Electric vehicle charging demand prediction method and device, terminal and medium

Abstract

The invention discloses an electric vehicle charging demand prediction method and device, a terminal and a medium, and the method comprises the steps: building a dynamic thermodynamic diagram reflecting the traffic flow state of each road network node according to traffic charging network data through a kernel density estimation mode, capturing the traffic flow change in real time through the dynamic thermodynamic diagram, and carrying out the prediction of the charging demand of an electric vehicle. The method comprises the steps of depicting road network topology constraints in combination with a graph convolutional network, performing feature alignment on a dynamic thermodynamic diagram and external factor data on the basis of time and space dimensions, performing cross-modal attention fusion on the features of the dynamic thermodynamic diagram and the features of the external factor data through multi-modal attention processing to obtain a fusion vector, and performing cross-modal attention fusion on the features of the dynamic thermodynamic diagram and the features of the external factor data on the basis of the fusion vector. The charging demand prediction value of each road network node is determined in combination with the preset space-time diagram neural network model, so that the problem of space-time granularity difference of multi-source data is solved, the robustness of demand prediction in emergencies and high fluctuation scenes is remarkably improved, and the accuracy of charging demand prediction in extreme scenes is improved.

Description

Technical Field

[0001] The present application relates to the technical field of load forecasting, and particularly to a method, device, terminal and medium for predicting electric vehicle charging demand. Background Art

[0002] With the rapid development of electric vehicles, predicting charging demand has become a key technology for optimizing the layout of charging piles and grid dispatching. Traditional methods mainly rely on historical charging data such as charging pile usage records and vehicle driving trajectories for prediction. However, this method is difficult to handle dynamic traffic scenarios such as sudden accidents and holiday peaks, as well as complex road network spatial relationships. As a result, existing charging load forecasting technologies have the technical problem of inaccurate prediction when facing some extreme scenarios with characteristics of dynamic traffic scenarios and complex road network spatial relationships. Summary of the Invention

[0003] The present application provides a method, device, terminal and medium for predicting electric vehicle charging demand, which is used to solve the technical problem of inaccurate prediction existing in existing charging load forecasting technologies when facing some extreme scenarios with characteristics of dynamic traffic scenarios and complex road network spatial relationships.

[0004] To solve the above technical problem, in the first aspect of the present application, a method for predicting electric vehicle charging demand is provided, including:

[0005] Collect traffic charging network data and external factor data;

[0006] According to the traffic charging network data, a dynamic heat map is constructed by means of kernel density estimation, where the dynamic heat map is used to reflect the traffic flow status of each road network node;

[0007] Based on the time and space dimensions, feature alignment is performed on the dynamic heat map and the external factor data, and through multi-modal attention processing, cross-modal attention fusion is performed on the features of the dynamic heat map and the features of the external factor data to obtain a fusion vector;

[0008] Based on the fusion vector, combined with a preset spatio-temporal graph neural network model, the charging demand prediction value of each road network node is determined.

[0009] Preferably, the traffic charging network data includes: vehicle GPS trajectories, traffic flow data, charging pile status, and road network topology.

[0010] Preferably, the external factor data includes: weather data, electricity price data, and historical demand data.

[0011] Preferably, the determining the charging demand prediction value of each road network node based on the fusion vector and combined with a preset spatio-temporal graph neural network model includes:

[0012] Based on the fusion vector, combined with the dynamic adjacency matrix, the fusion vector is subjected to neighborhood aggregation processing through a spatial graph convolution unit to obtain the spatial dependence features between the road network nodes, where the dynamic critical matrix is used to reflect the similarity between different road network nodes;

[0013] Based on the spatial dependence features, time convolution processing is performed through a time gating unit to obtain the spatio-temporal dependence features of each road network node;

[0014] Determine the predicted charging demand value of each road network node according to the spatio-temporal dependence features.

[0015] Preferably, the calculation formula of the dynamic critical matrix is specifically:

[0016]

[0017] In the formula, The value of the dynamic adjacency matrix between road network node i and road network node j at time t, and are the heat map values of road network node i and road network node j at time t respectively, is the scaling coefficient of the heat value difference.

[0018] Preferably, the calculation formula for generating the dynamic heat map is specifically:

[0019]

[0020] In the formula, is the heat value of road network node i at time t, is the topological distance between road network node i and vehicle j, is the dynamic traffic bandwidth coefficient at time t, is the Gaussian kernel function, and n is the number of vehicles.

[0021] Preferably, after constructing the dynamic heat map, it further includes:

[0022] The dynamic heat map is optimized through a graph convolutional neural network.

[0023] Meanwhile, a second aspect of the present application provides an electric vehicle charging demand prediction device, including:

[0024] A data acquisition unit for acquiring traffic charging network data and external factor data;

[0025] A heat map construction unit for constructing a dynamic heat map according to the traffic charging network data by means of kernel density estimation, where the dynamic heat map is used to reflect the traffic flow status of each road network node;

[0026] A feature fusion unit, configured to perform feature alignment on the dynamic heat map and the external factor data based on the time and space dimensions, and through multi-modal attention processing, perform cross-modal attention fusion on the features of the dynamic heat map and the features of the external factor data to obtain a fusion vector;

[0027] A charging demand prediction unit, configured to determine the charging demand prediction values of each road network node based on the fusion vector and in combination with a preset spatio-temporal graph neural network model.

[0028] A third aspect of the present application provides an electric vehicle charging demand prediction terminal, including: a memory and a processor;

[0029] The memory is used to store program code, and the program code is used to implement an electric vehicle charging demand prediction method provided in the first aspect of the present application;

[0030] The processor is used to read and execute the program code.

[0031] A fourth aspect of the present application provides a computer-readable storage medium, in which program code is stored, and the program code is used to be read and executed by a processor to implement an electric vehicle charging demand prediction method provided in the first aspect of the present application.

[0032] It can be seen from the above technical solutions that the present application has the following advantages:

[0033] The solution provided by the present application constructs a dynamic heat map reflecting the traffic flow status of each road network node through kernel density estimation based on traffic charging network data, uses the dynamic heat map to capture traffic flow changes in real time, combines graph convolutional networks to depict road network topological constraints, and then based on the time and space dimensions, performs feature alignment on the dynamic heat map and external factor data, and through multi-modal attention processing, performs cross-modal attention fusion on the features of the dynamic heat map and the features of the external factor data, so as to effectively integrate the context information of heterogeneous data, obtain a fusion vector, and based on the fusion vector, in combination with a preset spatio-temporal graph neural network model, determine the charging demand prediction values of each road network node, overcoming the problem of spatio-temporal granularity differences in multi-source data, significantly improving the robustness of demand prediction in emergency and high-fluctuation scenarios, and further improving the accuracy of charging demand prediction in the face of extreme scenarios. Description of the Drawings

[0034] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0035] Figure 1 It is a schematic flowchart of an embodiment of a method for predicting the charging demand of electric vehicles provided by the present application.

[0036] Figure 2 It is an overall framework diagram of an embodiment of a method for predicting the charging demand of electric vehicles provided by the present application.

[0037] Figure 3 It is a schematic flowchart for refining step 104 of a method for predicting the charging demand of electric vehicles provided by the present application.

[0038] Figure 4 It is a logical framework diagram of a dynamic spatio-temporal graph neural network prediction model in a method for predicting the charging demand of electric vehicles provided by the present application.

[0039] Figure 5 It is a schematic structural diagram of an embodiment of a device for predicting the charging demand of electric vehicles provided by the present application.

[0040] Figure 6 It is a schematic structural diagram of an embodiment of a terminal for predicting the charging demand of electric vehicles provided by the present application. Detailed implementation manners

[0041] The embodiments of the present application provide a method, a device, a terminal, and a medium for predicting the charging demand of electric vehicles, which are used to solve the technical problem that the existing charging load prediction technology is inaccurate in predicting in the face of some extreme scenarios with characteristics of dynamic traffic scenarios and complex road network spatial relationships.

[0042] To make the invention purpose, features, and advantages of the present application more obvious and understandable, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the embodiments described below are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0043] First, a detailed description of an embodiment of a method for predicting the charging demand of electric vehicles provided by the present application is as follows:

[0044] Please refer to Figure 1 andFigure 2 , a method for predicting the charging demand of an electric vehicle provided by this application includes:

[0045] Step 101, collect traffic charging network data and external factor data;

[0046] It should be noted that first are dynamic or static data such as traffic charging network data and external factor data. Among them, traffic charging network data includes: vehicle GPS trajectories, traffic flow data, charging pile status, and road network topology. External factor data includes: weather data, electricity price data, and historical demand data. The complete data collection information is shown in Table 1:

[0047]

[0048] By Map-Matching, map the GPS trajectories to specific road network nodes, and more real vehicle distribution and road section occupancy can be obtained; charging piles and traffic cameras can provide near-real-time status data. Combining factors such as weather and electricity price can reflect the impact of the external environment on the charging demand. After this step, an initial multi-source spatio-temporal input matrix and an adjacency matrix can be obtained, where N represents the number of road network nodes and D represents the feature dimension of a single node.

[0049] Step 102, construct a dynamic heat map according to the traffic charging network data by means of kernel density estimation;

[0050] Among them, the dynamic heat map is used to reflect the traffic flow status of each road network node;

[0051] It should be noted that based on the data obtained in the previous step, generate a dynamic heat map on the basis of these data to initially depict the traffic flow, congestion degree, and potential charging demand distribution of road network nodes. The dynamic heat map of this embodiment can be generated by a heat map generation algorithm based on adaptive kernel density estimation (AKDE).

[0052]

[0053] In the formula, is the heat value of node i at time t; is the topological distance between node i and vehicle j. In this embodiment, the topological distance is adopted to be different from the simple Euclidean distance, which is more in line with the constraints of the actual road network: if there are multiple roads connecting two nodes and the road section grade is high, the distance may be small; if detours are required or the congestion is serious, the distance increases to reflect the real traffic flow path and traffic conditions; is the dynamic bandwidth coefficient, and its value can be adjusted flexibly according to the actual traffic flow changes. During peak hours, the traffic flow variance is large, and the bandwidth can be appropriately increased to improve the smoothness of the kernel function. During off-peak hours, the bandwidth is reduced to focus on local hotspots and highlight local differences; is the Gaussian kernel function, which is used to map distance into density contribution.

[0054] Step 103: Based on the time and space dimensions, perform feature alignment on the dynamic heat map and external factor data. Through multi-modal attention processing, perform cross-modal attention fusion on the features of the dynamic heat map and the external factor data to obtain a fusion vector;

[0055] It should be noted that the dynamic heat map obtained in the previous step is aligned with the external factor data in the time and space dimensions; and the most critical external factors are highlighted through the attention mechanism to enrich the input information of the prediction model. After obtaining the optimized heat map, the spatio-temporal feature extraction and alignment strategy is shown in Table 2:

[0056]

[0057] After resampling to the same time step, the model can be processed in parallel ; The adoption of environmental feature normalization (such as electricity price) is to prevent the digital magnitude from being too large or too small from affecting the training stability.

[0058] After that, a multi-modal attention mechanism is introduced to highlight the relative importance of different modalities at different moments, and cross-modal attention fusion is performed on the dynamic heat map features (road traffic information) and external factor data such as weather, electricity price, and historical demand, so as to dynamically highlight the most relevant features and suppress redundant information at the model input stage. Specifically as follows:

[0059] 1. Multi-modal input: For node i at time t, through the processing of the previous steps, several modal feature vectors unified to the same time step and node index have been obtained. Among them, the representations of some modal feature vectors involved in this embodiment are as follows:

[0060]

[0061] Among them, is the traffic flow feature of node i at time t; is the weather feature of node i at time t, such as temperature, humidity, etc.; represents the electricity price information of node i at time t; represents the historical demand feature of node i at time t, indicating the time series information of past demands.

[0062] 2. Mapping to the Q / K / V Space: To enable the attention mechanism to recognize the importance of each modality, we use trainable linear transformations for different modalities respectively. The specific calculation formula is as follows:

[0063]

[0064] Among them, m represents the modality type (such as traffic, weather, price, etc.), are all trainable linear transformation parameters of modality m and are trainable parameters; is the feature vector of the m-th modality of node i at time t.

[0065] 3. Attention Calculation: Combine the Q, K, and V of different modalities to obtain their weighted output. The weighted calculation formula is as follows:

[0066]

[0067] Based on this weighted output, the modal features with the greatest influence at this time / node are given higher weights. Among them, Q, K, and V are the query, key, and value matrices from different modalities respectively, representing the information of different modalities; is the feature dimension scaling factor, which is used to stabilize the size of the inner product and avoid too large or too small gradients; is the standard softmax function, which is used to normalize the output.

[0068] Next, the above weighted output is concatenated to obtain , which represents the fused multi-modal feature vector of node i at time t. Summarize all N nodes to obtain , where is the dimension obtained after multi-modal fusion.

[0069] Step 104: Based on the fused vector, combined with the preset spatio-temporal graph neural network model, determine the predicted charging demand values of each road network node.

[0070] It should be noted that based on the fused vector obtained in step 103, a dynamic spatio-temporal neural network model is used to predict the charging load demand to determine the predicted charging demand values of each road network node, so as to reasonably dispatch power resources according to the predicted charging demand results.

[0071] More specifically, as Figure 3 shown, step 104 provided in this embodiment specifically includes:

[0072] Step 1041: Based on the fused vector, combined with the dynamic adjacency matrix, perform neighborhood aggregation processing on the fused vector through the spatial graph convolution unit to obtain the spatial dependence features between road network nodes;

[0073] Among them, the dynamic critical matrix is used to reflect the similarity between different road network nodes;

[0074] Step 1042: Based on the spatial dependence features, perform temporal convolution processing through a temporal gating unit to obtain the spatio-temporal dependence features of each road network node;

[0075] Step 1043: Determine the predicted charging demand value of each road network node according to the spatio-temporal dependence features.

[0076] It should be noted that, as Figure 4 described above, first, based on the fusion vector obtained in Step 103, combined with the dynamic adjacency matrix constructed based on the dynamic heat map, perform neighborhood aggregation processing on the fusion vector through a spatial graph convolution unit to obtain the spatial dependence features between road network nodes. Among them, the construction expression of the dynamic adjacency matrix in this embodiment is specifically:

[0077]

[0078] In the formula, The value of the dynamic adjacency matrix between road network node i and road network node j at time t, and are the heat map values of road network node i and road network node j at time t respectively, is the scaling coefficient of the heat value difference, which is used to control the influence of the heat difference between nodes on the connection weight. It changes in the time dimension, allowing the association between nodes to be dynamically adjusted according to the heat difference of the traffic flow.

[0079] After completing the dynamic adjacency matrix, use spatial graph convolution and utilize to perform neighborhood aggregation and extract local spatial dependence relationships. The specific operation expression is as follows:

[0080]

[0081] Among them, is the dynamic adjacency matrix at time t, representing the spatial dependence between nodes, is the fusion vector matrix at time t, is the graph convolution operation, which is used to aggregate the information of adjacent nodes.

[0082] After completing the spatial graph convolution, perform temporal gating processing. The specific operation expression is as follows:

[0083]

[0084] Among them, is the result of the spatial convolution at time t, that is, the spatial dependence feature, is the weight of the convolution kernel, is the expansion coefficient, K is the convolution kernel size, which indicates how many time steps of data need to be integrated in each convolution operation. This operation can capture long sequence dependencies and take into account parallel computing efficiency.

[0085] Then, through a layer of perceptron mapping, we get the The charging demand forecast value at the moment:

[0086]

[0087] in, is the predicted charging demand, indicating the time The corresponding demand forecast value of each node; is the result of time gating processing at time t, that is, the spatiotemporal dependency feature, is a mapping function used to map the spatiotemporal features to the final predicted values, wherein the mapping relationship between the spatiotemporal features and the predicted values ​​can be specifically determined through model training.

[0088] Among them, the training examples of graph convolutional networks and time-gated networks are as follows:

[0089] a. In the pre-training GCN part, the heat map is first fully combined with the road network information.

[0090] b. In the end-to-end fine-tuning stage, multimodal attention, dynamic adjacency, and TCN are trained together to open up the overall data flow.

[0091] Through the preset loss function, the first GCN pre-training can ensure that the road network information is reasonably captured; the final loss takes into account both numerical accuracy and spatial coherence, making the prediction results more accurate and reasonable;

[0092] The expression of the loss function is as follows:

[0093]

[0094] in, represents the predicted charging demand vector output by the model, is the actual observed charging demand vector, MAE is the mean absolute error, which measures the prediction accuracy; Spatial-Smoothness is a spatial smoothness constraint, which encourages the prediction values ​​between adjacent nodes not to jump suddenly; α and β are the weights for controlling different objectives.

[0095] The real-time update mechanism can adopt the edge-cloud collaborative architecture and deploy the pruned or quantized DSTGNN model at the edge (RSU) to reduce the number of parameters and computing overhead; when encountering a sudden change in the heat map ( ) timely alarm and local short-term forecasting, where Let \(\theta\) be the mutation threshold of the heatmap. If the vehicle distribution or demand exceeds the preset change amplitude within a short period of time, it is regarded as an abnormal or sudden situation, triggering the edge response strategy. In the cloud, the monitoring results and local features from the edge are received, and incremental learning is performed to update the weights of the main model, and the latest weights are sent to the edge to complete synchronization. Through this collaborative mode, on the one hand, the edge realizes minute-level rapid prediction and local early warning, and on the other hand, the cloud performs batch or incremental training on a larger scale of historical data, enabling the model to continuously adapt to the evolution of traffic and demand distribution.

[0096]

[0097] Where is the updated model weight; is the current model weight; is the learning rate, which controls the step size of weight update; is the gradient of the loss function, which is used to guide how the model adjusts the weights; is the mutation region data from the edge, which serves as the input for incremental learning.

[0098] Through the above steps, the patent solution successfully combines the dynamic heatmap with the graph neural network, realizes the deep integration of vehicle distribution and road network topology, finely depicts the dynamic evolution of charging demand in space; and overcomes the problem of spatio-temporal granularity differences in multi-source data through temporal sequence technologies such as multi-modal attention and TCN, significantly improving the robustness of demand prediction in emergency and high-fluctuation scenarios; finally, cloud-edge collaboration and lightweight models are adopted to achieve real-time monitoring and emergency response of charging demand, effectively shortening the closed-loop cycle of perception-decision-update, and the entire system architecture is as Figure 2 shown. In practical applications, this solution can provide minute-level charging demand prediction and dynamic hot spot identification functions, helping the power grid and operators to more efficiently layout charging piles and balance the power load.

[0099] Furthermore, the solution provided in this embodiment may further include the following steps after step 102 and before step 103:

[0100] Step 1021: Optimize the dynamic heatmap through a graph convolutional neural network.

[0101] It should be noted that after obtaining the dynamic heatmap through step 102, the initial heatmap can also be optimized through a Graph Convolutional Network (GCN). As shown in the following expression of the optimization process, GCN can optimize the accuracy and adaptability of the heatmap by mining the complex topological relationships in spatial data. The expression of the optimization process is as follows:

[0102] .

[0103] Among them, is the set or matrix of heat values of the dynamic heat map before optimization at time t, is the set or matrix of heat values of the dynamic heat map after optimization at time t. L is the road network adjacency matrix, which represents the actual road connections and weights between nodes (such as highways / ordinary roads). The higher the value, the closer the association between nodes; is a graph convolution operation used to fuse the spatial relationships between nodes, and generally can be set to 3 layers.

[0104] It can be understood that if the method of this embodiment includes step 1021, the subsequent steps related to the heat value steps, all use the optimized to replace the original , for example, the expression of the dynamic adjacency matrix mentioned in step 1041 can be correspondingly adjusted to:

[0105]

[0106] The above is a detailed description of an embodiment of a method for predicting electric vehicle charging demand provided by this application. The following is a detailed description of an embodiment of an apparatus for predicting electric vehicle charging demand provided by this application.

[0107] Please refer to Figure 5 , an embodiment of an apparatus for predicting electric vehicle charging demand provided by this application includes:

[0108] A data acquisition unit 201, configured to acquire traffic charging network data and external factor data;

[0109] A heat map construction unit 202, configured to construct a dynamic heat map based on the traffic charging network data through kernel density estimation, where the dynamic heat map is used to reflect the traffic flow status of each road network node;

[0110] A feature fusion unit 203, configured to perform feature alignment on the dynamic heat map and external factor data based on the time and space dimensions, and perform cross-modal attention fusion on the features of the dynamic heat map and the external factor data through multi-modal attention processing to obtain a fusion vector;

[0111] A charging demand prediction unit 204, configured to determine the charging demand prediction value of each road network node based on the fusion vector in combination with a preset spatio-temporal graph neural network model.

[0112] In addition, this application also provides a detailed description of related embodiments such as an electric vehicle charging demand prediction terminal and a computer-readable storage medium.

[0113] AsFigure 6 As shown, an embodiment of an electric vehicle charging demand prediction terminal provided by the present application, its implementation types include but are not limited to: personal computers, industrial computers, servers, and embedded intelligent devices. The main components of the terminal include: a memory 33 and a processor 31, and the memory 33 and the processor 31 can be connected through a communication bus 34;

[0114] The memory 33 is used to store program codes, and the program codes are used to implement an electric vehicle charging demand prediction method as in the above embodiment;

[0115] The processor 31 is used to read and execute the program codes.

[0116] The fourth aspect of the present application provides a computer-readable storage medium, in which program codes are stored, and the program codes are used to be read and executed by a processor to implement an electric vehicle charging demand prediction method as in the above embodiment.

[0117] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described terminal, device, and unit can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0118] In several embodiments provided by the present application, it should be understood that the disclosed terminal, device, and method can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0119] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances, so that the embodiments of the present application described here, for example, can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0120] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Here, A and B can be singular or plural. The character " / " generally indicates an "or" relationship between the associated objects before and after. "At least one (item) of the following" or its similar expressions refer to any combination of these items, including any combination of single item (item) or plural items (items). For example, at least one (item) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0121] The unit described as a separate component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0122] In addition, each functional unit in various embodiments of the present invention can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0123] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs, etc., which can store program codes.

[0124] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for predicting the charging demand of an electric vehicle, characterized in that, Including: Collect traffic charging network data and external factor data; According to the traffic charging network data, construct a dynamic heat map by means of kernel density estimation, wherein the dynamic heat map is used to reflect the traffic flow status of each road network node; Based on the time and space dimensions, perform feature alignment on the dynamic heat map and the external factor data, and through multi-modal attention processing, perform cross-modal attention fusion on the features of the dynamic heat map and the external factor data to obtain a fusion vector; Based on the fusion vector, combined with a preset spatio-temporal graph neural network model, determine the predicted charging demand value of each road network node.

2. The electric vehicle charging demand prediction method according to claim 1, characterized in that The traffic charging network data includes: vehicle GPS trajectories, traffic flow data, charging pile status, and road network topology.

3. A method for predicting the charging demand of an electric vehicle according to claim 1, characterized in that, The external factor data includes: weather data, electricity price data, and historical demand data.

4. A method for predicting the charging demand of an electric vehicle according to claim 1, characterized in that, The determining the predicted charging demand value of each road network node based on the fusion vector and combined with a preset spatio-temporal graph neural network model includes: Based on the fusion vector, combined with a dynamic adjacency matrix, perform neighborhood aggregation processing on the fusion vector through a spatial graph convolution unit to obtain the spatial dependence features between the road network nodes, wherein the dynamic critical matrix is used to reflect the similarity between different road network nodes; Based on the spatial dependence features, perform time convolution processing through a time gating unit to obtain the spatio-temporal dependence features of each road network node; Determine the predicted charging demand value of each road network node according to the spatio-temporal dependence features.

5. A method for predicting the charging demand of an electric vehicle according to claim 4, characterized in that, The specific calculation formula of the dynamic critical matrix is: In the formula, The dynamic adjacency matrix value between road network node i and road network node j at time t, and Are respectively the heat map values of road network node i and road network node j at time t, Is the scaling coefficient of the heat value difference.

6. The electric vehicle charging demand prediction method according to claim 1, wherein The specific calculation formula for generating the dynamic heat map is: Wherein, is the heat value of road network node i at time t, is the topological distance between road network node i and vehicle j, is the dynamic traffic bandwidth coefficient at time t, is the Gaussian kernel function, and n is the number of vehicles.

7. A method for predicting the charging demand of an electric vehicle according to claim 1, wherein After constructing the dynamic heat map, it further includes: Optimize the dynamic heat map through a graph convolutional neural network.

8. An electric vehicle charging demand prediction device, characterized in that, Including: A data acquisition unit for collecting traffic charging network data and external factor data; A heat map construction unit for constructing a dynamic heat map according to the traffic charging network data by means of kernel density estimation, wherein the dynamic heat map is used to reflect the traffic flow status of each road network node; A feature fusion unit for performing feature alignment on the dynamic heat map and the external factor data based on the time and space dimensions, and through multi-modal attention processing, performing cross-modal attention fusion on the features of the dynamic heat map and the external factor data to obtain a fusion vector; A charging demand prediction unit for determining the predicted charging demand value of each road network node based on the fusion vector and combined with a preset spatio-temporal graph neural network model.

9. An electric vehicle charging demand prediction terminal, characterized in that, Including: A memory and a processor; The memory is used to store program codes, and the program codes are used to implement an electric vehicle charging demand prediction method according to any one of claims 1 to 7; The processor is used to read and execute the program codes.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores program codes, and the program codes are used to be read and executed by the processor to implement an electric vehicle charging demand prediction method according to any one of claims 1 to 7.

Images