Vehicle speed prediction method

By performing feature extraction and graph convolution processing on the vehicle's historical vehicle speed and acceleration information, combined with similarity analysis and discriminator optimization, the problem of poor vehicle speed prediction is solved, and more accurate future vehicle speed prediction and energy management strategy optimization is achieved.

CN119078854APending Publication Date: 2024-12-06CHONGQING SELIS PHOENIX INTELLIGENT INNOVATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411392544.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-08
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing vehicle speed prediction methods lack a reasonable modeling mechanism in intelligent driving, resulting in poor accuracy in vehicle speed prediction.

Method used

By processing the historical vehicle speed and acceleration information of the target vehicle, global features and aggregate features are extracted, and the graph convolutional network and similarity analysis are used to predict future vehicle speeds, and the prediction performance is optimized in combination with the discriminator and attention function.

Benefits of technology

It improves the accuracy and reference value of vehicle speed prediction, helps target vehicles to formulate more effective energy management strategies, optimize power output, and reduce energy consumption.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119078854A_ABST
    Figure CN119078854A_ABST
Patent Text Reader

Abstract

The invention discloses a vehicle speed prediction method. The method can be applied to the technical field of intelligent driving, and specifically comprises the following steps: processing historical vehicle speed information of a target vehicle at each historical moment in a historical time period according to global vehicle speed characteristics of the target vehicle in the historical time period to obtain vehicle speed aggregation characteristics of the target vehicle at each historical moment; processing the historical acceleration information of the target vehicle at each historical moment according to the global acceleration characteristics of the target vehicle at the historical time periods to obtain acceleration aggregation characteristics of the target vehicle at each historical moment; and predicting the future vehicle speed of the target vehicle at the future moment according to the first similarity between the vehicle speed aggregation features, the vehicle speed aggregation features, the second similarity between the acceleration aggregation features and the acceleration aggregation features. Based on the method and the device, the finally predicted future vehicle speed is more accurate, and the predicted future vehicle speed is higher in reference value.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and in particular to a vehicle speed prediction method. Background Art

[0002] With the rapid development of intelligent driving technology, the value of vehicle speed prediction technology in intelligent driving has become increasingly obvious. Vehicle speed prediction has many values ​​in intelligent driving. It not only increases the safety of the car during driving, but also facilitates the energy management strategy (EMS) system to obtain the vehicle speed based on the prediction and determine the energy management strategy at the future moment, so as to configure the power distribution between the vehicle power sources and reduce energy consumption.

[0003] With the rapid development of artificial intelligence technology, the current vehicle speed prediction process also applies deep learning methods, such as deep belief networks, stacked autoencoder neural networks, etc. However, this vehicle speed prediction method lacks a reasonable modeling mechanism, resulting in poor accuracy of vehicle speed prediction. Summary of the invention

[0004] Based on this, it is necessary to provide a vehicle speed prediction method that can accurately predict the future vehicle speed in response to the above technical problems.

[0005] In a first aspect, the present application provides a vehicle speed prediction method, the method comprising:

[0006] According to the global speed characteristics of the target vehicle in the historical period, the historical speed information of the target vehicle at each historical moment in the historical period is processed to obtain the speed aggregation characteristics of the target vehicle at each historical moment;

[0007] According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment;

[0008] The future vehicle speed of the target vehicle at a future time is predicted based on the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features.

[0009] In one embodiment, based on the global vehicle speed characteristics of the target vehicle in the historical period, the historical vehicle speed information of the target vehicle at each historical moment in the historical period is processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment, including:

[0010] Extract the historical speed information of the target vehicle at each historical moment in the historical period to obtain the speed value at each historical moment;

[0011] According to the speed values ​​at each historical moment, determine the global speed characteristics of the target vehicle in the historical period;

[0012] According to the global vehicle speed characteristics, the vehicle speed values ​​at each historical moment are processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment.

[0013] In one embodiment, determining the global vehicle speed characteristics of the target vehicle in the historical period according to the vehicle speed values ​​at each historical moment includes:

[0014] According to the vehicle speed values ​​at each historical moment, a historical vehicle speed matrix is ​​constructed;

[0015] Transpose the historical vehicle speed matrix to obtain a transposed matrix of the historical vehicle speed matrix;

[0016] The product of the historical speed matrix and the transposed matrix is ​​used as the global speed feature of the target vehicle in the historical period.

[0017] In one embodiment, the vehicle speed values ​​at each historical moment are processed according to the global vehicle speed feature to obtain the vehicle speed aggregation feature of the target vehicle at each historical moment, including:

[0018] Preprocessing the global vehicle speed feature to obtain a preprocessed global vehicle speed feature; wherein the preprocessing at least includes normalization processing;

[0019] The preprocessed global vehicle speed features and the vehicle speed values ​​at each historical moment are input into the graph convolutional network to obtain the speed aggregation features of the target vehicle at each historical moment.

[0020] In one embodiment, according to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment, including:

[0021] Feature extraction is performed on the historical acceleration information of the target vehicle at each historical moment in the historical period to obtain the acceleration value at each historical moment;

[0022] According to the acceleration values ​​at each historical moment, the global acceleration characteristics of the target vehicle in the historical period are determined;

[0023] According to the global acceleration characteristics, the acceleration values ​​at each historical moment are processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment.

[0024] In one embodiment, determining the global acceleration characteristics of the target vehicle in the historical period according to the acceleration values ​​at each historical moment includes:

[0025] Construct a historical acceleration matrix based on the Euclidean distance between the acceleration values ​​at each historical moment;

[0026] The historical acceleration matrix is ​​used as the global acceleration feature of the target vehicle in the historical period.

[0027] In one embodiment, the acceleration values ​​at each historical moment are processed according to the global acceleration feature to obtain the acceleration aggregation feature of the target vehicle at each historical moment, including:

[0028] The global acceleration features and the acceleration values ​​at each historical moment are input into the graph convolutional network to obtain the acceleration aggregation features of the target vehicle at each historical moment;

[0029] The method further includes:

[0030] Based on the discriminator, the prediction performance is determined according to the predicted future speed of the target vehicle at the future time and the actual speed of the target vehicle at the future time;

[0031] When the prediction performance does not meet the performance requirements, the weight coefficients in the graph convolutional network are updated.

[0032] In one embodiment, predicting the future speed of the target vehicle at a future time according to the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features, includes:

[0033] According to the first similarity between the vehicle speed aggregate features, selecting the first vehicle speed aggregate feature and the second vehicle speed aggregate feature corresponding to the greatest first similarity from the vehicle speed aggregate features;

[0034] According to the second similarity between the acceleration aggregation features, the first acceleration aggregation feature and the second acceleration aggregation feature corresponding to the greatest second similarity are selected from the acceleration aggregation features;

[0035] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature, the second vehicle speed aggregation feature, the first acceleration aggregation feature and the second acceleration aggregation feature, as well as each vehicle speed aggregation feature and each acceleration aggregation feature.

[0036] In one embodiment, predicting the future speed of the target vehicle at a future time according to the first vehicle speed aggregate feature, the second vehicle speed aggregate feature, the first acceleration aggregate feature and the second acceleration aggregate feature, and each vehicle speed aggregate feature and each acceleration aggregate feature includes:

[0037] Replacing the second vehicle speed aggregation feature in each vehicle speed aggregation feature with the first acceleration aggregation feature to obtain a first vehicle speed aggregation feature group;

[0038] Replacing the first vehicle speed aggregation feature in each vehicle speed aggregation feature with the second acceleration aggregation feature to obtain a second vehicle speed aggregation feature group;

[0039] Replacing the first acceleration aggregation feature in each acceleration aggregation feature with the first velocity aggregation feature to obtain a first acceleration aggregation feature group;

[0040] Replacing the second acceleration aggregation feature in each acceleration aggregation feature with the second velocity aggregation feature to obtain a second acceleration aggregation feature group;

[0041] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature group, the second vehicle speed aggregation feature group, the first acceleration aggregation feature group, and the second acceleration aggregation feature group.

[0042] In one embodiment, predicting the future speed of the target vehicle at a future time according to the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group includes:

[0043] Performing graph convolution processing on each feature group of the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group to obtain a feature value corresponding to each feature group;

[0044] Based on the attention function, determine the attention coefficient of each feature group;

[0045] Determine the prediction results of each feature group according to the feature values ​​and attention coefficients corresponding to each feature group;

[0046] Based on the prediction results of each feature group, the future speed of the target vehicle at the future time is predicted.

[0047] In a second aspect, the present application provides a vehicle speed prediction method, the device comprising:

[0048] The first determination module is used to process the historical speed information of the target vehicle at each historical moment in the historical period according to the global speed characteristics of the target vehicle in the historical period, and obtain the speed aggregation characteristics of the target vehicle at each historical moment;

[0049] The second determination module is used to process the historical acceleration information of the target vehicle at each historical moment according to the global acceleration characteristics of the target vehicle in the historical period, and obtain the acceleration aggregation characteristics of the target vehicle at each historical moment;

[0050] The prediction module is used to predict the future speed of the target vehicle at a future time based on the first similarity between the vehicle speed aggregation features and the vehicle speed aggregation features, and the second similarity between the acceleration aggregation features and the acceleration aggregation features.

[0051] In a third aspect, the present application further provides a computer device, the computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the following steps when executing the computer program:

[0052] According to the global speed characteristics of the target vehicle in the historical period, the historical speed information of the target vehicle at each historical moment in the historical period is processed to obtain the speed aggregation characteristics of the target vehicle at each historical moment;

[0053] According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment;

[0054] The future vehicle speed of the target vehicle at a future time is predicted based on the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features.

[0055] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0056] According to the global speed characteristics of the target vehicle in the historical period, the historical speed information of the target vehicle at each historical moment in the historical period is processed to obtain the speed aggregation characteristics of the target vehicle at each historical moment;

[0057] According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment;

[0058] The future vehicle speed of the target vehicle at a future time is predicted based on the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features.

[0059] In a fifth aspect, the present application further provides a computer program product, the computer program product comprising a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0060] According to the global speed characteristics of the target vehicle in the historical period, the historical speed information of the target vehicle at each historical moment in the historical period is processed to obtain the speed aggregation characteristics of the target vehicle at each historical moment;

[0061] According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment;

[0062] The future vehicle speed of the target vehicle at a future time is predicted based on the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features.

[0063] The above-mentioned vehicle speed prediction method processes the historical vehicle speed information of the target vehicle at each historical moment in the historical period according to the global vehicle speed characteristics of the target vehicle in the historical period, and obtains the vehicle speed aggregation characteristics of the target vehicle at each historical moment. According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment. According to the first similarity between each vehicle speed aggregation characteristic and each vehicle speed aggregation characteristic, and the second similarity between each acceleration aggregation characteristic and each acceleration aggregation characteristic, the future vehicle speed of the target vehicle at a future moment is predicted. Compared with the traditional vehicle speed prediction method, when predicting the future vehicle speed of the target vehicle at a future moment, the present application first combines the global vehicle speed characteristics of the target vehicle in the historical period and the historical vehicle speed information at each historical moment to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment. Then, the global acceleration characteristics in the historical period and the acceleration aggregation characteristic information at each historical moment are combined to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment. Finally, not only the aggregation features of each vehicle speed and each acceleration aggregation feature are considered, but also the first similarity between the aggregation features of each vehicle speed and the second similarity between the aggregation features of each acceleration are considered to predict the vehicle speed at the future moment, so that the final predicted future vehicle speed is more accurate and has more reference value, which can facilitate the target vehicle to determine the energy management strategy at the future moment based on the predicted future vehicle speed, optimize the power output of the target vehicle at the future moment, and reduce energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 A diagram of an application environment of a vehicle speed prediction method provided in this embodiment;

[0065] Figure 2 A schematic flow chart of a first vehicle speed prediction method provided in this embodiment;

[0066] Figure 3 A schematic diagram of a process for determining the vehicle speed aggregation characteristics of a target vehicle at each historical moment provided in this embodiment;

[0067] Figure 4 A schematic diagram of a process for determining the acceleration aggregation characteristics of a target vehicle at each historical moment provided in this embodiment;

[0068] Figure 5 A schematic diagram of a process for predicting the future speed of a target vehicle at a future time provided in this embodiment;

[0069] Figure 6 A schematic diagram of the principle of predicting future vehicle speed based on feature groups provided in this embodiment;

[0070] Figure 7 A schematic flow chart of a second vehicle speed prediction method provided in this embodiment;

[0071] Figure 8 A schematic diagram of the principle of the discriminator provided in this embodiment;

[0072] Fig. 9 A schematic flow chart of a third vehicle speed prediction method provided in this embodiment;

[0073] Fig.10 A structural block diagram of a vehicle speed prediction device provided in this embodiment;

[0074] Fig.11 This is a diagram of the internal structure of the computer device provided in this embodiment. DETAILED DESCRIPTION

[0075] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0076] The path updating method provided in the embodiment of the present application can be applied to the situation of how to predict vehicle speed, and is particularly applicable to the situation of how to predict the future speed of a vehicle. Optionally, the vehicle speed prediction method can be executed by the target vehicle, or can be executed by a server. For example, the target vehicle / server can predict the future vehicle speed based on a preset vehicle speed prediction logic, combined with the historical vehicle speed information and historical acceleration information of the target vehicle at historical moments. Furthermore, the vehicle speed prediction method can also be executed by the interaction between the target vehicle and the server, that is, the method can be applied to the following situations: Figure 1In the application environment shown. Wherein, when there is a need for vehicle speed prediction, the server 104 interacts with the target vehicle 102, obtains the historical vehicle speed information and historical acceleration information of the target vehicle 102 at each historical moment from the target vehicle 102, and processes the historical vehicle speed information of the target vehicle 102 at each historical moment in the historical period according to the global vehicle speed characteristics of the target vehicle 102 in the historical period, and obtains the vehicle speed aggregation characteristics of the target vehicle 102 at each historical moment. And according to the global acceleration characteristics of the target vehicle 102 in the historical period, the historical acceleration information of the target vehicle 102 at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle 102 at each historical moment. Finally, the server 104 predicts the future vehicle speed of the target vehicle 102 at a future moment according to the first similarity between each vehicle speed aggregation characteristic and each vehicle speed aggregation characteristic, and the second similarity between each acceleration aggregation characteristic and each acceleration aggregation characteristic.

[0077] The server 104 can be implemented as an independent server or a server cluster composed of multiple servers. The target vehicle 102 can be an ordinary car, an electric car, or a vehicle with an automatic driving function. The present application does not limit the type of the target vehicle.

[0078] In one embodiment, Figure 2 is a flow chart of a vehicle speed prediction method provided according to an embodiment of the present application, in which the method is applied to Figure 1 Taking the server in as an example, the method includes the following steps:

[0079] S201, based on the global vehicle speed characteristics of the target vehicle in the historical period, the historical vehicle speed information of the target vehicle at each historical moment in the historical period is processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment.

[0080] Among them, the target vehicle refers to the vehicle for which future speed prediction is required. The historical period refers to the period before the current period. The global speed feature refers to the feature that can reflect the speed change of the target vehicle in the historical period as a whole. The historical moment refers to the moment in the historical period. The historical speed information refers to the speed-related information of the target vehicle at the historical moment. The speed aggregation feature refers to the feature obtained by aggregating the global speed feature in the historical period and the historical speed information of each historical moment in the historical period.

[0081] Optionally, in this embodiment, a historical vehicle speed matrix may be constructed based on historical vehicle speed information of the target vehicle at each historical moment in the historical period, and the historical vehicle speed matrix may be used as a global vehicle speed feature of the target vehicle in the historical period.

[0082] Optionally, in this embodiment, the speed fusion value of the target vehicle in the historical period can be determined according to the historical speed information of the target vehicle at each historical moment in the historical period, and the speed fusion value can be used as the global speed feature of the target vehicle in the historical period. The speed fusion value can be the average speed value, the accumulated speed value, etc.

[0083] An optional implementation of this embodiment is: inputting the global vehicle speed characteristics of the target vehicle in the historical period and the historical vehicle speed information of the target vehicle at each historical moment in the historical period into the neural network model, and the neural network model outputs the vehicle speed aggregation characteristics of the target vehicle at each historical moment.

[0084] Another optional implementation of this embodiment is: according to the global vehicle speed characteristics of the target vehicle in the historical period, the historical vehicle speed information of the target vehicle at each historical moment in the historical period is aggregated to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment.

[0085] S202, according to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment.

[0086] Among them, the global acceleration feature refers to the feature that can reflect the acceleration change of the target vehicle in the time period as a whole. The historical acceleration information refers to the acceleration-related information of the target vehicle at the historical moment. The acceleration aggregation feature refers to the feature obtained by aggregating the global acceleration feature in the time period and the historical acceleration information at each historical moment in the historical period.

[0087] Optionally, in this embodiment, a historical acceleration matrix may be constructed based on historical acceleration information of the target vehicle at each historical moment in the historical period, and the historical acceleration matrix may be used as a global acceleration feature of the target vehicle in the historical period.

[0088] Optionally, in this embodiment, the acceleration fusion value of the target vehicle in the historical period can be determined according to the historical acceleration information of the target vehicle at each historical moment in the historical period, and the acceleration fusion value can be used as the global acceleration feature of the target vehicle in the historical period. The acceleration fusion value can be the acceleration mean value, the acceleration accumulation value, etc.

[0089] An optional implementation of this embodiment is: inputting the global acceleration characteristics of the target vehicle in the historical period and the historical acceleration information of the target vehicle at each historical moment in the historical period into the neural network model, and the neural network model outputs the acceleration aggregation characteristics of the target vehicle at each historical moment.

[0090] Another optional implementation of this embodiment is: according to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment in the historical period is aggregated to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment.

[0091] S203, predicting the future vehicle speed of the target vehicle at a future time according to the first similarity between each vehicle speed aggregation feature and the second similarity between each acceleration aggregation feature and the acceleration aggregation feature.

[0092] The first similarity refers to the similarity between the aggregated features of each vehicle speed. The second similarity refers to the similarity between the aggregated features of each acceleration. The future vehicle speed refers to the predicted vehicle speed value of the target vehicle at a future time.

[0093] An optional implementation of the present embodiment is: inputting the first similarity between each vehicle speed aggregation feature and each vehicle speed aggregation feature, and the second similarity between each acceleration aggregation feature and each acceleration aggregation feature into the neural network model, and the neural network model predicts and outputs the future vehicle speed of the target vehicle at the future moment.

[0094] Another optional implementation of this embodiment is: In this embodiment, based on the first similarity between each vehicle speed aggregation feature and each vehicle speed aggregation feature, the two vehicle speed aggregation features with the highest first similarity are determined. Based on the second similarity between each acceleration aggregation feature and each acceleration aggregation feature, the two acceleration aggregation features with the highest second similarity are determined. The two vehicle speed aggregation features with the highest first similarity and the two acceleration aggregation features with the highest second similarity are eliminated. Based on the neural network model, the future vehicle speed of the target vehicle at the future moment is predicted based on the eliminated vehicle speed aggregation features and acceleration aggregation features.

[0095] The above-mentioned vehicle speed prediction method processes the historical vehicle speed information of the target vehicle at each historical moment in the historical period according to the global vehicle speed characteristics of the target vehicle in the historical period, and obtains the vehicle speed aggregation characteristics of the target vehicle at each historical moment. According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment. According to the first similarity between each vehicle speed aggregation characteristic and each vehicle speed aggregation characteristic, and the second similarity between each acceleration aggregation characteristic and each acceleration aggregation characteristic, the future vehicle speed of the target vehicle at a future moment is predicted. Compared with the traditional vehicle speed prediction method, when predicting the future vehicle speed of the target vehicle at a future moment, the present application first combines the global vehicle speed characteristics of the target vehicle in the historical period and the historical vehicle speed information at each historical moment to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment. Then, the global acceleration characteristics in the historical period and the acceleration aggregation characteristic information at each historical moment are combined to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment. Finally, not only the aggregation features of each vehicle speed and each acceleration aggregation feature are considered, but also the first similarity between the aggregation features of each vehicle speed and the second similarity between the aggregation features of each acceleration are considered to predict the vehicle speed at the future moment, so that the final predicted future vehicle speed is more accurate and has more reference value, which can facilitate the target vehicle to determine the energy management strategy at the future moment based on the predicted future vehicle speed, optimize the power output of the target vehicle at the future moment, and reduce energy consumption.

[0096] In one embodiment, in order to more accurately determine the vehicle speed aggregation characteristics of the target vehicle at each historical moment in the historical period, such as Figure 3 As shown, an optional implementation of S201 includes:

[0097] S301, extracting features from historical vehicle speed information of the target vehicle at each historical moment in a historical period to obtain the vehicle speed value at each historical moment.

[0098] The vehicle speed value refers to the driving speed of the target vehicle extracted from the historical vehicle speed information.

[0099] Optionally, in this embodiment, the historical vehicle speed information at each historical moment in the historical period can be feature extracted through a convolutional network to obtain the vehicle speed value at each historical moment. Since the historical vehicle speed information is one-dimensional data, the convolutional network can adopt a one-dimensional convolutional network.

[0100] S302: Determine the global vehicle speed characteristics of the target vehicle in the historical period according to the vehicle speed values ​​at each historical moment.

[0101] Optionally, in this embodiment, a historical vehicle speed matrix is ​​constructed according to the vehicle speed values ​​at each historical moment. The historical vehicle speed matrix is ​​transposed to obtain a transposed matrix of the historical vehicle speed matrix. The product of the historical vehicle speed matrix and the transposed matrix is ​​used as the global vehicle speed feature of the target vehicle in the historical period.

[0102] Optionally, in this embodiment, the global vehicle speed characteristic of the target vehicle in the historical period can be determined by the following formula (1):

[0103]

[0104] Among them, A in formula (1) V1 H represents the historical vehicle speed adjacency matrix. In this application, the historical vehicle speed adjacency matrix is ​​used as the global vehicle speed feature of the target vehicle in the historical period; V1 Represents the historical vehicle speed matrix; Represents the transposed matrix.

[0105] S303: Process the vehicle speed values ​​at each historical moment according to the global vehicle speed characteristics to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment.

[0106] Optionally, in this embodiment, the global vehicle speed feature is preprocessed to obtain the preprocessed global vehicle speed feature; wherein the preprocessing includes at least normalization processing. The preprocessed global vehicle speed feature and the vehicle speed value at each historical moment are input into the graph convolutional network to obtain the vehicle speed aggregation feature of the target vehicle at each historical moment.

[0107] Optionally, in this embodiment, the vehicle speed aggregation feature at each historical moment may be determined by the following formula (2):

[0108]

[0109] Among them, H 1i represents the aggregated feature of vehicle speed at time i; σ represents the Relu activation function; represents the global vehicle speed feature after preprocessing; X li represents the vehicle speed at time i; W i Represents the weight coefficient of the graph convolutional network.

[0110] In this embodiment, feature extraction is performed on the historical speed information of the target vehicle at each historical moment in the historical period to obtain the speed value at each historical moment. Based on the speed value at each historical moment, the global speed feature of the target vehicle in the historical period is determined. Based on the global speed feature, the speed value at each historical moment is processed to obtain the speed aggregation feature of the target vehicle at each historical moment. Based on this embodiment, the obtained global speed feature is more accurate, and based on the graph convolution network of the present application, the global speed feature and the historical speed feature at each historical moment are subjected to graph convolution aggregation, and the obtained speed aggregation feature at each historical moment is also more accurate, which effectively improves the accuracy of the prediction result.

[0111] In one embodiment, in order to more accurately determine the acceleration aggregation characteristics of the target vehicle at each historical moment in the historical period, such as Figure 4 As shown, an optional implementation of S202 includes:

[0112] S401, extracting features from historical acceleration information of the target vehicle at each historical moment in a historical period to obtain acceleration values ​​at each historical moment.

[0113] The acceleration value refers to the acceleration of the target vehicle at a historical moment obtained by extracting features based on historical acceleration information. It should be noted that the historical acceleration information includes both acceleration and other related information.

[0114] Optionally, in this embodiment, based on a convolutional network, feature extraction is performed on the historical acceleration information of the target vehicle at each historical moment in the historical period to obtain the acceleration value at each historical moment. Since the historical acceleration information is one-dimensional data, the present application can use a one-dimensional convolutional network to feature extract the historical acceleration information of the target vehicle at each historical moment in the historical period to obtain the acceleration value at each historical moment.

[0115] S402: Determine the global acceleration characteristics of the target vehicle in the historical period according to the acceleration values ​​at each historical moment.

[0116] Optionally, in this embodiment, a historical acceleration matrix is ​​constructed according to the Euclidean distance between the acceleration values ​​at each historical moment. The historical acceleration matrix is ​​used as the global acceleration feature of the target vehicle in the historical period. In this embodiment, an optional implementation method for constructing the historical acceleration matrix according to the Euclidean distance between the acceleration values ​​at each historical moment is: respectively calculate the Euclidean distance between the acceleration value at each historical moment and the acceleration value at other historical moments, and then obtain an n*n historical acceleration matrix.

[0117] Optionally, in this embodiment, the Euclidean distance between the acceleration values ​​at each historical moment can be determined by the following formula (3):

[0118]

[0119] Among them, E in formula (3) ij represents the Euclidean distance between the acceleration values ​​at time i and time j; x(i, n) represents the acceleration value at time i; x(j, n) represents the acceleration value at time j.

[0120] S403, processing the acceleration values ​​at each historical moment according to the global acceleration characteristics to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment.

[0121] Optionally, in this embodiment, the global acceleration feature is preprocessed, and the preprocessed global acceleration feature and the acceleration value at each historical moment are input into the graph convolution network to obtain the acceleration aggregation feature of the target vehicle at each historical moment. The preprocessing method adopts but is not limited to normalization processing.

[0122] Optionally, in this embodiment, the acceleration aggregation feature of the target vehicle at each historical moment can be obtained based on the following formula (4) combined with the preprocessed global acceleration feature and the acceleration value at each historical moment:

[0123]

[0124] Among them, H in formula (4) 2i represents the acceleration aggregation feature at time i; σ represents the Relu activation function; represents the acceleration adjacency matrix. In this application, the acceleration adjacency matrix is ​​replaced by the historical acceleration matrix, that is, this feature also represents the global acceleration feature; X 2i represents the acceleration value at time i; W i Represents the weight coefficient of the graph convolutional network.

[0125] In this embodiment, feature extraction is performed on the historical acceleration information of the target vehicle at each historical moment in the historical period to obtain the acceleration value at each historical moment. According to the Euclidean distance between the acceleration values ​​at each historical moment, a historical acceleration matrix is ​​constructed, and the historical acceleration matrix is ​​used as the global acceleration feature of the target vehicle in the historical period. According to the global acceleration feature, the acceleration value at each historical moment is processed to obtain the acceleration aggregation feature of the target vehicle at each historical moment. In this embodiment, the Euclidean distance between the acceleration values ​​at each historical moment is used to determine the global acceleration feature of the target vehicle in the historical period, which increases the accuracy of the global acceleration feature, and based on the graph convolution network of the present application, the global acceleration feature and the acceleration value at each historical moment are combined to determine the acceleration aggregation feature at each historical moment, so that the obtained acceleration aggregation feature is more accurate, which increases the accuracy of predicting the future vehicle speed at future moments.

[0126] In one embodiment, in order to increase the accuracy of predicting the future speed of the target vehicle at a future time, as Figure 5 As shown, an optional implementation of S203 includes:

[0127] S501 , according to the first similarities between the vehicle speed aggregate features, selecting a first vehicle speed aggregate feature and a second vehicle speed aggregate feature corresponding to the greatest first similarity from the vehicle speed aggregate features.

[0128] The first vehicle speed aggregation feature refers to one of the pair of vehicle speed aggregation features with the largest first similarity. The second vehicle speed aggregation feature refers to the other of the pair of vehicle speed aggregation features with the largest first similarity.

[0129] Optionally, in this embodiment, the first similarities between the vehicle speed aggregation features are sorted, and a pair of vehicle speed aggregation features corresponding to the first similarity with the greatest similarity are selected as the first vehicle speed aggregation feature and the second vehicle speed aggregation feature, respectively.

[0130] Optionally, in this embodiment, the first similarity between the vehicle speed aggregation features may be determined by the following formula (5):

[0131]

[0132] Wherein, S1(i,j) in formula (5) represents the first similarity between the vehicle speed aggregation feature at time i and the vehicle speed aggregation feature at time j; H 1i represents the vehicle speed aggregation characteristics at time i; H 1j Represents the vehicle speed aggregation characteristics at time j.

[0133] S502: According to the second similarities between the acceleration aggregation features, a first acceleration aggregation feature and a second acceleration aggregation feature corresponding to the greatest second similarity are selected from the acceleration aggregation features.

[0134] The first acceleration aggregation feature refers to one of the acceleration aggregation features in the pair of acceleration aggregation features with the second greatest similarity. The second acceleration aggregation feature refers to the other acceleration aggregation feature in the pair of acceleration aggregation features with the second greatest similarity.

[0135] Optionally, in this embodiment, the second similarities between the acceleration aggregation features are sorted, and a pair of acceleration aggregation features corresponding to the second similarity with the greatest similarity are selected as the first acceleration aggregation feature and the second acceleration aggregation feature, respectively.

[0136] Optionally, in this embodiment, the second similarity between the acceleration aggregation features can be determined by the following formula (6):

[0137]

[0138] Wherein, S2(i,j) in formula (6) represents the second similarity between the acceleration aggregation feature at time i and the acceleration aggregation feature at time j; H 2i represents the acceleration aggregation characteristics at time i; H 2j Represents the acceleration aggregation characteristics at time j.

[0139] S503, predicting the future vehicle speed of the target vehicle at a future time based on the first vehicle speed aggregation feature, the second vehicle speed aggregation feature, the first acceleration aggregation feature and the second acceleration aggregation feature, as well as each vehicle speed aggregation feature and each acceleration aggregation feature.

[0140] Optionally, in this embodiment, the second vehicle speed aggregation feature in each vehicle speed aggregation feature is replaced by the first acceleration aggregation feature to obtain a first vehicle speed aggregation feature group. The first vehicle speed aggregation feature in each vehicle speed aggregation feature is replaced by the second acceleration aggregation feature to obtain a second vehicle speed aggregation feature group. The first acceleration aggregation feature in each acceleration aggregation feature is replaced by the first speed aggregation feature to obtain a first acceleration aggregation feature group. The second acceleration aggregation feature in each acceleration aggregation feature is replaced by the second speed aggregation feature to obtain a second acceleration aggregation feature group. Based on the first vehicle speed aggregation feature group, the second vehicle speed aggregation feature group, the first acceleration aggregation feature group and the second acceleration aggregation feature group, the future vehicle speed of the target vehicle at the future moment is predicted.

[0141] In this embodiment, an optional implementation method for predicting the future speed of the target vehicle at a future time is as follows based on the first vehicle speed aggregation feature group, the second vehicle speed aggregation feature group, the first acceleration aggregation feature group, and the second acceleration aggregation feature group: Figure 6 As shown, graph convolution is performed on each feature group in the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group to obtain the feature value corresponding to each feature group. Based on the attention function, the attention coefficient of each feature group is determined. According to the feature value and attention coefficient corresponding to each feature group, the prediction result of each feature group is determined. According to the prediction result of each feature group, the future speed of the target vehicle at the future time is predicted. Specifically, as Figure 6 As shown in the figure, for each feature group, the feature group is first input into the graph convolution network, and the graph convolution network outputs the feature value. The feature value is also input into the one-dimensional convolution network for feature convolution processing, and the result after the one-dimensional convolution layer processing is input into the attention function (i.e. Figure 6The attention coefficient of the feature group is determined by the attention function, and the output result of the feature group is obtained by multiplying the eigenvalue corresponding to the feature group and the attention coefficient corresponding to the feature group. Finally, the output results of each feature group are fused (for example, summed) to obtain the future speed of the target vehicle at the future time (i.e. Figure 6 H out ).

[0142] In this embodiment, according to the first similarity between the vehicle speed aggregation features, the first vehicle speed aggregation feature and the second vehicle speed aggregation feature corresponding to the maximum first similarity are selected from the vehicle speed aggregation features. According to the second similarity between the acceleration aggregation features, the first acceleration aggregation feature and the second acceleration aggregation feature corresponding to the maximum second similarity are selected from the acceleration aggregation features. According to the first vehicle speed aggregation feature, the second vehicle speed aggregation feature, the first acceleration aggregation feature and the second acceleration aggregation feature, as well as the vehicle speed aggregation features and the acceleration aggregation features, the graph convolution network and the attention function are combined to make the predicted future vehicle speed of the target vehicle at the future moment more accurate.

[0143] In this embodiment, the second vehicle speed aggregation feature in each vehicle speed aggregation feature is replaced by the first acceleration aggregation feature to obtain a first vehicle speed aggregation feature group. The first vehicle speed aggregation feature in each vehicle speed aggregation feature is replaced by the second acceleration aggregation feature to obtain a second vehicle speed aggregation feature group. The first acceleration aggregation feature in each acceleration aggregation feature is replaced by the first speed aggregation feature to obtain a first acceleration aggregation feature group. The second acceleration aggregation feature in each acceleration aggregation feature is replaced by the second speed aggregation feature to obtain a second acceleration aggregation feature group. According to the first vehicle speed aggregation feature group, the second vehicle speed aggregation feature group, the first acceleration aggregation feature group and the second acceleration aggregation feature group, the future vehicle speed of the target vehicle at a future moment is predicted. Based on this embodiment, not only the accuracy of the prediction result is increased, but also the repeated transmission of information is effectively avoided, and the differences and diversity between the features in the feature group are increased, thereby increasing the fluency and generalization performance of the prediction process.

[0144] In one embodiment, in order to further improve the accuracy of the prediction results, Figure 7 As shown, an optional implementation of a vehicle speed prediction method includes:

[0145] S701, based on the discriminator, determining prediction performance according to the predicted future speed of the target vehicle at the future time and the actual speed of the target vehicle at the future time.

[0146] Among them, the discriminator refers to the network model used to identify the accuracy of predicting future vehicle speed.

[0147] Optional, such as Figure 8 As shown, in this embodiment, the predicted future speed of the target vehicle at the future moment and the actual speed of the target vehicle at the future moment are input into the discriminator, which discriminates the prediction result and outputs the prediction performance.

[0148] In this embodiment, the objective function of the discriminator is shown in formula (7):

[0149]

[0150] In formula (7), the independent variable x represents the actual speed of the target vehicle at the future time, and the logarithmic function does not change the monotonicity of the function. represents the probability of classifying a real sample as real; D(G(z)) represents the probability of classifying a fake sample as real; correspondingly, 1-D(G(z)) represents the probability of classifying a fake sample as fake. For the discriminator, the higher this ability is, the better the discriminator effect is. Therefore, the parameter optimization process of the discriminator is a process of maximizing the objective function, that is, using It means that when the discriminator parameters are fixed, the first term of the objective function becomes a constant. When the generator parameters are adjusted, only the second term changes. The second term represents the ability to classify fake samples as fake. For the vehicle speed prediction method of this application, this ability should be as small as possible, so that the generator can deceive the discriminator. All optimization of the generator is a process of minimizing the objective function, that is, using express.

[0151] In this embodiment, if the discriminator outputs a false result, it indicates that the prediction result is inaccurate, and accordingly, the prediction performance of the present application is worse. On the contrary, if the discriminator outputs a true result, it indicates that the prediction result is accurate, and accordingly, the prediction performance of the present application is more accurate.

[0152] S702: When the prediction performance does not meet the performance requirements, update the weight coefficients in the graph convolutional network.

[0153] An optional implementation of this embodiment is: when the prediction performance does not meet the performance requirements, the discriminator controls the updating of the weight coefficients in the graph convolutional network, thereby realizing intelligent updating of the weight coefficients of the graph convolutional network.

[0154] Another optional implementation of this embodiment is: when the prediction performance does not meet the performance requirements, the weight coefficients in the graph convolutional network are updated manually or in combination with other tools.

[0155] In this embodiment, based on the discriminator, the prediction performance is determined according to the predicted future speed of the target vehicle at the future moment and the actual speed of the target vehicle at the future moment. In the case where the predicted performance does not meet the performance requirements, the weight coefficients in the graph convolutional network are updated. Based on this embodiment, not only can the prediction performance of the present application be automatically determined, but also the weight coefficients in the graph convolutional network can be updated in the case where the predicted performance does not meet the performance requirements, thereby realizing the update of the prediction method, further improving the accuracy of the prediction results, and realizing a closed-loop optimization process based on the prediction results and the actual results, thereby improving the optimization efficiency of the prediction performance.

[0156] In one embodiment, if Fig. 9 As shown, another optional implementation of the vehicle speed prediction method includes:

[0157] S901, extracting features from historical vehicle speed information of the target vehicle at each historical moment in a historical period to obtain the vehicle speed value at each historical moment.

[0158] S902: construct a historical vehicle speed matrix according to the vehicle speed values ​​at each historical moment.

[0159] S903, performing transposition processing on the historical vehicle speed matrix to obtain a transposed matrix of the historical vehicle speed matrix.

[0160] S904: The product of the historical vehicle speed matrix and the transposed matrix is ​​used as the global vehicle speed feature of the target vehicle in the historical period.

[0161] S905, preprocessing the global vehicle speed feature to obtain the preprocessed global vehicle speed feature, wherein the preprocessing at least includes normalization processing.

[0162] S906: Input the preprocessed global vehicle speed features and the vehicle speed values ​​at each historical moment into the graph convolutional network to obtain the vehicle speed aggregation features of the target vehicle at each historical moment.

[0163] S907, extracting features from historical acceleration information of the target vehicle at each historical moment in the historical period to obtain acceleration values ​​at each historical moment.

[0164] S908: construct a historical acceleration matrix according to the Euclidean distances between the acceleration values ​​at each historical moment.

[0165] S909: Use the historical acceleration matrix as the global acceleration feature of the target vehicle in the historical period.

[0166] S910, inputting the global acceleration features and the acceleration values ​​at each historical moment into the graph convolutional network to obtain the acceleration aggregation features of the target vehicle at each historical moment.

[0167] S911, according to the first similarities between the vehicle speed aggregate features, select a first vehicle speed aggregate feature and a second vehicle speed aggregate feature corresponding to the greatest first similarity from the vehicle speed aggregate features.

[0168] S912: According to the second similarities between the acceleration aggregation features, select the first acceleration aggregation feature and the second acceleration aggregation feature corresponding to the greatest second similarity from among the acceleration aggregation features.

[0169] S913: Replace the second vehicle speed aggregation feature in each vehicle speed aggregation feature with the first acceleration aggregation feature to obtain a first vehicle speed aggregation feature group.

[0170] S914: Replace the first vehicle speed aggregation feature in each vehicle speed aggregation feature with the second acceleration aggregation feature to obtain a second vehicle speed aggregation feature group.

[0171] S915: Replace the first acceleration aggregation feature in each acceleration aggregation feature with the first velocity aggregation feature to obtain a first acceleration aggregation feature group.

[0172] S916: Replace the second acceleration aggregation feature in each acceleration aggregation feature with the second velocity aggregation feature to obtain a second acceleration aggregation feature group.

[0173] S917, performing graph convolution processing on each feature group of the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group to obtain a feature value corresponding to each feature group.

[0174] S918, based on the attention function, determine the attention coefficient of each feature group.

[0175] S919, determining the prediction result of each feature group according to the feature value and attention coefficient corresponding to each feature group.

[0176] S920: predict the future speed of the target vehicle at a future time based on the prediction results of each feature group.

[0177] S921, based on the discriminator, determine the prediction performance according to the predicted future speed of the target vehicle at the future time and the actual speed of the target vehicle at the future time.

[0178] S922: When the prediction performance does not meet the performance requirements, update the weight coefficients in the graph convolutional network.

[0179] The vehicle speed prediction method of this embodiment processes the historical vehicle speed information of the target vehicle at each historical moment in the historical period according to the global vehicle speed characteristics of the target vehicle in the historical period, and obtains the vehicle speed aggregation characteristics of the target vehicle at each historical moment. According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment. According to the first similarity between each vehicle speed aggregation characteristic and each vehicle speed aggregation characteristic, and the second similarity between each acceleration aggregation characteristic and each acceleration aggregation characteristic, the future vehicle speed of the target vehicle at a future moment is predicted. Compared with the traditional vehicle speed prediction method, when predicting the future vehicle speed of the target vehicle at a future moment, the present application first combines the global vehicle speed characteristics of the target vehicle in the historical period and the historical vehicle speed information at each historical moment to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment. Then, the global acceleration characteristics in the historical period and the acceleration aggregation characteristic information at each historical moment are combined to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment. Finally, not only the aggregation features of each vehicle speed and each acceleration aggregation feature are considered, but also the first similarity between the aggregation features of each vehicle speed and the second similarity between the aggregation features of each acceleration are considered to predict the vehicle speed at the future moment, so that the final predicted future vehicle speed is more accurate and has more reference value, which can facilitate the target vehicle to determine the energy management strategy at the future moment based on the predicted future vehicle speed, optimize the power output of the target vehicle at the future moment, and reduce energy consumption.

[0180] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0181] Based on the same inventive concept, the embodiment of the present application also provides a vehicle speed prediction device for implementing the vehicle speed prediction method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more vehicle speed prediction device embodiments provided below can refer to the limitations of the vehicle speed prediction method above, and will not be repeated here.

[0182] In one embodiment, by Fig.10 FIG. 2 shows a structural block diagram of a vehicle speed prediction device in one embodiment. Fig.10 As shown, a vehicle speed prediction device 1 is provided, the device comprising: a first determination module 10, a second determination module 20 and a prediction module 30, wherein:

[0183] The first determination module 10 is used to process the historical speed information of the target vehicle at each historical moment in the historical period according to the global speed characteristics of the target vehicle in the historical period, and obtain the speed aggregation characteristics of the target vehicle at each historical moment;

[0184] The second determination module 20 is used to process the historical acceleration information of the target vehicle at each historical moment according to the global acceleration characteristics of the target vehicle in the historical period, and obtain the acceleration aggregation characteristics of the target vehicle at each historical moment;

[0185] The prediction module 30 is used to predict the future speed of the target vehicle at a future time according to the first similarity between the vehicle speed aggregation features and the vehicle speed aggregation features, and the second similarity between the acceleration aggregation features and the acceleration aggregation features.

[0186] In one embodiment, the Fig.10 The first determining module 10 is further specifically used for:

[0187] Extract the historical speed information of the target vehicle at each historical moment in the historical period to obtain the speed value at each historical moment;

[0188] According to the speed values ​​at each historical moment, determine the global speed characteristics of the target vehicle in the historical period;

[0189] According to the global vehicle speed characteristics, the vehicle speed values ​​at each historical moment are processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment.

[0190] In one embodiment, the Fig.10 The first determining module 10 is further specifically used for:

[0191] According to the vehicle speed values ​​at each historical moment, a historical vehicle speed matrix is ​​constructed;

[0192] Transpose the historical vehicle speed matrix to obtain a transposed matrix of the historical vehicle speed matrix;

[0193] The product of the historical speed matrix and the transposed matrix is ​​used as the global speed feature of the target vehicle in the historical period.

[0194] In one embodiment, the Fig.10 The first determining module 10 is further specifically used for:

[0195] Preprocessing the global vehicle speed feature to obtain a preprocessed global vehicle speed feature; wherein the preprocessing at least includes normalization processing;

[0196] The preprocessed global vehicle speed features and the vehicle speed values ​​at each historical moment are input into the graph convolutional network to obtain the speed aggregation features of the target vehicle at each historical moment.

[0197] In one embodiment, the Fig.10 The second determining module 20 is further specifically used for:

[0198] Feature extraction is performed on the historical acceleration information of the target vehicle at each historical moment in the historical period to obtain the acceleration value at each historical moment;

[0199] According to the acceleration values ​​at each historical moment, the global acceleration characteristics of the target vehicle in the historical period are determined;

[0200] According to the global acceleration characteristics, the acceleration values ​​at each historical moment are processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment.

[0201] In one embodiment, the Fig.10 The second determining module 20 is further specifically used for:

[0202] Construct a historical acceleration matrix based on the Euclidean distance between the acceleration values ​​at each historical moment;

[0203] The historical acceleration matrix is ​​used as the global acceleration feature of the target vehicle in the historical period.

[0204] In one embodiment, the Fig.10 The second determining module 20 is further specifically used for:

[0205] The global acceleration features and the acceleration values ​​at each historical moment are input into the graph convolutional network to obtain the acceleration aggregation features of the target vehicle at each historical moment;

[0206] Based on the above embodiments, Fig.10 The vehicle speed prediction device 1 further comprises:

[0207] The third determination module is used to determine the prediction performance based on the discriminator according to the predicted future speed of the target vehicle at the future moment and the actual speed of the target vehicle at the future moment.

[0208] The update module is used to update the weight coefficients in the graph convolutional network when the prediction performance does not meet the performance requirements.

[0209] In one embodiment, the Fig.10 The prediction module 30 is also specifically used for:

[0210] According to the first similarity between the vehicle speed aggregate features, selecting the first vehicle speed aggregate feature and the second vehicle speed aggregate feature corresponding to the greatest first similarity from the vehicle speed aggregate features;

[0211] According to the second similarity between the acceleration aggregation features, the first acceleration aggregation feature and the second acceleration aggregation feature corresponding to the greatest second similarity are selected from the acceleration aggregation features;

[0212] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature, the second vehicle speed aggregation feature, the first acceleration aggregation feature and the second acceleration aggregation feature, as well as each vehicle speed aggregation feature and each acceleration aggregation feature.

[0213] In one embodiment, the Fig.10 The prediction module 30 is also specifically used for:

[0214] Replacing the second vehicle speed aggregation feature in each vehicle speed aggregation feature with the first acceleration aggregation feature to obtain a first vehicle speed aggregation feature group;

[0215] Replacing the first vehicle speed aggregation feature in each vehicle speed aggregation feature with the second acceleration aggregation feature to obtain a second vehicle speed aggregation feature group;

[0216] Replacing the first acceleration aggregation feature in each acceleration aggregation feature with the first velocity aggregation feature to obtain a first acceleration aggregation feature group;

[0217] Replacing the second acceleration aggregation feature in each acceleration aggregation feature with the second velocity aggregation feature to obtain a second acceleration aggregation feature group;

[0218] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature group, the second vehicle speed aggregation feature group, the first acceleration aggregation feature group, and the second acceleration aggregation feature group.

[0219] In one embodiment, the Fig.10 The prediction module 30 is also specifically used for:

[0220] Performing graph convolution processing on each feature group of the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group to obtain a feature value corresponding to each feature group;

[0221] Based on the attention function, determine the attention coefficient of each feature group;

[0222] Determine the prediction results of each feature group according to the feature values ​​and attention coefficients corresponding to each feature group;

[0223] Based on the prediction results of each feature group, the future speed of the target vehicle at the future time is predicted.

[0224] Each module in the above vehicle speed prediction device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module.

[0225] In one embodiment, a computer device is provided. The computer device may be a platform side, and its internal structure diagram may be as shown in FIG. Fig.11 As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store vehicle parameter information. The network interface of the computer device is used to communicate with an external user side through a network connection. When the computer program is executed by the processor, a vehicle speed prediction method is implemented.

[0226] Those skilled in the art will understand that Fig.11 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. Specifically, the computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0227] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0228] According to the global speed characteristics of the target vehicle in the historical period, the historical speed information of the target vehicle at each historical moment in the historical period is processed to obtain the speed aggregation characteristics of the target vehicle at each historical moment;

[0229] According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment;

[0230] The future vehicle speed of the target vehicle at a future time is predicted based on the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features.

[0231] In one embodiment, when the processor executes the computer program, the following steps are further implemented: according to the global vehicle speed characteristics of the target vehicle in the historical period, the historical vehicle speed information of the target vehicle at each historical moment in the historical period is processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment, including:

[0232] Extract the historical speed information of the target vehicle at each historical moment in the historical period to obtain the speed value at each historical moment;

[0233] According to the speed values ​​at each historical moment, determine the global speed characteristics of the target vehicle in the historical period;

[0234] According to the global vehicle speed characteristics, the vehicle speed values ​​at each historical moment are processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment.

[0235] In one embodiment, when the processor executes the computer program, the following steps are further implemented: determining the global vehicle speed characteristics of the target vehicle in the historical period according to the vehicle speed values ​​at each historical moment, including:

[0236] According to the vehicle speed values ​​at each historical moment, a historical vehicle speed matrix is ​​constructed;

[0237] Transpose the historical vehicle speed matrix to obtain a transposed matrix of the historical vehicle speed matrix;

[0238] The product of the historical speed matrix and the transposed matrix is ​​used as the global speed feature of the target vehicle in the historical period.

[0239] In one embodiment, when the processor executes the computer program, the following steps are further implemented: according to the global vehicle speed feature, the vehicle speed value at each historical moment is processed to obtain the vehicle speed aggregation feature of the target vehicle at each historical moment, including:

[0240] Preprocessing the global vehicle speed feature to obtain a preprocessed global vehicle speed feature; wherein the preprocessing at least includes normalization processing;

[0241] The preprocessed global vehicle speed features and the vehicle speed values ​​at each historical moment are input into the graph convolutional network to obtain the speed aggregation features of the target vehicle at each historical moment.

[0242] In one embodiment, when the processor executes the computer program, the following steps are further implemented: according to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment, including:

[0243] Feature extraction is performed on the historical acceleration information of the target vehicle at each historical moment in the historical period to obtain the acceleration value at each historical moment;

[0244] According to the acceleration values ​​at each historical moment, the global acceleration characteristics of the target vehicle in the historical period are determined;

[0245] According to the global acceleration characteristics, the acceleration values ​​at each historical moment are processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment.

[0246] In one embodiment, when the processor executes the computer program, the following steps are further implemented: determining the global acceleration characteristics of the target vehicle in the historical period according to the acceleration values ​​at each historical moment, including:

[0247] Construct a historical acceleration matrix based on the Euclidean distance between the acceleration values ​​at each historical moment;

[0248] The historical acceleration matrix is ​​used as the global acceleration feature of the target vehicle in the historical period.

[0249] In one embodiment, when the processor executes the computer program, the following steps are further implemented: according to the global acceleration feature, the acceleration value at each historical moment is processed to obtain the acceleration aggregation feature of the target vehicle at each historical moment, including:

[0250] The global acceleration features and the acceleration values ​​at each historical moment are input into the graph convolutional network to obtain the acceleration aggregation features of the target vehicle at each historical moment;

[0251] The method further includes:

[0252] Based on the discriminator, the prediction performance is determined according to the predicted future speed of the target vehicle at the future time and the actual speed of the target vehicle at the future time;

[0253] When the prediction performance does not meet the performance requirements, the weight coefficients in the graph convolutional network are updated.

[0254] In one embodiment, when the processor executes the computer program, the following steps are further implemented: predicting the future speed of the target vehicle at a future time according to the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features, including:

[0255] According to the first similarity between the vehicle speed aggregate features, selecting the first vehicle speed aggregate feature and the second vehicle speed aggregate feature corresponding to the greatest first similarity from the vehicle speed aggregate features;

[0256] According to the second similarity between the acceleration aggregation features, the first acceleration aggregation feature and the second acceleration aggregation feature corresponding to the greatest second similarity are selected from the acceleration aggregation features;

[0257] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature, the second vehicle speed aggregation feature, the first acceleration aggregation feature and the second acceleration aggregation feature, as well as each vehicle speed aggregation feature and each acceleration aggregation feature.

[0258] In one embodiment, when the processor executes the computer program, the following steps are further implemented: predicting the future speed of the target vehicle at a future time according to the first vehicle speed aggregate feature, the second vehicle speed aggregate feature, the first acceleration aggregate feature and the second acceleration aggregate feature, and each vehicle speed aggregate feature and each acceleration aggregate feature, including:

[0259] Replacing the second vehicle speed aggregation feature in each vehicle speed aggregation feature with the first acceleration aggregation feature to obtain a first vehicle speed aggregation feature group;

[0260] Replacing the first vehicle speed aggregation feature in each vehicle speed aggregation feature with the second acceleration aggregation feature to obtain a second vehicle speed aggregation feature group;

[0261] Replacing the first acceleration aggregation feature in each acceleration aggregation feature with the first velocity aggregation feature to obtain a first acceleration aggregation feature group;

[0262] Replacing the second acceleration aggregation feature in each acceleration aggregation feature with the second velocity aggregation feature to obtain a second acceleration aggregation feature group;

[0263] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature group, the second vehicle speed aggregation feature group, the first acceleration aggregation feature group, and the second acceleration aggregation feature group.

[0264] In one embodiment, when the processor executes the computer program, the following steps are further implemented: predicting the future speed of the target vehicle at a future time according to the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group, including:

[0265] Performing graph convolution processing on each feature group of the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group to obtain a feature value corresponding to each feature group;

[0266] Based on the attention function, determine the attention coefficient of each feature group;

[0267] Determine the prediction results of each feature group according to the feature values ​​and attention coefficients corresponding to each feature group;

[0268] Based on the prediction results of each feature group, the future speed of the target vehicle at the future time is predicted.

[0269] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0270] According to the global speed characteristics of the target vehicle in the historical period, the historical speed information of the target vehicle at each historical moment in the historical period is processed to obtain the speed aggregation characteristics of the target vehicle at each historical moment;

[0271] According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment;

[0272] The future vehicle speed of the target vehicle at a future time is predicted based on the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features.

[0273] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: according to the global vehicle speed characteristics of the target vehicle in the historical period, the historical vehicle speed information of the target vehicle at each historical moment in the historical period is processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment, including:

[0274] Extract the historical speed information of the target vehicle at each historical moment in the historical period to obtain the speed value at each historical moment;

[0275] According to the speed values ​​at each historical moment, determine the global speed characteristics of the target vehicle in the historical period;

[0276] According to the global vehicle speed characteristics, the vehicle speed values ​​at each historical moment are processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment.

[0277] In one embodiment, when the processor executes the computer program, the following steps are further implemented: determining the global vehicle speed characteristics of the target vehicle in the historical period according to the vehicle speed values ​​at each historical moment, including:

[0278] According to the vehicle speed values ​​at each historical moment, a historical vehicle speed matrix is ​​constructed;

[0279] Transpose the historical vehicle speed matrix to obtain a transposed matrix of the historical vehicle speed matrix;

[0280] The product of the historical speed matrix and the transposed matrix is ​​used as the global speed feature of the target vehicle in the historical period.

[0281] In one embodiment, when the processor executes the computer program, the following steps are further implemented: according to the global vehicle speed feature, the vehicle speed value at each historical moment is processed to obtain the vehicle speed aggregation feature of the target vehicle at each historical moment, including:

[0282] Preprocessing the global vehicle speed feature to obtain a preprocessed global vehicle speed feature; wherein the preprocessing at least includes normalization processing;

[0283] The preprocessed global vehicle speed features and the vehicle speed values ​​at each historical moment are input into the graph convolutional network to obtain the speed aggregation features of the target vehicle at each historical moment.

[0284] In one embodiment, according to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment, including:

[0285] Feature extraction is performed on the historical acceleration information of the target vehicle at each historical moment in the historical period to obtain the acceleration value at each historical moment;

[0286] According to the acceleration values ​​at each historical moment, the global acceleration characteristics of the target vehicle in the historical period are determined;

[0287] According to the global acceleration characteristics, the acceleration values ​​at each historical moment are processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment.

[0288] In one embodiment, when the processor executes the computer program, the following steps are further implemented: determining the global acceleration characteristics of the target vehicle in the historical period according to the acceleration values ​​at each historical moment, including:

[0289] Construct a historical acceleration matrix based on the Euclidean distance between the acceleration values ​​at each historical moment;

[0290] The historical acceleration matrix is ​​used as the global acceleration feature of the target vehicle in the historical period.

[0291] In one embodiment, when the processor executes the computer program, the following steps are further implemented: according to the global acceleration feature, the acceleration value at each historical moment is processed to obtain the acceleration aggregation feature of the target vehicle at each historical moment, including:

[0292] The global acceleration features and the acceleration values ​​at each historical moment are input into the graph convolutional network to obtain the acceleration aggregation features of the target vehicle at each historical moment;

[0293] The method further includes:

[0294] Based on the discriminator, the prediction performance is determined according to the predicted future speed of the target vehicle at the future time and the actual speed of the target vehicle at the future time;

[0295] When the prediction performance does not meet the performance requirements, the weight coefficients in the graph convolutional network are updated.

[0296] In one embodiment, when the processor executes the computer program, the following steps are further implemented: predicting the future speed of the target vehicle at a future time according to the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features, including:

[0297] According to the first similarity between the vehicle speed aggregate features, selecting the first vehicle speed aggregate feature and the second vehicle speed aggregate feature corresponding to the greatest first similarity from the vehicle speed aggregate features;

[0298] According to the second similarity between the acceleration aggregation features, the first acceleration aggregation feature and the second acceleration aggregation feature corresponding to the greatest second similarity are selected from the acceleration aggregation features;

[0299] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature, the second vehicle speed aggregation feature, the first acceleration aggregation feature and the second acceleration aggregation feature, as well as each vehicle speed aggregation feature and each acceleration aggregation feature.

[0300] In one embodiment, when the processor executes the computer program, the following steps are further implemented: predicting the future speed of the target vehicle at a future time according to the first vehicle speed aggregate feature, the second vehicle speed aggregate feature, the first acceleration aggregate feature and the second acceleration aggregate feature, and each vehicle speed aggregate feature and each acceleration aggregate feature, including:

[0301] Replacing the second vehicle speed aggregation feature in each vehicle speed aggregation feature with the first acceleration aggregation feature to obtain a first vehicle speed aggregation feature group;

[0302] Replacing the first vehicle speed aggregation feature in each vehicle speed aggregation feature with the second acceleration aggregation feature to obtain a second vehicle speed aggregation feature group;

[0303] Replacing the first acceleration aggregation feature in each acceleration aggregation feature with the first velocity aggregation feature to obtain a first acceleration aggregation feature group;

[0304] Replacing the second acceleration aggregation feature in each acceleration aggregation feature with the second velocity aggregation feature to obtain a second acceleration aggregation feature group;

[0305] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature group, the second vehicle speed aggregation feature group, the first acceleration aggregation feature group, and the second acceleration aggregation feature group.

[0306] In one embodiment, when the processor executes the computer program, the following steps are further implemented: predicting the future speed of the target vehicle at a future time according to the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group, including:

[0307] Performing graph convolution processing on each feature group of the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group to obtain a feature value corresponding to each feature group;

[0308] Based on the attention function, determine the attention coefficient of each feature group;

[0309] Determine the prediction results of each feature group according to the feature values ​​and attention coefficients corresponding to each feature group;

[0310] Based on the prediction results of each feature group, the future speed of the target vehicle at the future time is predicted.

[0311] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0312] According to the global speed characteristics of the target vehicle in the historical period, the historical speed information of the target vehicle at each historical moment in the historical period is processed to obtain the speed aggregation characteristics of the target vehicle at each historical moment;

[0313] According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment;

[0314] The future vehicle speed of the target vehicle at a future time is predicted based on the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features.

[0315] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: according to the global vehicle speed characteristics of the target vehicle in the historical period, the historical vehicle speed information of the target vehicle at each historical moment in the historical period is processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment, including:

[0316] Extract the historical speed information of the target vehicle at each historical moment in the historical period to obtain the speed value at each historical moment;

[0317] According to the speed values ​​at each historical moment, determine the global speed characteristics of the target vehicle in the historical period;

[0318] According to the global vehicle speed characteristics, the vehicle speed values ​​at each historical moment are processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment.

[0319] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: determining the global vehicle speed characteristics of the target vehicle in the historical period according to the vehicle speed values ​​at each historical moment, including:

[0320] According to the vehicle speed values ​​at each historical moment, a historical vehicle speed matrix is ​​constructed;

[0321] Transpose the historical vehicle speed matrix to obtain a transposed matrix of the historical vehicle speed matrix;

[0322] The product of the historical speed matrix and the transposed matrix is ​​used as the global speed feature of the target vehicle in the historical period.

[0323] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: according to the global vehicle speed feature, the vehicle speed value at each historical moment is processed to obtain the vehicle speed aggregation feature of the target vehicle at each historical moment, including:

[0324] Preprocessing the global vehicle speed feature to obtain a preprocessed global vehicle speed feature; wherein the preprocessing at least includes normalization processing;

[0325] The preprocessed global vehicle speed features and the vehicle speed values ​​at each historical moment are input into the graph convolutional network to obtain the speed aggregation features of the target vehicle at each historical moment.

[0326] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: according to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment, including:

[0327] Feature extraction is performed on the historical acceleration information of the target vehicle at each historical moment in the historical period to obtain the acceleration value at each historical moment;

[0328] According to the acceleration values ​​at each historical moment, the global acceleration characteristics of the target vehicle in the historical period are determined;

[0329] According to the global acceleration characteristics, the acceleration values ​​at each historical moment are processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment.

[0330] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: determining the global acceleration characteristics of the target vehicle in the historical period according to the acceleration values ​​at each historical moment, including:

[0331] Construct a historical acceleration matrix based on the Euclidean distance between the acceleration values ​​at each historical moment;

[0332] The historical acceleration matrix is ​​used as the global acceleration feature of the target vehicle in the historical period.

[0333] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: according to the global acceleration feature, the acceleration value at each historical moment is processed to obtain the acceleration aggregation feature of the target vehicle at each historical moment, including:

[0334] The global acceleration features and the acceleration values ​​at each historical moment are input into the graph convolutional network to obtain the acceleration aggregation features of the target vehicle at each historical moment;

[0335] The method further includes:

[0336] Based on the discriminator, the prediction performance is determined according to the predicted future speed of the target vehicle at the future time and the actual speed of the target vehicle at the future time;

[0337] When the prediction performance does not meet the performance requirements, the weight coefficients in the graph convolutional network are updated.

[0338] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: predicting the future speed of the target vehicle at a future time according to the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features, including:

[0339] According to the first similarity between the vehicle speed aggregate features, selecting the first vehicle speed aggregate feature and the second vehicle speed aggregate feature corresponding to the greatest first similarity from the vehicle speed aggregate features;

[0340] According to the second similarity between the acceleration aggregation features, the first acceleration aggregation feature and the second acceleration aggregation feature corresponding to the greatest second similarity are selected from the acceleration aggregation features;

[0341] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature, the second vehicle speed aggregation feature, the first acceleration aggregation feature and the second acceleration aggregation feature, as well as each vehicle speed aggregation feature and each acceleration aggregation feature.

[0342] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: predicting the future speed of the target vehicle at a future time according to the first vehicle speed aggregate feature, the second vehicle speed aggregate feature, the first acceleration aggregate feature and the second acceleration aggregate feature, and each vehicle speed aggregate feature and each acceleration aggregate feature, including:

[0343] Replacing the second vehicle speed aggregation feature in each vehicle speed aggregation feature with the first acceleration aggregation feature to obtain a first vehicle speed aggregation feature group;

[0344] Replacing the first vehicle speed aggregation feature in each vehicle speed aggregation feature with the second acceleration aggregation feature to obtain a second vehicle speed aggregation feature group;

[0345] Replacing the first acceleration aggregation feature in each acceleration aggregation feature with the first velocity aggregation feature to obtain a first acceleration aggregation feature group;

[0346] Replacing the second acceleration aggregation feature in each acceleration aggregation feature with the second velocity aggregation feature to obtain a second acceleration aggregation feature group;

[0347] The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregation feature group, the second vehicle speed aggregation feature group, the first acceleration aggregation feature group, and the second acceleration aggregation feature group.

[0348] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: predicting the future speed of the target vehicle at a future time according to the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group, including:

[0349] Performing graph convolution processing on each feature group of the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group to obtain a feature value corresponding to each feature group;

[0350] Based on the attention function, determine the attention coefficient of each feature group;

[0351] Determine the prediction results of each feature group according to the feature values ​​and attention coefficients corresponding to each feature group;

[0352] Based on the prediction results of each feature group, the future speed of the target vehicle at the future time is predicted.

[0353] It should be noted that the data involved in this application (including but not limited to messages and information in the vehicle speed prediction process) are all data that have been fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0354] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0355] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0356] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A vehicle speed prediction method, characterized in that: The method comprises: According to the global vehicle speed characteristics of the target vehicle in the historical period, the historical vehicle speed information of the target vehicle at each historical moment in the historical period is processed to obtain the vehicle speed aggregation characteristics of the target vehicle at each historical moment; According to the global acceleration characteristics of the target vehicle in the historical period, the historical acceleration information of the target vehicle at each historical moment is processed to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment; The future vehicle speed of the target vehicle at a future time is predicted based on the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features.

2. The method according to claim 1, characterized in that The process of processing the historical speed information of the target vehicle at each historical moment in the historical period according to the global speed feature of the target vehicle in the historical period to obtain the speed aggregation feature of the target vehicle at each historical moment includes: Extract the historical speed information of the target vehicle at each historical moment in the historical period to obtain the speed value at each historical moment; Determine the global vehicle speed characteristics of the target vehicle in the historical period according to the vehicle speed values ​​at each historical moment; According to the global vehicle speed feature, the vehicle speed value at each historical moment is processed to obtain the vehicle speed aggregation feature of the target vehicle at each historical moment.

3. The method according to claim 2, characterized in that Determining the global vehicle speed characteristics of the target vehicle in the historical period according to the vehicle speed values ​​at each historical moment includes: According to the vehicle speed values ​​at each historical moment, a historical vehicle speed matrix is ​​constructed; Transposing the historical vehicle speed matrix to obtain a transposed matrix of the historical vehicle speed matrix; The product of the historical vehicle speed matrix and the transposed matrix is ​​used as the global vehicle speed feature of the target vehicle in the historical period.

4. The method according to claim 2, characterized in that: The process of processing the vehicle speed values ​​at each historical moment according to the global vehicle speed feature to obtain the vehicle speed aggregation feature of the target vehicle at each historical moment includes: Preprocessing the global vehicle speed feature to obtain a preprocessed global vehicle speed feature; wherein the preprocessing at least includes normalization processing; The preprocessed global vehicle speed features and the vehicle speed values ​​at each historical moment are input into the graph convolutional network to obtain the vehicle speed aggregation features of the target vehicle at each historical moment.

5. The method according to claim 1, characterized in that The process of processing the historical acceleration information of the target vehicle at each historical moment according to the global acceleration characteristics of the target vehicle in the historical period to obtain the acceleration aggregation characteristics of the target vehicle at each historical moment includes: Feature extraction is performed on the historical acceleration information of the target vehicle at each historical moment in the historical period to obtain the acceleration value at each historical moment; Determining the global acceleration characteristics of the target vehicle in the historical period according to the acceleration values ​​at each historical moment; According to the global acceleration feature, the acceleration value at each historical moment is processed to obtain the acceleration aggregation feature of the target vehicle at each historical moment.

6. The method according to claim 5, characterized in that Determining the global acceleration characteristics of the target vehicle in the historical period according to the acceleration values ​​at each historical moment includes: Construct a historical acceleration matrix based on the Euclidean distance between the acceleration values ​​at each historical moment; The historical acceleration matrix is ​​used as the global acceleration feature of the target vehicle in the historical period.

7. The method according to claim 5, characterized in that The step of processing the acceleration values ​​at each historical moment according to the global acceleration feature to obtain the acceleration aggregation feature of the target vehicle at each historical moment includes: Inputting the global acceleration feature and the acceleration value at each historical moment into a graph convolutional network to obtain the acceleration aggregation feature of the target vehicle at each historical moment; Accordingly, the method further includes: Determining prediction performance based on the discriminator according to the predicted future speed of the target vehicle at the future time and the actual speed of the target vehicle at the future time; When the prediction performance does not meet the performance requirement, the weight coefficients in the graph convolutional network are updated.

8. The method according to claim 1, characterized in that The predicting of the future speed of the target vehicle at a future time according to the first similarity between the vehicle speed aggregate features and the vehicle speed aggregate features, and the second similarity between the acceleration aggregate features and the acceleration aggregate features, comprises: According to the first similarity between the vehicle speed aggregate features, selecting the first vehicle speed aggregate feature and the second vehicle speed aggregate feature corresponding to the greatest first similarity from the vehicle speed aggregate features; According to the second similarity between the acceleration aggregation features, the first acceleration aggregation feature and the second acceleration aggregation feature corresponding to the greatest second similarity are selected from the acceleration aggregation features; The future vehicle speed of the target vehicle at a future moment is predicted based on the first vehicle speed aggregation feature, the second vehicle speed aggregation feature, the first acceleration aggregation feature, the second acceleration aggregation feature, and each vehicle speed aggregation feature and each acceleration aggregation feature.

9. The method according to claim 8, characterized in that The predicting the future vehicle speed of the target vehicle at a future time according to the first vehicle speed aggregate feature, the second vehicle speed aggregate feature, the first acceleration aggregate feature, the second acceleration aggregate feature, and each vehicle speed aggregate feature and each acceleration aggregate feature includes: Replacing the second vehicle speed aggregation feature in each vehicle speed aggregation feature with the first acceleration aggregation feature to obtain a first vehicle speed aggregation feature group; Replacing the first vehicle speed aggregate feature in each vehicle speed aggregate feature with the second acceleration aggregate feature to obtain a second vehicle speed aggregate feature group; Replacing the first acceleration aggregation feature in each acceleration aggregation feature with the first speed aggregation feature to obtain a first acceleration aggregation feature group; Replacing the second acceleration aggregation feature in each acceleration aggregation feature with the second speed aggregation feature to obtain a second acceleration aggregation feature group; The future vehicle speed of the target vehicle at a future time is predicted based on the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group.

10. The method according to claim 9, characterized in that The predicting the future vehicle speed of the target vehicle at a future time according to the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group includes: Performing graph convolution processing on each feature group of the first vehicle speed aggregate feature group, the second vehicle speed aggregate feature group, the first acceleration aggregate feature group, and the second acceleration aggregate feature group to obtain a feature value corresponding to each feature group; Based on the attention function, determine the attention coefficient of each feature group; Determine the prediction results of each feature group according to the feature values ​​and attention coefficients corresponding to each feature group; The future speed of the target vehicle at a future time is predicted based on the prediction results of each feature group.