Vehicle speed prediction method and device, electronic equipment and medium
The multi-layer graph structure is constructed through graph convolution network, and the spatial information of vehicle speed data is extracted, which solves the problem that existing vehicle speed prediction models ignore spatial correlation, improves vehicle speed prediction accuracy, and provides an improvement for the energy management strategy of extended-range electric vehicles. Accurate design foundation.
Patent Information
- Application Number
- CN202411278219.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-12
- Publication Date
- 2024-12-27
AI Technical Summary
The existing vehicle speed prediction model mainly relies on long-term memory networks and gated cycle units, ignoring the spatial correlation between vehicle speed data, resulting in limited prediction accuracy and affecting the energy management strategy design of extended-range electric vehicles.
A graph convolution network is used to build a multi-layer graph structure. Through graph convolution operations with different times and node connections, the spatial information of vehicle speed data is extracted, thereby improving the accuracy of vehicle speed prediction. The specific steps include building multiple graph structures, performing graph convolution operations, extracting key features, and combining similarity matrix and features connected by different numbers of nodes to predict vehicle speed.
By considering the spatial correlation of vehicle speed data, the accuracy of vehicle speed prediction is improved, providing a more accurate basis for the predictive energy management strategy of extended-range electric vehicles.
Smart Images
Figure CN119190048A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of vehicles, and in particular to a vehicle speed prediction method, device, electronic equipment and medium. Background Art
[0002] In the field of transportation, the automotive power system is accelerating its transformation to electrification, but pure electric vehicles still face problems such as limited mileage, inconvenient charging, and battery safety. Extended-range electric vehicles that convert fuel to electricity and are fully electric have emerged. The energy management strategy is the core control algorithm of extended-range electric vehicles, and the quality of the algorithm directly affects the performance of the entire vehicle. Research on energy management strategies for extended-range electric vehicles, especially predictive energy management strategies based on instantaneous optimization algorithms, is of great significance to the performance improvement and optimization of extended-range electric vehicles.
[0003] Speed prediction is an important component of predictive energy management strategy. After speed prediction, the vehicle driving state within the predicted sight distance is equivalent to known. It is input into the control strategy optimization module and optimized using the global optimization algorithm or instantaneous optimization algorithm to obtain the optimal solution within the predicted sight distance. The control instruction selection module selects the first item of the optimized solution and outputs it to the controlled object for control. The remaining items are discarded. At this point, the optimization control of the current prediction step is completed. The subsequent control is advanced by continuously updating the time domain and repeating the above steps. However, most of the existing speed prediction models use long short-term memory networks (LSTMs) and gated recurrent units (GRU) to extract the time dependency of historical speed data, but ignore the influence of spatial correlation between speed data, resulting in limited accuracy of the speed prediction model. Therefore, designing an accurate speed prediction model is of great significance to the design of predictive energy management strategy for extended-range vehicles. Summary of the invention
[0004] The object of the present invention is to provide a vehicle speed prediction method, device, electronic equipment and medium to improve the vehicle speed prediction accuracy.
[0005] In order to achieve the above object, the present invention adopts the following technical solutions:
[0006] In a first aspect, the present invention provides a vehicle speed prediction method, comprising:
[0007] According to the vehicle speed data of each time period, a first graph structure of each time period is constructed;
[0008] According to the first graph structure, a second graph structure is constructed; the second graph structure includes key features generated by graph convolutions of different times; a graph convolution operation is performed on the second graph structure to obtain a first output feature;
[0009] According to the first graph structure, a third graph structure is constructed; the third graph structure includes key features generated by connecting different numbers of nodes; a graph convolution operation is performed on the third graph structure to obtain a second output feature;
[0010] The vehicle speed is predicted based on the first output feature and the second output feature.
[0011] As a further preferred technical solution, the first graph structure of each time period is constructed according to the vehicle speed data of each time period, including:
[0012] The vehicle speed data of each time period is extracted using a multi-layer perceptron to obtain initial features;
[0013] A first graph structure is constructed according to the initial features.
[0014] As a further preferred technical solution, constructing a second graph structure according to the first graph structure includes:
[0015] Extracting key nodes according to the first graph structure;
[0016] Performing at least three graph convolution operations on the first graph structure, and extracting a first feature of the key node from an output feature of each graph convolution operation;
[0017] A second graph structure is constructed according to the first feature of the key node.
[0018] As a further preferred technical solution, constructing a second graph structure according to the first feature of the key node includes:
[0019] Calculating the similarity of the first feature of the key node to obtain a first similarity adjacency matrix;
[0020] A second graph structure is constructed according to the first feature of the key node and the first similarity adjacency matrix.
[0021] As a further preferred technical solution, constructing a third graph structure according to the first graph structure includes:
[0022] Extracting key nodes according to the first graph structure;
[0023] Determining an updated first graph structure according to an adjacency matrix of the first graph structure and a set scale; the set scale is used to characterize the amount of aggregated information;
[0024] Performing a graph convolution operation on the updated first graph structure to extract a second feature of the key node in the output feature;
[0025] A third graph structure is constructed according to the second feature of the key node.
[0026] As a further preferred technical solution, constructing a third graph structure according to the second feature of the key node includes:
[0027] Calculating the similarity of the second feature of the key node to obtain a second similarity adjacency matrix;
[0028] A third graph structure is constructed according to the second feature of the key node and the second similarity adjacency matrix.
[0029] As a further preferred technical solution, predicting the vehicle speed according to the first output feature and the second output feature includes:
[0030] Determine a historical vehicle speed feature according to the first output feature and the second output feature; the historical vehicle speed feature is used to represent historical vehicle speed information;
[0031] The vehicle speed is predicted according to the historical vehicle speed characteristics.
[0032] In a second aspect, the present invention provides a vehicle speed prediction device, comprising:
[0033] A first graph structure building module, used to build a first graph structure for each time period according to the vehicle speed data for each time period;
[0034] A first output module is used to construct a second graph structure according to the first graph structure; the second graph structure includes key features generated by graph convolutions of different times; and perform a graph convolution operation on the second graph structure to obtain a first output feature;
[0035] A second output module is used to construct a third graph structure according to the first graph structure; the third graph structure includes key features generated by connecting different numbers of nodes; and perform a graph convolution operation on the third graph structure to obtain a second output feature;
[0036] A prediction module is used to predict the vehicle speed according to the first output feature and the second output feature.
[0037] In a third aspect, the present invention provides an electronic device, comprising:
[0038] at least one processor, and a memory communicatively coupled to at least one of the processors;
[0039] The memory stores instructions that can be executed by at least one of the processors, and the instructions are executed by at least one of the processors so that at least one of the processors can perform the above method.
[0040] In a fourth aspect, the present invention provides a computer-readable storage medium, on which computer instructions are stored, and the computer instructions are used to enable the computer to execute the above method.
[0041] Compared with the prior art, the present invention has the following beneficial effects:
[0042] The vehicle speed prediction method provided by the present invention constructs a first graph structure for each time period according to the vehicle speed data of each time period; constructs a second graph structure according to the first graph structure; the second graph structure includes key features generated by different times of graph convolution; performs graph convolution operation on the second graph structure to obtain a first output feature; constructs a third graph structure according to the first graph structure; the third graph structure includes key features generated by different numbers of node connections; performs graph convolution operation on the third graph structure to obtain a second output feature; predicts the vehicle speed according to the first output feature and the second output feature. The second graph structure constructed in the method includes key features generated by different times of graph convolution, taking into account the feature changes of key nodes aggregating different jump ranges in the graph convolution recursive operation; the constructed third graph structure includes key features generated by different numbers of node connections, taking into account the feature changes brought about by the key nodes being connected to different numbers of nodes, and extracts the spatial information of the vehicle speed information in each time period through two modes, thereby improving the accuracy of vehicle speed prediction, and laying a foundation for the design of subsequent predictive energy management strategies for extended-range vehicles or other types of new energy vehicles. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0044] Figure 1 It is a schematic flow chart of the vehicle speed prediction method provided by the present invention;
[0045] Figure 2 is a schematic diagram of a first graph structure construction process provided by the present invention;
[0046] Figure 3 Schematic diagram of the graph convolution process with different jump ranges provided by the present invention;
[0047] Figure 4 It is a schematic diagram of the feature aggregation process of key nodes in different hop ranges provided by the present invention;
[0048] Figure 5 It is a structural schematic diagram of the vehicle speed prediction device provided by the present invention;
[0049] Figure 6 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION
[0050] The following is a description of exemplary embodiments of the present application in conjunction with the accompanying drawings, including various details of the embodiments of the present application to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.
[0051] Example 1
[0052] Figure 1 It is a flow chart of a vehicle speed prediction method provided in this embodiment. The method can be executed by a vehicle speed prediction device. The device can be composed of software and / or hardware and is generally integrated in an electronic device. The electronic device can be a vehicle controller, etc. For ease of understanding, each step in the control method of this embodiment is executed by the vehicle controller.
[0053] like Figure 1 As shown, this embodiment provides a vehicle speed prediction method, comprising the following steps:
[0054] S110: Construct a first graph structure for each time period according to the vehicle speed data for each time period.
[0055] Optionally, before constructing the first graph structure for each time period based on the vehicle speed data for each time period, the following steps are also included: acquiring vehicle speed data under different working conditions, dividing the vehicle speed data into time periods, and obtaining vehicle speed data for each time period.
[0056] Optionally, constructing the first graph structure of each time period according to the vehicle speed data of each time period includes:
[0057] The vehicle speed data of each time period is extracted using a multi-layer perceptron to obtain initial features;
[0058] A first graph structure is constructed according to the initial features.
[0059] The input of the graph convolutional network mainly includes the graph structure and the corresponding features. For the graph structure, we use G = (V, A, X), where V represents the set of nodes in the graph, A represents the adjacency matrix in the graph, and X is the feature corresponding to the node. In this embodiment, the speed data of each time period (i.e., the historical speed) is first constructed by graph construction. For example, the speed data X with an input of 30s is divided into sections according to T = 10s, and three sections of historical speed data with a time length of 10 seconds are obtained, which are represented by t1 = X1, t2 = X2, and t3 = X3. t1 is the speed data of 0-10s, t2 is the speed data of 10-20s, and t3 is the speed data of 20-30s. For the convenience of expression, the subsequent steps and formulas involved in this embodiment are all explained by taking the above three time periods as examples. Then, the speed data in each time period is graph constructed, and features are extracted by multi-layer perceptron (MLP) respectively, H i =MLP(X i ), where i = 1, 2, 3, X i Represents the vehicle speed data in three time periods respectively, H i is the feature extracted by MLP in the corresponding time period. Then each extracted feature matrix is multiplied by its transposed matrix to obtain the adjacency matrix corresponding to the vehicle speed data in each time period. The node features are X1, X2, and X3. Thus, the graph structures corresponding to the three time periods can be obtained: the graph structure of the time period t1: G1 = (V1, A1, H1), the graph structure of the time period t2: G2 = (V2, A2, H2), and the graph structure of the time period t3: G3 = (V3, A3, H3). The schematic diagram of the first graph structure construction process is as follows Figure 2 shown.
[0060] S120. Construct a second graph structure according to the first graph structure; the second graph structure includes key features generated by graph convolutions of different orders; perform a graph convolution operation on the second graph structure to obtain a first output feature.
[0061] This step can be completed based on graph convolution modules with different hop ranges.
[0062] Optionally, constructing a second graph structure according to the first graph structure includes:
[0063] Extracting key nodes according to the first graph structure;
[0064] Performing at least three graph convolution operations on the first graph structure, and extracting a first feature of the key node from an output feature of each graph convolution operation;
[0065] A second graph structure is constructed according to the first feature of the key node.
[0066] Optionally, constructing a second graph structure according to the first feature of the key node includes:
[0067] Calculating the similarity of the first feature of the key node to obtain a first similarity adjacency matrix;
[0068] A second graph structure is constructed according to the first feature of the key node and the first similarity adjacency matrix.
[0069] Specifically, this step extracts key nodes from the adjacency matrix in the first graph structure. Each row of the adjacency matrix represents a node, and each column in this row represents the connection weight between this node and other nodes. The larger the connection weight constructed by the above features, the closer the relationship between this node and other nodes. Therefore, all columns corresponding to each row are summed up and then sorted to obtain a node index with the largest sum. Since the weight value of the connection between this node and other nodes is the largest, this node is regarded as a key node. Figure 2 For the graph structure in , the three adjacency matrices can obtain the indexes of the three key nodes, which are recorded as id1, id2, and id3 respectively.
[0070] The second graph structure constructed in this embodiment takes into account the feature changes of key nodes in different jump ranges in the graph convolution recursive operation. At least three graph convolutions are performed on each first graph structure (such as G1, G2, and G3). Each graph convolution operation can aggregate the node features on the corresponding jump. This embodiment will be described later using three graph convolutions as an example. Figure 3 As shown in the figure, the first graph convolution can aggregate the node features on 1 hop, the second can aggregate the features on 2 hop nodes, and the third can aggregate the features on 3 hop nodes. The calculation process is as follows:
[0071]
[0072] Among them, i=1,2,3, corresponding to the three time periods t1, t2, and t3 respectively; l=1,2,3, representing 1, 2, and 3 graph convolutions respectively; Represents the output features after the first graph convolution. For example, the output features of the first graph convolution are The output feature of the second graph convolution is The output feature of the third graph convolution is ReLU is the activation function, For the adjacency matrix A i Perform symmetric normalization operation, is the input feature calculated by the first layer of convolution, W i l It is the weight matrix of the first graph convolution in each time period.
[0073] After each graph convolution, the above key nodes will have features within different hop ranges. Therefore, the features of the key nodes at the positions of the output features of each graph convolution in each time period are extracted. In this way, three key features (i.e., the first features of the key nodes) generated by different numbers of graph convolutions are obtained in each time period, which are recorded as: The first feature of the key node in the t1 period is recorded as: The first feature of the key node in the t2 time period is recorded as: The first feature of the key node in the t3 period is recorded as: The three key features in each time period can be regarded as three node features. Then, the similarity of the three key features in each time period is calculated according to the cosine similarity, and the first similarity adjacency matrix A can be obtained. ik , where i = 1, 2, and 3 correspond to three time periods. In this way, the key features of different graph convolution times (i.e., the first features of key nodes) are combined into a new graph structure (i.e., the second graph structure). Therefore, the new graph structure corresponding to the three time periods can be written as follows: Graph structure for time period t1: The structure of the graph in time period t2: The structure of the graph for time period t3: Then, the second graph structure is aggregated by graph convolution, which can better mine the similarities in graph convolution recursive propagation and aggregate information at different levels. The aggregation process of key node features in different hop ranges is as follows: Figure 4 The aggregation calculation process is as follows:
[0074]
[0075] Where i=1,2,3 corresponds to three time periods, H ik Contains H 1k , H 2k , H 3k , respectively represent the features corresponding to t1, t2, and t3, W i is the corresponding weight matrix.
[0076] S130. Construct a third graph structure according to the first graph structure; the third graph structure includes key features generated by connecting different numbers of nodes; and perform a graph convolution operation on the third graph structure to obtain a second output feature.
[0077] This step can be completed based on graph convolution modules with different numbers of nodes.
[0078] Optionally, constructing a third graph structure according to the first graph structure includes:
[0079] Extracting key nodes according to the first graph structure;
[0080] Determining an updated first graph structure according to an adjacency matrix of the first graph structure and a set scale; the set scale is used to characterize the amount of aggregated information;
[0081] Performing a graph convolution operation on the updated first graph structure to extract a second feature of the key node in the output feature;
[0082] A third graph structure is constructed according to the second feature of the key node.
[0083] Optionally, constructing a third graph structure according to the second feature of the key node includes:
[0084] Calculating the similarity of the second feature of the key node to obtain a second similarity adjacency matrix;
[0085] A third graph structure is constructed according to the second feature of the key node and the second similarity adjacency matrix.
[0086] The above second graph structure only considers the feature changes of different jump ranges of key node aggregation in the graph convolution recursive operation, but ignores the feature changes brought by the key node connecting with different numbers of nodes. Therefore, the third graph structure constructed in step S130 considers the feature changes brought by the key node connecting with different numbers of nodes: first, the adjacency matrix in the above t1 time period graph structure: G1 = (V1, A1, H1), t2 time period graph structure: G2 = (V2, A2, H2), t3 time period graph structure: G3 = (V3, A3, H3) is sorted from the smallest to the largest according to the max-k sorting mechanism, and A is sorted. i The first P values of each row in are set to 1 (1 represents connection), and the rest are set to 0 (0 represents no connection). By taking different values of P, we can get adjacency matrices with different numbers of node connections. The values of P are 3, 6, and 9. Therefore, three adjacency matrices will be generated in each time period, which can be expressed as A iP , where P is 3, 6, and 9. Therefore, the graph structure of each time period can be rewritten as follows: t1 time period graph structure: G1 = (V1, A 1P , H1), t2 time period graph structure: G2 = (V2, A 2P , H2), t3 time period graph structure: G3 = (V3, A 3P , H3), the above rewritten graph structure is the updated first graph structure, and P is the set scale. Then the graph convolution aggregation operation is performed, and the calculation process is as follows:
[0087]
[0088] Among them, H iPwhere i = 1, 2, 3, P = 3, 6, 9. The explanation is as follows: time t1 corresponds to [H 13 ,H 16 ,H 19 ], t2 corresponds to [H 23 ,H 26 ,H 29 ], t3 corresponds to [H 33 ,H 36 ,H 39 ].
[0089] Then the above key nodes id1, id2, and id3 will have features after aggregation of different numbers of nodes, so from H iP The features of the key nodes are extracted from the , so that three key features with different numbers of node connections (i.e., the second features of the key nodes) are obtained in each time period, denoted as: H iPk , the second feature of the key node in the time period t1 is recorded as: [H 13k ,H 16k ,H 19k ], the second feature of the key node in the time period t2 is recorded as: [H 23k ,H 26k ,H 29k ], the second feature of the key node in the t3 time period is recorded as: [H 33k ,H 36k ,H 39k ]. Then, similar to S120, the three key features in each time period can be regarded as three node features, and then the similarity of the three key features in each time period is calculated according to the cosine similarity, and the similarity adjacency matrix A can be obtained. iPk , where i = 1, 2, 3 correspond to three time periods. In this way, the key features generated by different numbers of node connections (i.e., the second features of the key nodes) are combined into a new graph structure (i.e., the third graph structure). Therefore, the new graph structure corresponding to the three time periods can be written as: Graph structure for time period t1: G 1Pk =(V 1Pk , A 1Pk , H 1Pk ), t2 time period graph structure: G 2Pk =(V 2Pk , A 2Pk , H 2Pk ), t3 time period graph structure: G 3Pk =(V 3Pk , A 3Pk , H 3Pk ). Then the third graph structure is subjected to graph convolution aggregation, which can better mine the similarities in the propagation of different numbers of nodes. In each time period, an aggregated feature is obtained, and H z1, H z2 , H z3 Represent the features corresponding to t1, t2, and t3 respectively.
[0090] S140: Predict a vehicle speed according to the first output characteristic and the second output characteristic.
[0091] Optionally, predicting the vehicle speed according to the first output feature and the second output feature includes:
[0092] Determine a historical vehicle speed feature according to the first output feature and the second output feature; the historical vehicle speed feature is used to represent historical vehicle speed information;
[0093] The vehicle speed is predicted according to the historical vehicle speed characteristics.
[0094] Optionally, the historical vehicle speed feature H out Calculate using the following formula: H out =(H 1k +H z1 )||(H 2k +H z2 )||(H 3k +H z3 ).
[0095] Optionally, the historical vehicle speed features are input into a fully connected layer to predict the vehicle speed.
[0096] The above-mentioned vehicle speed prediction method constructs a first graph structure for each time period according to the vehicle speed data of each time period; constructs a second graph structure according to the first graph structure; the second graph structure includes key features generated by different times of graph convolution; performs graph convolution operation on the second graph structure to obtain a first output feature; constructs a third graph structure according to the first graph structure; the third graph structure includes key features generated by different numbers of node connections; performs graph convolution operation on the third graph structure to obtain a second output feature; predicts vehicle speed according to the first output feature and the second output feature. The second graph structure constructed in the method includes key features generated by different times of graph convolution, taking into account the feature changes of different jump ranges of key node aggregation in the graph convolution recursive operation; the constructed third graph structure includes key features generated by different numbers of node connections, taking into account the feature changes brought about by the key nodes being connected to different numbers of nodes, and extracts spatial information of vehicle speed information in each time period through two modes, thereby improving the accuracy of vehicle speed prediction, and laying a foundation for the design of subsequent predictive energy management strategies for extended-range vehicles or other types of new energy vehicles.
[0097] Example 2
[0098] like Figure 5 As shown, this embodiment provides a vehicle speed prediction device, including:
[0099] A first graph structure building module 201, used to build a first graph structure for each time period according to the vehicle speed data for each time period;
[0100] A first output module 202 is used to construct a second graph structure according to the first graph structure; the second graph structure includes key features generated by graph convolutions of different times; and perform a graph convolution operation on the second graph structure to obtain a first output feature;
[0101] A second output module 203 is used to construct a third graph structure according to the first graph structure; the third graph structure includes key features generated by connecting different numbers of nodes; and perform a graph convolution operation on the third graph structure to obtain a second output feature;
[0102] The prediction module 204 is used to predict the vehicle speed according to the first output feature and the second output feature.
[0103] The device is used to execute the above method, and thus has at least functional modules and beneficial effects corresponding to the above method.
[0104] Example 3
[0105] like Figure 6 As shown, this embodiment provides an electronic device, including:
[0106] at least one processor; and
[0107] a memory communicatively connected to at least one of the processors; wherein,
[0108] The memory stores instructions executable by at least one of the processors, and the instructions are executed by at least one of the processors to enable at least one of the processors to perform the above method. At least one processor in the electronic device can perform the above method, and thus has at least the same advantages as the above method.
[0109] Optionally, the electronic device also includes an interface for connecting various components, including a high-speed interface and a low-speed interface. Various components are interconnected using different buses and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the electronic device, including instructions stored in or on a memory to display graphical information of a GUI (Graphical User Interface) on an external input / output device (such as a display device coupled to an interface). In other embodiments, if necessary, multiple processors can be used together with multiple memories, and / or multiple buses can be used together with multiple memories. Similarly, multiple electronic devices can be connected (for example, as a server array, a group of blade servers, or a multi-processor system), and each device provides some necessary operations. Figure 6 A processor 301 is taken as an example.
[0110] The memory 302, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the vehicle speed prediction method in the embodiment of the present invention (for example, the first graph structure building module, the first output module, the second output module, and the prediction module in the vehicle speed prediction device). The processor 301 executes various functional applications and data processing of the device by running the software programs, instructions and modules stored in the memory 302, that is, realizes the above-mentioned vehicle speed prediction method.
[0111] The memory 302 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 302 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 302 may further include a memory remotely arranged relative to the processor 301, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0112] The electronic device may further include: an input device 303 and an output device 304. The processor 301, the memory 302, the input device 303 and the output device 304 may be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.
[0113] The input device 303 can receive input digital or character information, and the output device 304 can include a display device, an auxiliary lighting device (e.g., LED), and a tactile feedback device (e.g., a vibration motor), etc. The display device can include, but is not limited to, a liquid crystal display (LCD), a light emitting diode (LED) display, and a plasma display. In some embodiments, the display device can be a touch screen.
[0114] Example 4
[0115] This embodiment provides a computer-readable storage medium, on which computer instructions are stored, and the computer instructions are used to enable the computer to execute the above method. The computer instructions on the computer-readable storage medium are used to enable the computer to execute the above method, and thus have at least the same advantages as the above method.
[0116] The medium in the present invention may adopt any combination of one or more computer-readable media. The medium may be a computer-readable signal medium or a computer-readable storage medium. The medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples of the medium (a non-exhaustive list) include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, the medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device, or device.
[0117] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.
[0118] The program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, RF (Radio Frequency), etc., or any suitable combination of the above.
[0119] Computer program code for performing the operations of the present invention may be written in one or more programming languages or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).
[0120] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this application can be executed in parallel, sequentially or in different orders, as long as the expected results of the technical solution disclosed in this application can be achieved, and this document is not limited here.
[0121] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.
Claims
1. A vehicle speed prediction method, characterized in that: include: According to the vehicle speed data of each time period, a first graph structure of each time period is constructed; According to the first graph structure, a second graph structure is constructed; the second graph structure includes key features generated by graph convolutions of different orders; Performing a graph convolution operation on the second graph structure to obtain a first output feature; According to the first graph structure, a third graph structure is constructed; the third graph structure includes key features generated by connecting different numbers of nodes; Performing a graph convolution operation on the third graph structure to obtain a second output feature; The vehicle speed is predicted based on the first output feature and the second output feature.
2. The vehicle speed prediction method according to claim 1, characterized in that: The step of constructing a first graph structure for each time period according to the vehicle speed data for each time period includes: The vehicle speed data of each time period is extracted using a multi-layer perceptron to obtain initial features; A first graph structure is constructed according to the initial features.
3. The vehicle speed prediction method according to claim 1, characterized in that: The step of constructing a second graph structure according to the first graph structure includes: Extracting key nodes according to the first graph structure; Performing at least three graph convolution operations on the first graph structure, and extracting a first feature of the key node from an output feature of each graph convolution operation; A second graph structure is constructed according to the first feature of the key node.
4. The vehicle speed prediction method according to claim 3, characterized in that: The step of constructing a second graph structure according to the first feature of the key node includes: Calculating the similarity of the first feature of the key node to obtain a first similarity adjacency matrix; A second graph structure is constructed according to the first feature of the key node and the first similarity adjacency matrix.
5. The vehicle speed prediction method according to claim 1, characterized in that: The step of constructing a third graph structure according to the first graph structure includes: Extracting key nodes according to the first graph structure; Determining an updated first graph structure according to an adjacency matrix of the first graph structure and a set scale; the set scale is used to characterize the amount of aggregated information; Performing a graph convolution operation on the updated first graph structure to extract a second feature of the key node in the output feature; A third graph structure is constructed according to the second feature of the key node.
6. The vehicle speed prediction method according to claim 5, characterized in that: The step of constructing a third graph structure according to the second feature of the key node includes: Calculating the similarity of the second feature of the key node to obtain a second similarity adjacency matrix; A third graph structure is constructed according to the second feature of the key node and the second similarity adjacency matrix.
7. The vehicle speed prediction method according to claim 1, characterized in that: The predicting the vehicle speed according to the first output feature and the second output feature includes: Determine a historical vehicle speed feature according to the first output feature and the second output feature; the historical vehicle speed feature is used to represent historical vehicle speed information; The vehicle speed is predicted according to the historical vehicle speed characteristics.
8. A vehicle speed prediction device, characterized in that: include: A first graph structure building module, used to build a first graph structure for each time period according to the vehicle speed data for each time period; A first output module, configured to construct a second graph structure according to the first graph structure; the second graph structure includes key features generated by graph convolutions of different orders; Performing a graph convolution operation on the second graph structure to obtain a first output feature; A second output module is used to construct a third graph structure according to the first graph structure; the third graph structure includes key features generated by connecting different numbers of nodes; Performing a graph convolution operation on the third graph structure to obtain a second output feature; A prediction module is used to predict the vehicle speed according to the first output feature and the second output feature.
9. An electronic device, characterized in that: include: at least one processor, and a memory communicatively coupled to at least one of the processors; The memory stores instructions executable by at least one of the processors, and the instructions are executed by at least one of the processors so that the at least one processor can execute the method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that: The medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method according to any one of claims 1 to 7.