Dynamic traffic optimization method and system based on multi-source data fusion

Abstract

The invention discloses a dynamic traffic optimization method and system based on multi-source data fusion, and belongs to the technical field of traffic optimization, and the method comprises the steps: obtaining a plurality of pieces of platform data, carrying out the preprocessing of the plurality of pieces of platform data, carrying out the classification operation, obtaining spatial dimension data and time dimension data, training a pre-constructed combined traffic prediction model by using the historical data of the adjacent stations and the historical data of the current station, and inputting the real-time data of the adjacent stations and the real-time data of the current station into the trained combined traffic prediction model to obtain future traffic flow data; and generating a corresponding traffic optimization scheme. According to the method, multi-source data are fused, the traffic conditions of the current station and the adjacent stations are comprehensively and accurately reflected, the future traffic condition is accurately predicted by adopting the combined traffic prediction model, and whether the traffic condition needs to be optimized or not is judged according to the predicted future traffic condition, so that the traffic operation efficiency is improved; the situation of traffic jam is avoided, and the travel experience of the vehicle owner is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of traffic optimization, and particularly relates to a dynamic traffic optimization method and system based on multi-source data fusion. Background Art

[0002] With the increase in the population and the continuous growth of the motor vehicle ownership, the lag of traffic infrastructure is difficult to meet the rapidly rising travel demand. The traffic congestion problem has become a common challenge faced by major cities around the world. Especially during holidays, the traffic flow speed on highways is significantly lower than the designed speed. Traditional traffic management methods, such as manual command and static traffic signal control, are difficult to meet the growing traffic demand. How to use intelligent means to achieve intelligent management of traffic has become a research hotspot in the traffic field. In recent years, due to the rapid development of technologies such as big data and artificial intelligence, new opportunities have been provided for the research and application of intelligent transportation systems. The intelligent transportation system realizes real-time monitoring, prediction, and optimal control of traffic flow by collecting and analyzing traffic data, thereby improving traffic operation efficiency, alleviating traffic congestion, and improving travel experience.

[0003] However, there are still some limitations in the data processing of existing intelligent transportation systems. Traditional traffic data has a single type and limited coverage, making it difficult to comprehensively reflect complex traffic conditions and prone to data blind spots. The current traffic data sources have spatio-temporal properties and non-linearity, making it difficult to accurately analyze the current traffic flow data. Moreover, in practical applications, traffic flow is affected by various factors, with high dynamics and uncertainty, resulting in a decline in prediction accuracy and inability to achieve real-time and accurate traffic prediction and optimal control.

[0004] Based on the above deficiencies, how to provide an effective technical solution to solve the problems of single data type, difficult to accurately analyze the current traffic flow data, and inability to achieve real-time and accurate traffic prediction and optimal control has become a difficult problem to be solved in the existing technology. Summary of the Invention

[0005] The purpose of the present invention is to provide a dynamic traffic optimization method and system based on multi-source data fusion to solve the above problems existing in the prior art.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions: In the first aspect, the present invention provides a dynamic traffic optimization method based on multi-source data fusion, including: Obtain multiple platform data, where the multiple platform data includes vehicle platform data, highway platform data, and weather platform data. The vehicle platform data includes vehicle information and driving information. The highway platform data includes vehicle congestion and real-time traffic flow. The weather platform data includes temperature data and humidity data, and preprocess the vehicle platform data, highway platform data, and weather platform data; Classify the preprocessed vehicle platform data, highway platform data, and weather platform data to obtain spatial dimension data and time dimension data. The spatial dimension data includes real-time data of adjacent stations and historical data of adjacent stations. The time dimension data includes real-time data of the current station and historical data of the current station; Input the historical data of adjacent stations and the historical data of the current station into a pre-constructed combined traffic prediction model for training to obtain a trained combined traffic prediction model. Input the real-time data of adjacent stations and the real-time data of the current station into the trained combined traffic prediction model to obtain a prediction result, where the prediction result is future traffic flow data; Based on the prediction result, determine whether to trigger a traffic warning mechanism. If triggered, generate a traffic optimization plan.

[0007] In a possible design, preprocessing the vehicle platform data, highway platform data, and weather platform data includes: Use the wavelet analysis threshold method to denoise the vehicle platform data, highway platform data, and weather platform data; Use the combined threshold filling method to fill in the missing values of the denoised vehicle platform data, highway platform data, and weather platform data; Normalize the vehicle platform data, highway platform data, and weather platform data after the missing value filling process.

[0008] In a possible design, the platform data includes a timestamp and the geographical location information of each highway station; classifying the preprocessed vehicle platform data, highway platform data, and weather platform data to obtain spatial dimension data and time dimension data includes: Extract the features of the preprocessed vehicle platform data, highway platform data, and weather platform data based on the timestamp to obtain real-time feature data and historical feature data. The real-time feature data includes real-time vehicle features, real-time highway features, and real-time weather features. The historical feature data includes historical vehicle features, historical highway features, and historical weather features; Based on the geographical location information of each highway station, divide the real-time highway features and historical highway features into stations to obtain the real-time highway features and historical highway features of multiple highway stations; Select a certain high-speed station as the current station, and obtain spatial dimension data and time dimension data based on the real-time vehicle characteristics, historical vehicle characteristics, real-time weather characteristics, and historical weather characteristics of the current station, as well as the real-time vehicle characteristics, historical vehicle characteristics, real-time weather characteristics, and historical weather characteristics of adjacent stations.

[0009] In a possible design, the combined traffic prediction model includes multiple hidden layers and an LSTM layer; the process of inputting the historical data of adjacent stations and the historical data of the current station into a pre-constructed combined traffic prediction model for training to obtain a trained combined traffic prediction model includes: Extract the spatial feature data of the historical data of adjacent stations to obtain historical spatial features, and input the historical spatial features into the hidden layer to obtain a historical spatial prediction result; Extract the time feature data of the historical data of the current station to obtain historical time features, and input the historical time features into the LSTM layer to obtain a historical time prediction result; Fuse the historical spatial prediction result and the historical time prediction result to obtain a historical prediction result; Calculate the error value between the historical prediction result and the historical data of the current station based on the historical prediction result, and correct the pre-constructed combined traffic prediction model according to the error value; Repeat the iterative training until the error value is less than a preset threshold to obtain a trained combined traffic prediction model.

[0010] In a possible design, inputting the historical spatial features into the hidden layer to obtain a historical time prediction result includes: Select multiple data from the historical spatial features and use the data as cluster centers; Calculate the distances between the historical spatial features and each cluster center, and classify the historical spatial features to the nearest cluster center according to the distances; Iteratively calculate the distances between the cluster centers to obtain the distances between each cluster center, use the cluster centers as the center points of the hidden layer, and update the weights of the hidden layer based on the distances between each center point of the hidden layer; Obtain a historical spatial prediction result based on the historical spatial features and the weights of the hidden layer.

[0011] In a possible design, the calculation expression for obtaining a historical spatial prediction result based on the historical spatial features and the weights of the hidden layer is: ; In the above formula, is the historical spatial prediction result; is the number of hidden layers; is the weight of the i-th hidden layer; is the historical spatial feature of the i-th layer; is the bias term.

[0012] In a possible design, the combined traffic prediction model is constructed based on the RBF model and the Bi-LSTM model.

[0013] In a second aspect, the present invention provides a dynamic traffic optimization system based on multi-source data fusion, including: An acquisition module, configured to acquire multiple platform data, where the multiple platform data includes vehicle platform data, highway platform data, and weather platform data, and preprocess the vehicle platform data, highway platform data, and weather platform data; A classification module, configured to classify the preprocessed vehicle platform data, highway platform data, and weather platform data to obtain spatial dimension data and temporal dimension data, where the spatial dimension data includes real-time data and historical data of adjacent stations, and the temporal dimension data includes real-time data and historical data of the current station; A prediction module, configured to input the historical data of adjacent stations and the historical data of the current station into a pre-constructed combined traffic prediction model for training to obtain a trained combined traffic prediction model, and input the real-time data of adjacent stations and the real-time data of the current station into the trained combined traffic prediction model to obtain a prediction result, where the prediction result is future traffic flow data; An execution module, configured to determine whether to trigger a traffic warning mechanism based on the prediction result, and if so, generate a traffic optimization plan.

[0014] In a third aspect, the present invention provides a dynamic traffic optimization device based on multi-source data fusion, including a memory, a processor, and a transceiver that are communicatively connected in sequence, where the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the dynamic traffic optimization method based on multi-source data fusion as described in the first aspect.

[0015] In a fourth aspect, the present invention provides a computer program product containing instructions, which, when run on a computer, cause the computer to execute the dynamic traffic optimization method based on multi-source data fusion as described in any one of the above.

[0016] The beneficial effects of the present invention are as follows: The present invention discloses a dynamic traffic optimization method, system and product based on multi-source data fusion, including obtaining multiple platform data, where the multiple platform data includes vehicle platform data, highway platform data and weather platform data, preprocessing the multiple platform data, classifying the preprocessed platform data to obtain spatial dimension data and temporal dimension data, where the spatial dimension data includes real-time data of adjacent stations and historical data of adjacent stations, and the temporal dimension data includes real-time data of the current station and historical data of the current station, inputting the historical data of adjacent stations and the historical data of the current station into a pre-constructed combined traffic prediction model for training to obtain a trained combined traffic prediction model, inputting the real-time data of adjacent stations and the real-time data of the current station into the trained combined traffic prediction model to obtain a prediction result, where the prediction result is future traffic flow data; judging whether to trigger a traffic warning mechanism based on the prediction result, and if triggered, generating a traffic optimization plan. The present invention fuses the data of multiple platforms, uses a variety of data types, can comprehensively and accurately reflect the traffic conditions of the current station and adjacent stations, and uses a combined traffic prediction model to accurately predict future traffic conditions, and judges whether it is necessary to optimize the traffic conditions according to the predicted future traffic conditions, so as to improve traffic operation efficiency, avoid traffic congestion, and improve the travel experience of car owners. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a flowchart of the dynamic traffic optimization method based on multi-source data fusion provided in the first aspect of this embodiment; Figure 2 is a module diagram of the dynamic traffic optimization system based on multi-source data fusion provided in the second aspect of this embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the present invention in combination with the drawings and the descriptions of the embodiments or the prior art. Obviously, the following descriptions of the structures of the drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. It should be noted here that the descriptions of these embodiments are used to help understand the present invention, but do not constitute a limitation to the present invention.

[0019] Embodiment: As Figure 1 shown, the first aspect of this embodiment provides a dynamic traffic optimization method based on multi-source data fusion, which can be but is not limited to being executed by a computer device or virtual machine with certain computing resources, such as an electronic device such as a personal computer or a smart phone, or a virtual machine; as Figure 1As shown, the dynamic traffic optimization method based on multi-source data fusion may, but is not limited to, include the following steps: S1. Obtain multiple platform data, where the multiple platform data includes vehicle platform data, highway platform data, and weather platform data, and preprocess the vehicle platform data, highway platform data, and weather platform data; Specifically, in this embodiment, preferably, the vehicle platform data includes vehicle information and driving information; the highway platform data includes vehicle congestion conditions and real-time traffic flow; the weather platform data includes weather data such as temperature and humidity, and the platform data also includes timestamps and geographical location information of each highway station.

[0020] Further, in step S1, preprocessing the vehicle platform data, highway platform data, and weather platform data includes: S101. Use the wavelet analysis threshold method to denoise the vehicle platform data, highway platform data, and weather platform data; Specifically, the principle of the wavelet analysis threshold method is to decompose the signal into wavelet coefficients of different frequencies and scales through wavelet transform, then apply threshold processing to the wavelet coefficients, and finally reconstruct the signal through inverse wavelet transform to achieve the purpose of denoising.

[0021] S102. Use the combined threshold filling method to fill in the missing values of the denoised vehicle platform data, highway platform data, and weather platform data; Specifically, using the combined threshold filling method to fill in the missing values of the denoised vehicle platform data, highway platform data, and weather platform data includes filling in the missing values using the least squares support vector machine model according to univariate data and multivariate data, and using the combined threshold filling method to fuse the multivariate filling and univariate filling methods to fill in the denoised vehicle platform data, highway platform data, and weather platform data.

[0022] S103. Normalize the vehicle platform data, highway platform data, and weather platform data after the missing value filling process.

[0023] Normalizing the data can improve the generalization ability and processing efficiency of the model, and improve the accuracy and practicality of the processing.

[0024] S2. Classify the preprocessed vehicle platform data, highway platform data, and weather platform data to obtain spatial dimension data and time dimension data. The spatial dimension data includes real-time data of adjacent stations and historical data of adjacent stations, and the time dimension data includes real-time data of the current station and historical data of the current station; Specifically, in step S2, the preprocessed vehicle platform data, highway platform data, and weather platform data are classified to obtain spatial dimension data and time dimension data, including: S201. Extract the features of the preprocessed vehicle platform data, highway platform data, and weather platform data based on timestamps to obtain real-time feature data and historical feature data. The real-time feature data includes real-time vehicle features, real-time highway features, and real-time weather features, and the historical feature data includes historical vehicle features, historical highway features, and historical weather features; S202. Divide the real-time highway features and historical highway features by the geographical location information of each highway station to obtain the real-time highway features and historical highway features of multiple highway stations; S203. Select a certain highway station as the current station, and obtain spatial dimension data and time dimension data based on the real-time vehicle features, historical vehicle features, real-time weather features, and historical weather features of the current station, as well as the real-time vehicle features, historical vehicle features, real-time weather features, and historical weather features of adjacent stations.

[0025] S3. Input the historical data of adjacent stations and the historical data of the current station into the pre-constructed combined traffic prediction model for training to obtain the trained combined traffic prediction model. Input the real-time data of adjacent stations and the real-time data of the current station into the trained combined traffic prediction model to obtain the prediction result, which is the future traffic flow data; In a possible design, the combined traffic prediction model is constructed based on the RBF (Radial Basis Function) model and the Bi-LSTM (Bidirectional Long Short-Term Memory) model. It includes an input layer, multiple hidden layers, an LSTM layer, and an output layer. The principle of the RBF model is to map data to a high-dimensional space, making the problem that is linearly inseparable in the low-dimensional space become linearly separable in the high-dimensional space. The principle of the Bi-LSTM model is to combine the outputs of the forward and backward LSTMs and can consider the context information of the sequence simultaneously.

[0026] Furthermore, in this embodiment, the FA algorithm (Firefly Algorithm) is used to optimize the parameters of the RBF model and the Bi-LSTM model, and a residual network is introduced into the RBF model to enhance the feature extraction ability of the model.

[0027] Specifically, in step S3, inputting the historical data of adjacent stations and the historical data of the current station into the pre-constructed combined traffic prediction model for training to obtain the trained combined traffic prediction model includes: S301. Extract the spatial feature data of the historical data of adjacent stations to obtain the historical spatial features, and input the historical spatial features into the hidden layer to obtain the historical spatial prediction result; Specifically, in step S301, inputting the historical spatial features into the hidden layer to obtain the historical spatial prediction result includes: In this embodiment, preferably, the k-means method is used to select the hidden layer centers.

[0028] S3011. Select multiple data from the historical spatial features and use the data as the clustering centers; S3012. Calculate the distances between the historical spatial features and each clustering center, and classify the historical spatial features to the nearest clustering center according to the distances; S3013. Iteratively calculate the distances between the clustering centers to obtain the distances between each clustering center, use the clustering centers as the hidden layer center points, and update the weights of the hidden layer based on the distances between each hidden layer center point; S3014. Obtain the historical spatial prediction result based on the historical spatial features and the weights of the hidden layer.

[0029] Further, the calculation expression for obtaining the historical spatial prediction result based on the historical spatial features and the weights of the hidden layer is: ; In the above formula, is the historical spatial prediction result; is the number of hidden layers; is the weight of the i-th hidden layer; is the historical spatial feature of the i-th layer; is the bias term.

[0030] S302. Extract the time feature data of the historical data of the current station to obtain the historical time features, and input the historical time features into the LSTM layer to obtain the historical time prediction result; Specifically, in the LSTM layer, the data is processed bidirectionally according to the time series. The historical time features are processed in the forward direction of the time series, and the historical time features are processed in the reverse direction of the time series. The processing results of the two are combined to obtain the historical time prediction result.

[0031] S303. Fuse the historical spatial prediction result and the historical time prediction result to obtain the historical prediction result; S304. Calculate the error value between the historical prediction result and the historical data of the current station, and correct the pre-constructed combined traffic prediction model according to the error value; S305. Repeatedly iterate the training until the error value is less than the preset threshold to obtain the trained combined traffic prediction model.

[0032] S4. Determine whether to trigger the traffic warning mechanism based on the prediction result. If triggered, generate a traffic optimization plan.

[0033] Specifically, in this embodiment, the traffic warning mechanism is to determine whether the future traffic flow data reaches a preset traffic volume threshold. If it reaches the preset traffic volume threshold, a corresponding traffic optimization plan will be generated, such as releasing real-time information, combining with the navigation software to guide the vehicle owners to choose other routes to drive, and the route can be planned in advance; controlling the ramp, dynamically adjusting the speed and quantity of vehicles entering the main line according to the traffic flow data, etc.

[0034] This embodiment discloses a dynamic traffic optimization method based on multi-source data fusion. By obtaining the data of multiple platforms, preprocessing the data of multiple platforms, classifying the preprocessed data, spatial dimension data and time dimension data are obtained. Among them, the spatial dimension data includes real-time data of adjacent stations and historical data of adjacent stations, and the time dimension data includes real-time data of the current station and historical data of the current station. Input the historical data of adjacent stations and the historical data of the current station into the combined traffic prediction model for training to obtain the trained combined traffic prediction model. Input the real-time data of adjacent stations and the real-time data of the current station into the trained combined traffic prediction model to obtain the predicted future traffic flow data. Determine whether to trigger the traffic warning mechanism based on the predicted future traffic flow data. If triggered, generate a corresponding traffic optimization plan. Fusing multi-source data can comprehensively reflect the complex traffic conditions, avoid the emergence of data blind spots, and use the trained combined traffic prediction model for prediction, processing from the time dimension and the spatial dimension, improving the prediction accuracy, realizing real-time dynamic prediction, and generating a corresponding optimization plan according to the predicted traffic flow to control the traffic flow and avoid congestion and other situations.

[0035] As Figure 2 shown, the second aspect of this embodiment provides a dynamic traffic optimization system based on multi-source data fusion, including: An acquisition module, configured to acquire data of multiple platforms. The data of multiple platforms includes vehicle platform data, highway platform data, and weather platform data, and preprocess the vehicle platform data, highway platform data, and weather platform data; A classification module, configured to classify the preprocessed vehicle platform data, highway platform data, and weather platform data to obtain spatial dimension data and time dimension data. The spatial dimension data includes real-time data of adjacent stations and historical data of adjacent stations, and the time dimension data includes real-time data of the current station and historical data of the current station; A prediction module, configured to input historical data of neighboring stations and historical data of the current station into a pre-built combined traffic prediction model for training to obtain a trained combined traffic prediction model, and input real-time data of neighboring stations and real-time data of the current station into the trained combined traffic prediction model to obtain a prediction result, where the prediction result is future traffic flow data; An execution module, configured to determine whether to trigger a traffic warning mechanism based on the prediction result, and if triggered, generate a traffic optimization plan.

[0036] In the third aspect of this embodiment, a dynamic traffic optimization device based on multi-source data fusion is provided. Instructions are stored on the computer-readable storage medium, and when the instructions run on a computer, they are used to execute the dynamic traffic optimization method based on multi-source data fusion as described in the first aspect of the embodiment. Among them, the computer-readable storage medium refers to a carrier for storing data, and may include, but is not limited to, computer-readable storage media such as floppy disks, optical discs, hard disks, flash memories, USB flash drives, and / or memory sticks. The computer may be a general-purpose computer, a dedicated computer, a computer network, or other programmable devices.

[0037] For the working process, working details, and technical effects of the computer-readable storage medium provided in the third aspect of this embodiment, reference may be made to the dynamic traffic optimization method based on multi-source data fusion as described in the first aspect of the embodiment, and details will not be elaborated here.

[0038] In the fourth aspect of this embodiment, a computer program product is provided, including a computer program or instructions, and when the computer program or the instructions are executed by a computer, they are used to implement the dynamic traffic optimization method based on multi-source data fusion as described in the first aspect of the embodiment.

[0039] For the working process, working details, and technical effects of the computer program product provided in the fourth aspect of this embodiment, reference may be made to the dynamic traffic optimization method based on multi-source data fusion as described in the first aspect of the embodiment, and details will not be elaborated here.

[0040] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A dynamic traffic optimization method based on multi-source data fusion, characterized in that Including: Obtain multiple platform data, where the multiple platform data includes vehicle platform data, highway platform data, and weather platform data. The vehicle platform data includes vehicle information and driving information. The highway platform data includes vehicle congestion conditions and real-time traffic flow. The weather platform data includes temperature data and humidity data, and preprocess the vehicle platform data, highway platform data, and weather platform data; Classify the preprocessed vehicle platform data, highway platform data, and weather platform data to obtain spatial dimension data and time dimension data. The spatial dimension data includes real-time data and historical data of adjacent stations. The time dimension data includes real-time data and historical data of the current station; Input the historical data of adjacent stations and the historical data of the current station into a pre-constructed combined traffic prediction model for training to obtain a trained combined traffic prediction model. Input the real-time data of adjacent stations and the real-time data of the current station into the trained combined traffic prediction model to obtain a prediction result, where the prediction result is future traffic flow data; Based on the prediction result, determine whether to trigger a traffic warning mechanism. If triggered, generate a traffic optimization plan.

2. The dynamic traffic optimization method based on multi-source data fusion according to claim 1, wherein Preprocessing the vehicle platform data, highway platform data, and weather platform data includes: Use the wavelet analysis threshold method to denoise the vehicle platform data, highway platform data, and weather platform data; Use the combined threshold filling method to fill in the missing values of the denoised vehicle platform data, highway platform data, and weather platform data; Normalize the vehicle platform data, highway platform data, and weather platform data after the missing value filling process.

3. The dynamic traffic optimization method based on multi-source data fusion according to claim 1, characterized in that, The platform data includes a timestamp and the geographical location information of each highway station; The classification of the preprocessed vehicle platform data, highway platform data, and weather platform data to obtain spatial dimension data and time dimension data includes: Extract the features of the preprocessed vehicle platform data, highway platform data, and weather platform data based on the timestamp to obtain real-time feature data and historical feature data. The real-time feature data includes real-time vehicle features, real-time highway features, and real-time weather features. The historical feature data includes historical vehicle features, historical highway features, and historical weather features; Based on the geographical location information of each highway station, divide the real-time highway features and historical highway features into stations to obtain the real-time highway features and historical highway features of multiple highway stations; Select a certain highway station as the current station, and based on the real-time vehicle features, historical vehicle features, real-time weather features, and historical weather features of the current station, as well as the real-time vehicle features, historical vehicle features, real-time weather features, and historical weather features of adjacent stations, obtain spatial dimension data and time dimension data.

4. The dynamic traffic optimization method based on multi-source data fusion according to claim 1, characterized in that The combined traffic prediction model includes multiple hidden layers and an LSTM layer. The input of the historical data of adjacent stations and the historical data of the current station into the pre-constructed combined traffic prediction model for training to obtain a trained combined traffic prediction model includes: Extract the spatial feature data of the historical data of neighboring stations to obtain historical spatial features, and input the historical spatial features into the hidden layer to obtain the historical spatial prediction result; Extract the temporal feature data of the historical data of the current station to obtain historical temporal features, and input the historical temporal features into the LSTM layer to obtain the historical temporal prediction result; Fuse the historical spatial prediction result and the historical temporal prediction result to obtain the historical prediction result; Calculate the error value between the historical prediction result and the historical data of the current station, and correct the pre-constructed combined traffic prediction model according to the error value; Repeat the iterative training until the error value is less than the preset threshold to obtain the trained combined traffic prediction model.

5. The dynamic traffic optimization method based on multi-source data fusion according to claim 4, characterized in that Input the historical spatial features into the hidden layer to obtain the historical temporal prediction result, including: Select multiple data from the historical spatial features and use the data as the clustering centers; Calculate the distances between the historical spatial features and each clustering center, and classify the historical spatial features to the nearest clustering center according to the distances; Iteratively calculate the distances between the clustering centers to obtain the distances between each clustering center, use the clustering centers as the hidden layer center points, and update the weights of the hidden layer based on the distances between each hidden layer center point; Obtain the historical spatial prediction result based on the historical spatial features and the weights of the hidden layer.

6. The dynamic traffic optimization method based on multi-source data fusion according to claim 5, wherein The calculation expression for obtaining the historical spatial prediction result based on the historical spatial features and the weights of the hidden layer is: ； In the above formula, is the historical space prediction result; is the number of hidden layers; is the weight of the i-th hidden layer; is the historical space feature of the i-th layer; is the bias term.

7. The dynamic traffic optimization method based on multi-source data fusion according to claim 1, characterized in that The combined traffic prediction model is constructed based on the RBF model and the Bi-LSTM model.

8. A dynamic traffic optimization system based on multi-source data fusion, characterized in that, Including: An acquisition module, configured to acquire multiple platform data, where the multiple platform data includes vehicle platform data, highway platform data, and weather platform data, and preprocess the vehicle platform data, highway platform data, and weather platform data; A classification module, configured to classify the preprocessed vehicle platform data, highway platform data, and weather platform data to obtain spatial dimension data and temporal dimension data, where the spatial dimension data includes real-time data and historical data of neighboring stations, and the temporal dimension data includes real-time data and historical data of the current station; A prediction module, configured to input the historical data of neighboring stations and the historical data of the current station into the pre-constructed combined traffic prediction model for training to obtain the trained combined traffic prediction model, and input the real-time data of neighboring stations and the real-time data of the current station into the trained combined traffic prediction model to obtain a prediction result, where the prediction result is future traffic flow data; An execution module, configured to determine whether to trigger a traffic warning mechanism based on the prediction result, and if triggered, generate a traffic optimization plan.

9. A dynamic traffic optimization device based on multi-source data fusion, characterized in that, Including a memory, a processor, and a transceiver that are communicatively connected in sequence, where the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the dynamic traffic optimization method based on multi-source data fusion according to any one of claims 1 to 7.

10. A computer program product comprising a computer program or instructions, characterized in that, The computer program or the instruction, when executed by a computer, implements the dynamic traffic optimization method based on multi-source data fusion according to any one of claims 1 to 7.

Images