Smart city operation data management system and method

Abstract

The invention relates to the field of smart city management, and provides a smart city operation data management system and method, which adopts a deep learning-based data processing and analysis technology to understand the time sequence change of the traffic flow of an intersection by performing time sequence analysis on the time queue data of the traffic flow of a target management intersection. And the time configuration of the traffic signal lamp is dynamically adjusted based on the time sequence change of the intersection traffic flow. Therefore, the traffic signal lamp time can be scientifically and reasonably configured according to the actual traffic flow, and the operation effect of an urban traffic network can be optimized.

Description

Technical Field

[0001] The present application relates to the field of smart city management, and more specifically, to a smart city operation data management system and method. Background Art

[0002] Smart city refers to the use of various information technologies to connect and integrate the city's systems and services to improve resource utilization efficiency, optimize urban management and services, and improve the quality of life of citizens. With the acceleration of global urbanization, urban population density continues to rise, the contradiction between resource supply and demand has become increasingly prominent, and traffic congestion, environmental pollution, energy waste and other issues have become common challenges faced by many cities. In this context, the construction of smart cities is particularly important. Taking urban traffic management as an example, traditional traffic management methods usually rely on the setting of traffic lights with fixed durations, which is difficult to adapt to the real-time dynamic changes in traffic flow. Specifically, during peak hours in the morning and evening, when the traffic volume is large, the green light time is too short, which cannot effectively guide traffic and causes serious congestion; while in the middle of the night or when the traffic volume is small, the green light time is too long, which will lead to resource waste and reduce traffic efficiency.

[0003] Therefore, a smart city operation data management solution is needed. Summary of the invention

[0004] In view of the shortcomings of the prior art, this application provides a smart city operation data management system and method, which adopts data processing and analysis technology based on deep learning, performs time series analysis on the time queue data of the traffic flow at the target management intersection to understand the time series changes of the traffic flow at the intersection, and dynamically adjusts the time configuration of the traffic lights based on the time series changes of the traffic flow at the intersection. In this way, it is possible to scientifically and reasonably configure the traffic light time according to the actual traffic flow, which is conducive to optimizing the operation effect of the urban transportation network.

[0005] According to one aspect of the present application, a smart city operation data management method is provided, which includes: Obtain the time queue data of the traffic flow at the target managed intersection; Performing data analysis on the time queue data of the traffic flow to obtain a fusion vector of the time series characteristics of the intersection traffic flow; The traffic lights at the target managed intersection are controlled according to the information in the fusion vector of the traffic flow time series characteristics at the intersection.

[0006] According to another aspect of the present application, a smart city operation data management system is provided, comprising: The intersection traffic flow data collection module is used to obtain the time queue data of the traffic flow at the target management intersection; A traffic flow data analysis module for an intersection, used for performing data analysis on the time queue data of the traffic flow to obtain a fusion vector of the time series characteristics of the traffic flow at the intersection; The traffic light control module is used to control the traffic light at the target management intersection according to the information in the fusion vector of the traffic flow time series characteristics at the intersection.

[0007] This application has significant technical effects due to the adoption of the above technical solutions: The smart city operation data management system and method provided by the present application adopts data processing and analysis technology based on deep learning, performs time series analysis on the time queue data of the traffic flow at the target management intersection to understand the time series changes of the traffic flow at the intersection, and dynamically adjusts the time configuration of the traffic lights based on the time series changes of the traffic flow at the intersection. In this way, it is possible to scientifically and reasonably configure the traffic light time according to the actual traffic flow, which is conducive to optimizing the operation effect of the urban transportation network. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 Flow chart of a smart city operation data management method according to an embodiment of the present application.

[0010] Figure 2 This is a flowchart of step S2 in the smart city operation data management method according to an embodiment of the present application.

[0011] Figure 3 This is a flowchart of step S3 in the smart city operation data management method according to an embodiment of the present application.

[0012] Figure 4 This is a system block diagram of a smart city operation data management system according to an embodiment of the present application. DETAILED DESCRIPTION

[0013] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0014] Based on the problems in the above-mentioned background technology, this application proposes a smart city operation data management method. Figure 1 Flow chart of the smart city operation data management method according to an embodiment of the present application. Figure 1 As shown, the smart city operation data management method according to the embodiment of the present application includes: S1, obtaining the time queue data of the traffic flow at the target management intersection; S2, performing data analysis on the time queue data of the traffic flow to obtain the intersection traffic flow time series feature fusion vector; S3, controlling the traffic lights at the target management intersection according to the information in the intersection traffic flow time series feature fusion vector.

[0015] In step S1, the time queue data of the traffic flow at the target management intersection is obtained. It should be understood that the obtained time queue data of the traffic flow specifically includes the statistical values ​​of the number of vehicles passing through the target management intersection in various directions (such as east-west, north-south, etc.) at different time points. For example, the specific number of various types of vehicles such as cars, buses, trucks, etc. passing through the intersection is recorded every 5 minutes to present the trend of traffic flow over time, which can intuitively reflect when is the peak period of traffic flow and when is the trough period. For example, the traffic flow data at the morning peak of 7-9 o'clock on weekdays may show a rapid increase and a large value, while the traffic flow in the late night period will be very low. Specifically, by analyzing the size of the traffic flow in each direction in the time queue data of the traffic flow, the changes in different time periods, etc., the distribution proportion of the green light time of the traffic signal in each direction is determined. For example, if the east-west traffic volume at a certain intersection during the morning rush hour is much greater than that in the north-south direction, then the generated time configuration plan will appropriately extend the green light duration in the east-west direction and shorten the green light duration in the north-south direction to ensure that vehicles in the direction of heavy traffic can pass more smoothly and avoid congestion, thereby dynamically and reasonably allocating signal light time according to actual traffic flow.

[0016] In step S2, data analysis is performed on the time queue data of the traffic flow to obtain a fusion vector of the time series feature of the intersection traffic flow. Specifically, Figure 2 FIG. 1 is a flow chart of step S2 in the smart city operation data management method according to an embodiment of the present application. Figure 2 As shown, the step S2 includes: S21, extracting time series features from the time queue data of the traffic flow to obtain a time series feature vector of the intersection traffic flow; S22, performing context association encoding on the time series feature vector of the intersection traffic flow to obtain a time series context association feature vector of the intersection traffic flow; S23, fusing the time series feature vector of the intersection traffic flow and the time series context association feature vector of the intersection traffic flow to obtain a time series feature fusion vector of the intersection traffic flow.

[0017] In step S21, the time queue data of the traffic flow is subjected to time series feature extraction to obtain a time series feature vector of the traffic flow at the intersection. Specifically, in the embodiment of the present application, the step S21 includes: inputting the time queue data of the traffic flow into a multi-scale traffic flow sequence encoder to obtain a time series feature vector of the traffic flow at the intersection. It should be understood that the traffic flow presents complex dynamic changes at different times and is affected by various factors, such as working days and holidays, different time periods in a day (morning peak, noon trough, evening peak, etc.), seasonal changes, temporary traffic control, emergencies, etc. It is difficult to directly grasp the regularity and key features contained therein by relying solely on the original traffic flow time queue data. For example, by only looking at the traffic flow values ​​recorded every 5 minutes, it is difficult to see at a glance the periodic fluctuations of traffic flow over a long period of time and the mutual influence relationship between traffic flows in different directions, so it is necessary to mine deeper information through time series feature extraction. By extracting time series features, it is possible to find out the inherent laws of traffic flow changes over time, such as identifying the peaks and troughs of traffic flow that occur at fixed time periods every day, similar trends of traffic flow changes on weekdays, and overall differences in traffic flow in different seasons. Specifically, the time queue data of the traffic flow is input into a multi-scale traffic flow sequence encoder for processing. In particular, in the present application, the multi-scale traffic flow sequence encoder is a convolutional neural network model comprising a first convolutional layer and a second convolutional layer, the first convolutional layer uses a one-dimensional convolutional kernel of a first scale, and the second convolutional layer uses a one-dimensional convolutional kernel of a second scale, and the first scale is different from the second scale. The law of change of traffic flow is complex, involving the interweaving of multiple time scales, and it is difficult for a single-scale convolutional kernel to fully describe its characteristics. The use of multi-scale convolutional kernels can generate rich and diverse feature representations, providing a more comprehensive basis for the subsequent generation of traffic light time configuration schemes. That is, the use of a convolutional kernel of the first scale can capture subtle changes in traffic flow in a short period of time, such as a sudden increase or decrease in traffic flow within a few hours, while a larger-scale convolutional kernel can focus on traffic flow trends over a longer period of time, such as a few days or even longer. By capturing features at multiple scales, we can more comprehensively understand the changing characteristics of traffic flow at different time dimensions, thereby avoiding missing important information by focusing on a single scale.

[0018] In particular, in a specific embodiment of the present application, the time queue data of the traffic flow is input into a multi-scale traffic flow sequence encoder to obtain the time series feature vector of the intersection traffic flow, including: arranging the time queue data of the traffic flow according to the time dimension to obtain a traffic flow time series input vector; inputting the traffic flow time series input vector into the first convolution layer of the multi-scale traffic flow sequence encoder to obtain a first-scale intersection traffic flow time series feature vector; inputting the traffic flow time series input vector into the second convolution layer of the multi-scale traffic flow sequence encoder to obtain a second-scale intersection traffic flow time series feature vector; cascading the first-scale intersection traffic flow time series feature vector and the second-scale intersection traffic flow time series feature vector to obtain the intersection traffic flow time series feature vector.

[0019] In step S22, the intersection traffic flow time series feature vector is context-associated encoded to obtain the intersection traffic flow time series context-associated feature vector. Specifically, in the embodiment of the present application, the step S22 includes: dividing the intersection traffic flow time series feature vector into a plurality of intersection traffic flow time series feature sub-vectors and inputting the time series feature context-associated capturer to obtain the intersection traffic flow time series context-associated feature vector. It should be understood that traffic flow does not exist in isolation in the time dimension, and the traffic flow situation at a certain moment is often closely related to the traffic conditions before and after. For example, when traffic congestion occurs at an intersection, the traffic flow size and vehicle queue length in the previous time period will directly affect the traffic efficiency at the current moment and the degree of aggregation of subsequent vehicles; for example, when the green light in a certain direction is on, vehicles begin to pass, and the smoothness of the traffic at this moment will depend on the remaining vehicles at the intersection at the end of the previous green light cycle and whether there is continuous traffic flow in the future. Therefore, relying solely on the time series feature vector of intersection traffic flow obtained by multi-scale feature extraction is not enough to fully grasp the contextual correlation of traffic flow in the time series. Further special correlation coding processing is needed to mine these implicit correlation information.

[0020] Specifically, in order to fully consider the contextual association characteristics of traffic flow, it is necessary to first divide the intersection traffic flow time series feature vector into multiple intersection traffic flow time series feature sub-vectors, and then input the divided multiple vectors into a time series feature context association capturer for processing. In particular, the time series feature context association capturer in the present application is a bidirectional long short-term memory neural network model. Here, by dividing the intersection traffic flow time series feature vector, it is possible to analyze the traffic flow association situation in a shorter time period represented by each sub-vector from a more microscopic level, and at the same time, it is also possible to combine multiple sub-vectors to reflect the overall association characteristics in a longer time period. The unique advantage of the bidirectional long short-term memory neural network model is that it can simultaneously consider the forward and reverse information of the sequence data, that is, to mine the association features in the time direction from the past to the present and from the present to the future. This means that it can capture the association between the traffic flow at a certain moment and the previous time period and the subsequent time period. In addition, traffic flow has complex dependencies over a long time range. For example, although there are several hours between the morning and evening peak hours on weekdays, the traffic congestion and vehicle backlog during the morning peak will affect the traffic flow during the evening peak, which may cause the evening peak to arrive earlier or increase congestion. With its special structure and memory units, the Bi-LSTM model can effectively learn this long-term dependency of traffic flow over a long time series. Even if there are time periods represented by multiple sub-vectors, it can also mine the correlation features between them, thereby providing a more forward-looking and global basis for generating traffic light time configuration schemes, so that the signal light duration configuration can better cope with the changing trend of traffic flow over a long period of time, and avoid unreasonable configuration due to focusing only on short-term correlations.

[0021] In step S23, the intersection traffic flow time series feature vector and the intersection traffic flow time series contextual association feature vector are fused to obtain the intersection traffic flow time series feature fusion vector. It should be understood that the traffic flow time series feature vector is extracted from the time queue data of the original traffic flow by a multi-scale traffic flow sequence encoder. It focuses on the basic change characteristics of traffic flow at different time scales, such as the peak, trough and periodicity of traffic flow at different time periods, which can reflect the inherent characteristics of traffic flow at the macro and micro levels in the time dimension. The traffic flow time series contextual association feature vector is a feature vector mined on the basis of considering the correlation between the time periods before and after the traffic flow with the help of a time series feature contextual association capturer (bidirectional long short-term memory neural network model). It focuses on the dependency between traffic flows, the coherence of dynamic changes and the mutual influence between the time periods, and reveals more of the implicit logical connection and interaction of traffic flow in the time series. Fusion of these two vectors is equivalent to combining the "static" features (relatively independent time series features) and "dynamic" features (context-related features) of traffic flow. This can comprehensively and three-dimensionally characterize changes in traffic flow from multiple angles, avoiding the one-sided understanding of traffic flow based on a single type of feature vector, thereby providing a more scientific and complete basis for the subsequent generation of traffic light time configuration plans.

[0022] In particular, in a specific embodiment of the present application, the intersection traffic flow time series feature vector and the intersection traffic flow time series context association feature vector are fused to obtain the intersection traffic flow time series feature fusion vector, including: associating the intersection traffic flow time series feature vector and the intersection traffic flow time series context association feature vector with the following fusion formula to obtain the intersection traffic flow time series feature fusion vector; wherein the fusion formula is: ;in, is the fusion vector of the time series features of the traffic flow at the intersection, is the time series feature vector of the traffic flow at the intersection, is the temporal contextual feature vector of the traffic flow at the intersection, represents the addition of the elements at corresponding positions of the intersection traffic flow time series feature vector and the intersection traffic flow time series context association feature vector, and It is a weighted parameter used to control the balance between the intersection traffic flow time series feature vector and the intersection traffic flow time series context association feature vector in the intersection traffic flow time series feature fusion vector.

[0023] In step S3, the traffic lights at the target managed intersection are controlled according to the information in the intersection traffic flow time series feature fusion vector. Specifically, Figure 3 FIG. 1 is a flow chart of step S3 in the smart city operation data management method according to an embodiment of the present application. Figure 3 As shown, step S3 includes: S31, performing feature aggregation screening based on spatial reconstruction on the intersection traffic flow time series feature fusion vector to obtain an optimized intersection traffic flow time series feature fusion vector; S32, passing the optimized intersection traffic flow time series feature fusion vector through a time configuration scheme generator to obtain a time configuration scheme for traffic lights; S33, controlling the traffic lights at the target management intersection based on the time configuration scheme of the traffic lights.

[0024] In step S31, the intersection traffic flow time series feature fusion vector is subjected to feature aggregation screening based on spatial reconstruction to obtain an optimized intersection traffic flow time series feature fusion vector. In particular, the original traffic flow data contains a large amount of dynamic information, such as periodic changes in vehicle flow (morning and evening peaks), sudden anomalies (accidents, temporary road closures, etc.), and seasonal trends (holiday traffic). In the process of feature extraction and fusion, relevant information may be lost or specific features may be weakened, which will cause the intersection traffic flow time series feature fusion vector to be unable to accurately reflect the dynamic characteristics of the intersection traffic flow, that is, the model does not accurately understand the traffic flow change pattern of the target management intersection, which will cause the generated traffic light time configuration scheme to be unreasonable, and further affect the control effect of the traffic light. Therefore, in the technical solution of the present application, it is necessary to perform feature aggregation screening based on spatial reconstruction on the intersection traffic flow time series feature fusion vector to obtain an optimized intersection traffic flow time series feature fusion vector.

[0025] Specifically, in an embodiment of the present application, the intersection traffic flow timing feature fusion vector is subjected to feature aggregation screening based on spatial reconstruction to obtain an optimized intersection traffic flow timing feature fusion vector, including: extracting the target time configuration scheme generation parameter matrix of the time configuration scheme generator; performing node decomposition on the target time configuration scheme generation parameter matrix in units of row vectors to obtain a set of target time configuration scheme generation parameter node encoding vectors; using each target time configuration scheme generation parameter node encoding vector in the set of target time configuration scheme generation parameter node encoding vectors as a wandering topological space, and performing topological spatial constraints on the intersection traffic flow timing feature fusion vector to obtain a set of constrained intersection traffic flow timing feature fusion vectors; calculating the positional mean vector of the set of constrained intersection traffic flow timing feature fusion vectors to obtain the optimized intersection traffic flow timing feature fusion vector.

[0026] More specifically, in an embodiment of the present application, each target time configuration scheme generating parameter node encoding vector in the set of the target time configuration scheme generating parameter node encoding vector is used as the wandering topological space, and the intersection traffic flow time series feature fusion vector is subjected to topological space constraints respectively to obtain a set of constrained intersection traffic flow time series feature fusion vectors, including: multiplying the intersection traffic flow time series feature fusion vector and the transposed vector of the target time configuration scheme generating parameter node encoding vector, calculating the natural exponential function value of the multiplication result to obtain the time configuration scheme generating weighted exponential response weight; calculating the intersection traffic flow time series feature fusion vector and the transposed vector of the target time configuration scheme generating parameter node encoding vector; calculating the intersection traffic flow time series feature fusion vector and the transposed vector of the target time configuration scheme generating parameter node encoding vector, and calculating the natural exponential function value of the multiplication result to obtain the weighted exponential response weight of the time configuration scheme; calculating the intersection traffic flow time series feature fusion vector and the transposed vector of the target time configuration scheme generating parameter node encoding vector ... The Euclidean distance between the intersection traffic flow time series feature fusion vector and the target time configuration scheme generation parameter node coding vector is obtained to obtain the time configuration scheme generation node coding distance value; the time configuration scheme generation node coding distance value and the intersection traffic flow time series feature fusion vector are point multiplied, and the natural exponential function value is calculated for each eigenvalue of the vector after the point multiplication to obtain the time configuration scheme generation distance guidance index feature vector; the time configuration scheme generation weighted exponential response weight and the time configuration scheme generation distance guidance index feature vector are point multiplied to obtain the constrained intersection traffic flow time series feature fusion vector.

[0027] In the embodiment of the present application, specifically, the feature aggregation screening based on spatial reconstruction of the intersection traffic flow time series feature fusion vector is performed to obtain the optimized intersection traffic flow time series feature fusion vector, including: processing the intersection traffic flow time series feature fusion vector with the following optimization formula to obtain the optimized intersection traffic flow time series feature fusion vector; wherein the optimization formula is: ;in, represents the target time configuration scheme generation parameter matrix, The first, second, and third nodes of the set of encoding vectors of the target time configuration scheme generation parameter nodes are represented. , The target time configuration scheme generates the parameter node encoding vector, represents the transpose of a vector, represents the natural exponential function, represents the intersection traffic flow time series feature fusion vector, represents matrix multiplication, It means point multiplication by position. Represents the calculation vector and vector The Euclidean distance between The first one represents the set of the time series feature fusion vectors of the traffic flow at the intersection after the constraint Constrained intersection traffic flow time series feature fusion vector, Represents the total number of sets of traffic flow time series feature fusion vectors at the constrained intersection, Represents the optimized intersection traffic flow time series feature fusion vector.

[0028] That is, in response to the above technical problems, in the technical solution of the present application, the intersection traffic flow time series feature fusion vector is subjected to feature aggregation screening based on spatial reconstruction. The process first extracts the key parameters for decision-making from the trained time configuration scheme generator. These parameters form a matrix in a high-dimensional space, and each row represents the weight or influencing factor on a different dimension. The parameter matrix generated by the target time configuration scheme can provide insight into the location and shape of the model decision boundary, and then infer which input features are most critical to the prediction results.

[0029] Next, the target time configuration scheme generation parameter matrix is ​​node-decomposed in units of row vectors to obtain a set of node encoding vectors of the target time configuration scheme generation parameters. Here, each row vector is a node in graph theory, which means that each set of parameters is now regarded as an entity with potential connectivity. This transformation allows the application of methods from graph theory and network science to explore the interactions between features. The node encoding vector not only carries information about the original parameters, but also implies knowledge about the topological structure of the entire system. Node decomposition further reveals the intrinsic connection pattern or structure of the data, so that the optimized feature vector can better adapt to new task requirements.

[0030] Then, each of the target time configuration schemes in the set of parameter node encoding vectors generated by the target time configuration scheme is used as the walking topological space, and the intersection traffic flow time series feature fusion vector is subjected to topological space constraints to obtain a set of constrained intersection traffic flow time series feature fusion vectors. "Walking" in the topological space defined by the node encoding vector is actually simulating an exploratory process, the purpose of which is to find feature transformations that can best maintain the structural characteristics of the original data. Each step determines the position of the next step based on the probability distribution of the current state. The topological space constraints ensure that even in different contexts, the feature representation retains certain invariances. At the same time, it can also promote cross-domain transfer learning because it emphasizes the universal relationship between features rather than the details of a specific field. In this way, the reconstruction of the feature space is achieved, making the optimized feature vector more compact and having better generalization capabilities.

[0031] Finally, the position-wise mean vector of the set of the constrained intersection traffic flow time series feature fusion vectors is calculated to obtain the optimized intersection traffic flow time series feature fusion vector. Calculating the mean vector is a statistical aggregation method used to integrate the optimal solution from multiple perspectives. The idea behind this step is to reduce the deviation caused by a single estimate by fusing the information provided by different sample points. The averaging process is equivalent to performing a soft voting, which enhances the expressiveness of common features and makes the optimized feature vector more stable and reliable.

[0032] In step S32, the optimized intersection traffic flow time series feature fusion vector is passed through a time configuration scheme generator to obtain a time configuration scheme for traffic lights. It should be understood that traffic flow is in a dynamic state of change at all times in reality and is affected by many factors, such as travel patterns in different time periods (morning, mid-evening peak hours, off-peak periods, etc.) and emergencies (traffic accidents, road construction, etc.). In order to dynamically generate a traffic light time configuration scheme that matches the traffic flow at the intersection based on the real-time and potential changes in traffic flow reflected by the optimized intersection traffic flow time series feature fusion vector, and to ensure that the light duration can flexibly adapt to the fluctuations in traffic flow, it is necessary to pass the optimized intersection traffic flow time series feature fusion vector through a time configuration scheme generator to obtain a time configuration scheme for traffic lights. The time configuration scheme generator is a model built based on machine learning. The input layer receives the optimized intersection traffic flow time series feature fusion vector, and performs complex nonlinear transformation and information integration on these input features through the hidden layer. According to the pre-set objective function, the output layer outputs the corresponding traffic light time configuration scheme. The scheme content covers specific parameters such as the duration setting of each phase of the traffic light and the switching cycle in different time periods.

[0033] In step S33, the traffic lights at the target management intersection are controlled based on the time configuration scheme of the traffic lights. It should be understood that, according to the time configuration scheme information of the traffic lights, a longer green light duration can be reasonably allocated to the traffic lights in different directions of the target management intersection, and the switching sequence and duration interval between the lights can be coordinated to ensure that vehicles can pass through the intersection efficiently and orderly, reducing the waiting time of vehicles and congestion at the intersection. In the late night and other periods when the traffic volume is very small, the green light duration can be shortened accordingly to avoid waste of resources, so that the configuration of the traffic lights always fits the actual traffic flow and achieves precise dynamic adjustment.

[0034] In summary, the smart city operation data management method based on the embodiment of the present application is explained, which adopts data processing and analysis technology based on deep learning, performs time series analysis on the time queue data of the traffic flow at the target management intersection to understand the time series changes of the traffic flow at the intersection, and dynamically adjusts the time configuration of the traffic light based on the time series changes of the traffic flow at the intersection. In this way, it is possible to scientifically and reasonably configure the traffic light time according to the actual traffic flow, which is conducive to optimizing the operation effect of the urban transportation network.

[0035] Figure 4 : is a system block diagram of a smart city operation data management system according to an embodiment of the present application. Figure 4 As shown, according to the smart city operation data management system 100 of the embodiment of the present application, it includes: an intersection traffic flow data acquisition module 110, which is used to obtain the time queue data of the traffic flow at the target management intersection; an intersection traffic flow data analysis module 120, which is used to perform data analysis on the time queue data of the traffic flow to obtain the intersection traffic flow time series feature fusion vector; a traffic light control module 130, which is used to control the traffic light at the target management intersection according to the information in the intersection traffic flow time series feature fusion vector.

[0036] Here, those skilled in the art will appreciate that the specific functions and operations of the various units and modules in the smart city operation data management system 100 have been described in detail above. Figures 1 to 3 The present invention has been described in detail in the description of the smart city operation data management method, and therefore, its repeated description will be omitted.

[0037] In summary, the smart city operation data management system 100 based on the embodiment of the present application is explained, which adopts data processing and analysis technology based on deep learning, performs time series analysis on the time queue data of the traffic flow at the target management intersection to understand the time series changes of the traffic flow at the intersection, and dynamically adjusts the time configuration of the traffic light based on the time series changes of the traffic flow at the intersection. In this way, it is possible to scientifically and reasonably configure the traffic light time according to the actual traffic flow, which is conducive to optimizing the operation effect of the urban transportation network.

[0038] Finally, the present application also relates to a computer-readable storage medium for implementing a smart city operation data management method according to one or more embodiments of the present application. The computer-readable storage medium mentioned herein includes various types of computer storage media, which may be any available medium that can be accessed by a general-purpose or special-purpose computer. For example, a computer-readable storage medium may include RAM, ROM, EPROM, E2PROM, registers, hard disks, removable disks, CD-ROMs or other optical disk storage, disk storage or other magnetic storage devices, or any other temporary or non-temporary medium that can be used to carry or store desired program code units in the form of instructions or data structures and can be accessed by a general-purpose or special-purpose computer, or a general-purpose or special-purpose processor. The description of the computer-readable storage medium according to the present application can refer to the explanation of the smart city operation data management method according to the present application, and no further description will be given.

[0039] The terms used herein are for the purpose of describing specific embodiments only and are not intended to limit the present application. As used herein, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprise and / or comprising" when used in this specification specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, element components, and / or groups thereof.

[0040] Those skilled in the art will appreciate that various example embodiments are shown and described herein, each example embodiment having certain features in a particular embodiment, but the present application is not limited thereto. On the contrary, the present application can be modified to incorporate any number of variants, changes, substitutions, combinations, sub-combinations, or equivalent arrangements not previously described but commensurate with the scope of the present application. In addition, although various embodiments of the present application have been described, it will be understood that aspects of the present application may include only some of the described embodiments. Therefore, the present application will not be considered to be limited by the foregoing description, but only by the scope of the appended claims.

Claims

1. A smart city operation data management method, characterized in that: include: Obtain the time queue data of the traffic flow at the target managed intersection; Performing data analysis on the time queue data of the traffic flow to obtain a fusion vector of the time series characteristics of the intersection traffic flow; The traffic lights at the target managed intersection are controlled according to the information in the fusion vector of the traffic flow time series characteristics at the intersection.

2. The smart city operation data management method according to claim 1, characterized in that: Performing data analysis on the time queue data of the traffic flow to obtain a time series feature fusion vector of the intersection traffic flow, including: Extracting time series features from the time queue data of the traffic flow to obtain a time series feature vector of the intersection traffic flow; Performing context-association coding on the intersection traffic flow time series feature vector to obtain the intersection traffic flow time series context-association feature vector; The intersection traffic flow time series feature vector and the intersection traffic flow time series context association feature vector are fused to obtain the intersection traffic flow time series feature fusion vector.

3. The smart city operation data management method according to claim 2, characterized in that: Extracting time series features from the time queue data of the traffic flow to obtain a time series feature vector of the intersection traffic flow includes: inputting the time queue data of the traffic flow into a multi-scale traffic flow sequence encoder to obtain the time series feature vector of the intersection traffic flow.

4. The smart city operation data management method according to claim 3, characterized in that: Context-associated encoding is performed on the intersection traffic flow time series feature vector to obtain the intersection traffic flow time series context-associated feature vector, including: dividing the intersection traffic flow time series feature vector into multiple intersection traffic flow time series feature sub-vectors and then inputting them into a time series feature context-associated capturer to obtain the intersection traffic flow time series context-associated feature vector.

5. The smart city operation data management method according to claim 4, characterized in that: The temporal feature context association capturer is a bidirectional long short-term memory neural network model, and the multi-scale traffic flow sequence encoder is a convolutional neural network model including a first convolutional layer and a second convolutional layer, wherein the first convolutional layer uses a one-dimensional convolutional kernel of a first scale, and the second convolutional layer uses a one-dimensional convolutional kernel of a second scale, and the first scale is different from the second scale.

6. The smart city operation data management method according to claim 5, characterized in that: According to the information in the intersection traffic flow time series feature fusion vector, the traffic light at the target management intersection is controlled, including: Performing feature aggregation screening based on spatial reconstruction on the intersection traffic flow time series feature fusion vector to obtain an optimized intersection traffic flow time series feature fusion vector; Passing the optimized intersection traffic flow time series feature fusion vector through a time configuration scheme generator to obtain a time configuration scheme for traffic lights; The traffic lights at the target managed intersection are controlled based on the time configuration scheme of the traffic lights.

7. The smart city operation data management method according to claim 6, characterized in that: The intersection traffic flow time series feature fusion vector is subjected to feature aggregation screening based on spatial reconstruction to obtain an optimized intersection traffic flow time series feature fusion vector, including: Extracting a target time configuration scheme generation parameter matrix of the time configuration scheme generator; Decomposing the target time configuration scheme generation parameter matrix into nodes in units of row vectors to obtain a set of target time configuration scheme generation parameter node encoding vectors; Using each target time configuration scheme parameter node encoding vector generated in the set of target time configuration scheme parameter node encoding vectors as the walking topological space, topological space constraints are respectively applied to the intersection traffic flow time series feature fusion vectors to obtain a set of constrained intersection traffic flow time series feature fusion vectors; The position mean vector of the set of the constrained intersection traffic flow time series feature fusion vectors is calculated to obtain the optimized intersection traffic flow time series feature fusion vector.

8. The smart city operation data management method according to claim 7, characterized in that: Each target time configuration scheme parameter node encoding vector in the set of target time configuration scheme parameter node encoding vectors is used as a walking topological space, and topological space constraints are respectively applied to the intersection traffic flow time series feature fusion vectors to obtain a set of constrained intersection traffic flow time series feature fusion vectors, including: After multiplying the intersection traffic flow time series feature fusion vector and the transposed vector of the target time configuration scheme generation parameter node encoding vector, the natural exponential function value of the multiplication result is calculated to obtain the time configuration scheme generation weighted exponential response weight; Calculating the Euclidean distance between the intersection traffic flow time series feature fusion vector and the target time configuration scheme generation parameter node coding vector to obtain the time configuration scheme generation node coding distance value; Performing a dot multiplication on the node encoding distance value generated by the time configuration scheme and the intersection traffic flow time series feature fusion vector, and calculating a natural exponential function value for each eigenvalue of the vector after the dot multiplication to obtain a time configuration scheme generated distance guidance index feature vector; The weighted index response weight generated by the time configuration scheme and the distance guidance index feature vector generated by the time configuration scheme are point-multiplied to obtain the constrained intersection traffic flow time series feature fusion vector.

9. A smart city operation data management system, characterized in that: include: The intersection traffic flow data collection module is used to obtain the time queue data of the traffic flow at the target management intersection; A traffic flow data analysis module for an intersection, used for performing data analysis on the time queue data of the traffic flow to obtain a fusion vector of the time series characteristics of the traffic flow at the intersection; The traffic light control module is used to control the traffic light at the target management intersection according to the information in the fusion vector of the traffic flow time series characteristics at the intersection.

10. The smart city operation data management system according to claim 9, characterized in that: The intersection traffic flow data analysis module includes: A time series feature extraction unit for traffic flow at an intersection, used for extracting time series features from the time queue data of the traffic flow to obtain a time series feature vector of traffic flow at the intersection; A traffic flow intersection time series feature context coding unit, used for performing context association coding on the traffic flow intersection time series feature vector to obtain a traffic flow intersection time series context association feature vector; A feature fusion unit is used to fuse the intersection traffic flow time series feature vector and the intersection traffic flow time series context association feature vector to obtain the intersection traffic flow time series feature fusion vector.

Images