mark it as the situation indicator value; based on the multi-index decision-making theory, analyze the situation indicator value and evaluate the urban traffic situation.

2. According to the method for evaluating urban traffic situation by integrating multi-source heterogeneous data of claim 1, it is characterized in that: The method for detecting anomalies in multi-source traffic data comprises: Acquire historical traffic data, where the historical traffic data is multi-source traffic data collected at historical moments and within a normal range; mark the multi-source traffic data collected in real time as real-time traffic data, use the real-time traffic data and the historical traffic data as analytical traffic data, and mark each data in the real-time traffic data as real-time data; and construct corresponding a isolation sets for each data in the multi-source traffic data according to the analytical traffic data, where the a isolation sets are all different, and a is an integer greater than 1; According to the isolation set, a group of analysis values ​​corresponding to each real-time data is calculated, and the analysis values ​​include actual values ​​and expected values; the actual value is calculated by counting the number of times each real-time data is isolated in the corresponding isolation set, and marking it as the number of isolations, and taking the number of isolations as the actual value of the corresponding real-time data; the expected value is calculated by counting and analyzing the number of each type of data in the traffic data, and marking it as the number of data; the correlation coefficient is preset, and the expected value corresponding to each real-time data is calculated according to the number of data and the correlation coefficient; the expression of the expected value is: c=2(log2(d-1)+θ); where c is the expected value, d is the number of data, and θ is the correlation coefficient; According to the analysis value, calculate a anomaly coefficient corresponding to each real-time data; the expression of the anomaly coefficient is: Where e is the anomaly coefficient and f is the actual value. The anomaly coefficients corresponding to each real-time data are added in sequence and then divided by a to obtain the anomaly score corresponding to each real-time data. A scoring threshold is preset and the anomaly score of each real-time data is compared with the scoring threshold. If the anomaly score is less than the scoring threshold, the corresponding real-time data is marked as anomaly data. If the anomaly score is greater than or equal to the scoring threshold, the corresponding real-time data is not marked.

3. The urban traffic situation assessment method integrating multi-source heterogeneous data according to claim 2 is characterized in that: The steps of constructing a corresponding isolation set for each data in the multi-source traffic data include: Step S101: acquiring a data range value according to the analyzed traffic data; the data range value includes a range value corresponding to each type of data in the multi-source traffic data, wherein the maximum value of the range value corresponding to each type of data is the maximum value of the corresponding data in the analyzed traffic data, and the minimum value of the range value corresponding to each type of data is the minimum value of the corresponding data in the analyzed traffic data; Step S102: randomly selecting a data not marked as selected data from the multi-source traffic data, and marking it as selected data; Step S103: randomly selecting a value in the range value corresponding to the selected data as an isolation point, isolating the selected data according to the isolation point, and obtaining two isolation subsets; wherein, if there is a value in the corresponding value of the selected data that is the same as the isolation point, the corresponding value is marked as the same value, and the same value is isolated into an isolation subset greater than the isolation point; Step S104: Obtain the subset range of each isolated subset, the maximum value of the subset range is the maximum value of the corresponding isolated subset, and the minimum value of the subset range is the minimum value of the corresponding isolated subset; randomly select a value as an isolation point in the subset range corresponding to each isolated subset, and obtain b isolation points, where b is the number of isolated subsets containing at least two values; isolate the corresponding isolated subsets according to the b isolation points, and obtain 2b isolated subsets; Step S105: looping step S104 until there is only one value in each isolated subset, the loop ends, the isolated subset corresponding to the selected data is taken as the isolated set, and the process proceeds to step S106; Step S106: looping steps S102 to S105 until all data in the multi-source traffic data are marked as selected data, the loop ends, and the isolation set corresponding to each data in the multi-source traffic data is obtained.

4. The urban traffic situation assessment method integrating multi-source heterogeneous data according to claim 3 is characterized in that: The method for constructing a space-time coordinate system comprises: Set different digital labels for different roads in the city and mark them as road labels; divide the city into grids according to different road labels to obtain g spatial grids, where g is an integer greater than 1, and the spatial grids correspond to the road labels one by one; set the time unit, divide the acquisition interval into grids according to the time unit, and obtain h time grids, where h is an integer greater than 1; set different digital labels for the h time grids in increasing order, and mark the time labels, and the range of the time labels is [1, h]; construct a spatiotemporal coordinate system according to the road labels corresponding to the g spatial grids and the time labels corresponding to the h time grids; wherein the horizontal coordinate of the spatiotemporal coordinate in the spatiotemporal coordinate system is the road label, and the vertical coordinate is the time label; The method for obtaining spatiotemporal traffic data comprises: According to the time of the road corresponding to each data in the multi-source traffic data, the corresponding road label and time label are obtained; the multi-source traffic data are mapped to the space-time coordinate system according to the corresponding road label and time label, and the space-time coordinates corresponding to each data in the multi-source traffic data are obtained; the data with the same space-time coordinates in the multi-source traffic data are regarded as a group of multi-source space-time data, the multi-source space-time data correspond to the space-time coordinates one by one, and all multi-source space-time data are merged into space-time traffic data.

5. The urban traffic situation assessment method integrating multi-source heterogeneous data according to claim 4 is characterized in that: The method for extracting traffic situation data comprises: Each data in the spatiotemporal traffic data is taken as a tensor x(i), x(i)∈R I×J( i )×T ; Where R is a real number, I is the spatial dimension, J(i) is the data feature dimension of the i-th tensor, T is the time dimension, k is the number of data in the spatiotemporal traffic data, i∈[1,k]; the spatial dimension, data feature dimension and time dimension are all dimensions corresponding to each tensor, where the spatial dimension is the road label and the time dimension is the time label; a corresponding rank is set for each dimension corresponding to each tensor and marked as the dimension rank; according to the dimension rank, each tensor is low-rank decomposed to obtain the core tensor and factor matrix corresponding to each tensor, and the factor matrix includes the spatial factor matrix, the data feature factor matrix and the time factor matrix; According to the core tensor and factor matrix corresponding to each tensor, a coupling decomposition objective function is constructed; the factor matrix corresponding to each tensor is optimized using the alternating least squares method to minimize the coupling decomposition objective function; the core tensor and factor matrix corresponding to each tensor are obtained when the optimization is completed, and are marked as optimized tensors and optimized matrices respectively; the optimized tensors and optimized matrices corresponding to each data in each group of multi-source spatiotemporal data are fused to obtain the traffic situation data corresponding to each group of multi-source spatiotemporal data.

6. The urban traffic situation assessment method integrating multi-source heterogeneous data according to claim 5 is characterized in that: The expression of low-rank decomposition is: x(i)≈D(i)×A + ×B2(i)×C3; Where D(i) is the core tensor corresponding to the ith tensor, A is the spatial factor matrix, B(i) is the data feature factor matrix of the ith tensor, C is the time factor matrix, and D(i)∈R r1×r2×r3 , A∈R r1×I , B(i)∈R r2×J(i) , C∈R r3×T , r1 is the dimension rank corresponding to the spatial dimension, r2 is the dimension rank corresponding to the data feature dimension, and r3 is the dimension rank corresponding to the time dimension; A + It represents the multiplication of the spatial factor matrix with the first dimension in the core tensor, B2 represents the multiplication of the data feature factor matrix with the second dimension in the core tensor, and C3 represents the multiplication of the time factor matrix with the third dimension in the core tensor.

7. The urban traffic situation assessment method integrating multi-source heterogeneous data according to claim 6 is characterized in that: The expression of the coupling decomposition objective function is: Where F is the coupling decomposition objective function, |||| F is the Frobenius norm, λ is the regularization parameter, λ≥0; The expression of traffic situation data is: In the formula, Y is the traffic situation data, D(p) is the core tensor corresponding to the p-th data of multi-source spatiotemporal data, is the road label corresponding to the multi-source spatiotemporal data, B(p) is the data feature factor matrix corresponding to the p-th data of the multi-source spatiotemporal data, is the time label corresponding to the multi-source spatiotemporal data, q is the number of data in the multi-source spatiotemporal data, p∈[1,q].

8. The urban traffic situation assessment method integrating multi-source heterogeneous data according to claim 7 is characterized in that: The situation assessment model includes m indicator assessment models, where m is the number of situation indicators; each indicator assessment model is a deep neural network model, the construction process of each indicator assessment model is consistent, and the indicator assessment model corresponds to the situation indicator one by one; Methods for building a situation assessment model include: Collect u groups of traffic situation data in advance, set corresponding situation index values ​​for all the u groups of traffic situation data, where u is an integer greater than 1, and convert the traffic situation data and the corresponding situation index values ​​into a corresponding set of feature vectors; use each set of feature vectors as input to a situation assessment model, the situation assessment model uses a set of predicted situation index values ​​corresponding to each group of traffic situation data as output, and uses the actual situation index value corresponding to each group of traffic situation data as a prediction target, where the actual situation index value is the pre-set situation index value corresponding to the traffic situation data; use minimizing the sum of prediction errors of all traffic situation data as a training target; train the situation assessment model until the sum of prediction errors reaches convergence and stops training; The extracted traffic situation data are input into each indicator evaluation model respectively to obtain the situation indicator value corresponding to each set of traffic situation data, that is, to obtain the situation indicator value corresponding to each spatiotemporal coordinate in the city.

9. The urban traffic situation assessment method integrating multi-source heterogeneous data according to claim 8 is characterized in that: The method for evaluating urban traffic situation comprises: Obtain historical indicator values, which are situation indicator values ​​obtained at historical moments; normalize each historical indicator value to obtain a standardized indicator value; treat standardized indicator values ​​of the same type as an indicator set, and the indicator set corresponds to the situation indicator one by one; count the number of standardized indicator values ​​with different values ​​in each indicator set, and mark them as the number of values; count the number of all standardized indicator values ​​in each indicator set, and mark them as the total number of values; divide the number of values ​​corresponding to each indicator set by the corresponding total number of values, and obtain the probability of occurrence of each standardized indicator value with different values ​​in each indicator set; calculate the randomness of each indicator set based on the probability of occurrence; add the randomness of each indicator set in turn to obtain the total randomness; divide the randomness of each indicator set by the total randomness to obtain the random probability of each indicator set; use the inverse of the random probability corresponding to each indicator set as the weight coefficient of the corresponding situation indicator; A related set is preset, and the related set includes the related direction between each situation indicator and the total traffic situation indicator. The total traffic situation indicator is the quantitative value of the urban traffic situation at different time and space coordinates in the city, and the related direction includes positive correlation and negative correlation. According to the related set, the related direction between each situation indicator and the total traffic situation indicator is obtained; each situation indicator value is normalized to obtain a situation reference value; each situation reference value is multiplied by a corresponding weight coefficient to obtain a situation weight value; the situation indicator that is positively correlated with the total traffic situation indicator is marked as a positive indicator, and the situation indicator that is negatively correlated with the total traffic situation indicator is marked as a negative indicator; the situation weight values ​​corresponding to all positive indicators are added in sequence to obtain a positive situation value, and the situation weight values ​​corresponding to all negative indicators are added in sequence to obtain a negative situation value; the situation positive value is subtracted from the situation negative value to obtain the total traffic situation indicator; The expression of randomness is: In the formula, H is the correlation, P v is the probability of occurrence of the vth standardized indicator in the indicator set, n is the number of standardized indicators with different values ​​in the indicator set, v∈[1,n].

10. An urban traffic situation assessment system integrating multi-source heterogeneous data, implementing an urban traffic situation assessment method integrating multi-source heterogeneous data as described in any one of claims 1 to 9, characterized in that: include: A data collection module is used to collect multi-source traffic data in real time, where the multi-source traffic data includes urban traffic data collected from different data sources; A data detection module, used to use an anomaly detection algorithm to perform anomaly detection on multi-source traffic data and delete the anomaly data from the multi-source traffic data; The data mapping module is used to construct a spatiotemporal coordinate system, map multi-source traffic data into the spatiotemporal coordinate system, and obtain spatiotemporal traffic data; The data fusion module is used to fuse the spatiotemporal traffic data using a heterogeneous data fusion algorithm and extract traffic situation data; The situation assessment module is used to build a situation assessment model, use the situation assessment model to analyze the traffic situation data, obtain the quantitative value of each situation indicator, and mark it as the situation indicator value; based on the multi-indicator decision-making theory, analyze the situation indicator value and evaluate the urban traffic situation.

Images