Yaw rate prediction method, device, equipment, readable storage medium and program product
By combining route characteristics and section feature data, the progressively trained yaw rate prediction model is used to improve the accuracy of route yaw rate prediction in the navigation system, and the problem of insufficient yaw rate prediction in the navigation system is solved, and the problem of insufficient yaw rate prediction in the navigation system is improved. User experience and service quality.
Patent Information
- Application Number
- CN202210334630.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-30
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2042-03-30
AI Technical Summary
The accuracy of route yaw rate prediction in the existing navigation system is insufficient, which affects the navigation experience and service quality.
By using a yaw rate prediction method, by determining the target route and the pre-trained yaw rate prediction model, the first yaw rate prediction sub-model and the second yaw rate prediction sub-model are used, combining the route characteristics and section feature data, Perform progressive training to improve prediction accuracy.
It effectively improves the accuracy of route yaw rate prediction and improves the user experience and service quality of the navigation platform.
Smart Images

Figure CN114742291B_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present disclosure relate to the technical field of traffic data processing, and in particular to a yaw rate prediction method, device, equipment, readable storage medium, and program product. Background Art
[0002] With the development and progress of science and technology and data technology, the demand for accurate data prediction is becoming more and more widespread. For example, in the field of navigation, the route deviation rate refers to the probability that a user will deviate from a certain route. The lower the route deviation rate, the more the user prefers this route or the more the user confirms that there is no problem with this route. Subsequently, the planning and presentation of navigation routes can be implemented based on routes with lower route deviation rates, thereby improving the user's navigation experience and the service quality of the navigation platform. Therefore, the accuracy of route deviation rate prediction plays a very important role in the planning of navigation routes. Summary of the invention
[0003] Embodiments of the present disclosure provide a yaw rate prediction method, apparatus, device, readable storage medium, and program product.
[0004] In a first aspect, a yaw rate prediction method is provided in an embodiment of the present disclosure.
[0005] Specifically, the yaw rate prediction method includes:
[0006] Determine a target route and a pre-trained yaw rate prediction model, wherein the yaw rate prediction model includes a first yaw rate prediction sub-model and a second yaw rate prediction sub-model;
[0007] Acquire first feature data and second feature data of the target route, and input the second feature data into a second yaw rate prediction sub-model to obtain third feature data, wherein the first feature data includes route features and / or navigation object features, the second feature data is segment feature data of segments included in the route, represented as an N×M feature matrix, wherein N represents the number of segments, and M represents the number of attributes of a certain segment, and the third feature data is an implicit feature obtained based on the second feature data;
[0008] The first characteristic data and the third characteristic data of the target route are input into the first yaw rate prediction sub-model to obtain a yaw rate prediction result for the target route.
[0009] In combination with the first aspect, in a first implementation manner of the first aspect of the embodiment of the present disclosure, the method further includes:
[0010] Determine a yaw rate prediction model to be trained that includes a first initial yaw rate prediction sub-model and a second initial yaw rate prediction sub-model;
[0011] Acquire a training data set, wherein the training data set includes first feature data, second feature data of the training data, and label data corresponding to the training data;
[0012] The yaw rate prediction model to be trained is progressively trained based on the training data set to obtain a trained yaw rate prediction model.
[0013] In combination with the first aspect and the first implementation manner of the first aspect, in a second implementation manner of the first aspect of the embodiment of the present disclosure, the progressive training of the yaw rate prediction model to be trained based on the training data set includes:
[0014] The second feature data of the training data is used as the input of the second initial yaw rate prediction sub-model, the output of the second initial yaw rate prediction sub-model and the first feature data of the training data are used together as the input of the first initial yaw rate prediction sub-model, the label data corresponding to the training data is used as the output of the first initial yaw rate prediction sub-model, the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model are trained, and a yaw rate prediction model consisting of the first yaw rate prediction sub-model and the second yaw rate prediction sub-model is obtained.
[0015] In combination with the first aspect and the above implementation of the first aspect, in a third implementation of the first aspect of the present disclosure, the method further includes:
[0016] The first characteristic data and the second characteristic data of the target route and the corresponding yaw rate prediction result data are added to the training data set as new training data to train the yaw rate prediction model.
[0017] In combination with the first aspect and the above-mentioned implementation manner of the first aspect, in a fourth implementation manner of the first aspect of the present disclosure, the first yaw rate prediction sub-model is a deep learning model, and the second yaw rate prediction sub-model is a convolutional model.
[0018] In a second aspect, a yaw rate prediction device is provided in an embodiment of the present disclosure.
[0019] Specifically, the yaw rate prediction device includes:
[0020] A determination module is configured to determine a target route and a pre-trained yaw rate prediction model, wherein the yaw rate prediction model includes a first yaw rate prediction sub-model and a second yaw rate prediction sub-model;
[0021] an acquisition module, configured to acquire first feature data and second feature data of the target route, and input the second feature data into a second yaw rate prediction sub-model to obtain third feature data, wherein the first feature data includes route features and / or navigation object features, the second feature data is segment feature data of a segment included in the route, represented as an N×M feature matrix, wherein N represents the number of segments, and M represents the number of attributes of a certain segment, and the third feature data is an implicit feature obtained based on the second feature data;
[0022] The prediction module is configured to input the first characteristic data and the third characteristic data of the target route into the first yaw rate prediction sub-model to obtain a yaw rate prediction result for the target route.
[0023] In combination with the second aspect, in a first implementation manner of the second aspect of the embodiment of the present disclosure, the apparatus further includes a training module, and the training module is configured to:
[0024] Determine a yaw rate prediction model to be trained that includes a first initial yaw rate prediction sub-model and a second initial yaw rate prediction sub-model;
[0025] Acquire a training data set, wherein the training data set includes first feature data, second feature data of the training data, and label data corresponding to the training data;
[0026] The yaw rate prediction model to be trained is progressively trained based on the training data set to obtain a trained yaw rate prediction model.
[0027] In a third aspect, an embodiment of the present disclosure provides an electronic device, including a memory and a processor, wherein the memory is used to store one or more computer instructions that support a yaw rate prediction device to perform the above-mentioned yaw rate prediction method, and the processor is configured to execute the computer instructions stored in the memory. The yaw rate prediction device may also include a communication interface for the yaw rate prediction device to communicate with other devices or a communication network.
[0028] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium for storing computer instructions used by a yaw rate prediction device, which includes computer instructions involved in the yaw rate prediction device for executing the above-mentioned yaw rate prediction method.
[0029] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including a computer program / instruction, wherein the computer program / instruction implements the steps of the above-mentioned yaw rate prediction method when executed by a processor.
[0030] The technical solution provided by the embodiments of the present disclosure may have the following beneficial effects:
[0031] The above technical solution is based on the additional feature data obtained from the second yaw rate prediction sub-model in the pre-trained yaw rate prediction model, combined with the basic feature data of the target route, to jointly participate in the prediction of the preset indicators of the target route. This technical solution can effectively improve the accuracy of the prediction of indicators such as the route yaw rate, thereby improving the user experience and the service quality of the data platform.
[0032] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the embodiments of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Other features, purposes and advantages of the embodiments of the present disclosure will become more apparent through the following detailed description of non-limiting embodiments in conjunction with the accompanying drawings. In the accompanying drawings:
[0034] Figure 1 A flowchart showing a yaw rate prediction method according to an embodiment of the present disclosure is shown;
[0035] Figure 2 An overall flow chart of a yaw rate prediction method according to an embodiment of the present disclosure is shown;
[0036] Figure 3 A structural block diagram of a yaw rate prediction device according to an embodiment of the present disclosure is shown;
[0037] Figure 4 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown;
[0038] Figure 5 It is a structural schematic diagram of a computer system suitable for implementing a yaw rate prediction method according to an embodiment of the present disclosure. DETAILED DESCRIPTION
[0039] Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily implement them. In addition, for the sake of clarity, parts not related to the description of the exemplary embodiments are omitted in the accompanying drawings.
[0040] In the embodiments of the present disclosure, it should be understood that terms such as "including" or "having" are intended to indicate the existence of features, numbers, steps, behaviors, components, parts, or a combination thereof disclosed in this specification, and are not intended to exclude the possibility of one or more other features, numbers, steps, behaviors, components, parts, or a combination thereof existing or being added.
[0041] It should also be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other. The embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0042] The technical solution provided by the embodiment of the present disclosure is based on the additional feature data obtained from the second yaw rate prediction sub-model in the pre-trained yaw rate prediction model, combined with the basic feature data of the target route, to jointly participate in the prediction of the preset indicators of the target route. This technical solution can effectively improve the accuracy of the prediction of indicators such as the route yaw rate, thereby improving the user experience and the service quality of the data platform.
[0043] Figure 1 A flow chart of a yaw rate prediction method according to an embodiment of the present disclosure is shown as follows: Figure 1 As shown, the yaw rate prediction method includes the following steps S101-S103:
[0044] In step S101, a target route and a pre-trained yaw rate prediction model are determined, wherein the yaw rate prediction model includes a first yaw rate prediction sub-model and a second yaw rate prediction sub-model;
[0045] In step S102, first feature data and second feature data of the target route are obtained, and the second feature data are input into a second yaw rate prediction sub-model to obtain third feature data, wherein the first feature data includes route features and / or navigation object features, the second feature data is segment feature data of the segments included in the route, represented as an N×M feature matrix, wherein N represents the number of segments, and M represents the number of attributes of a certain segment, and the third feature data is an implicit feature obtained based on the second feature data;
[0046] In step S103, the first characteristic data and the third characteristic data of the target route are input into the first yaw rate prediction sub-model to obtain a yaw rate prediction result for the target route.
[0047] As mentioned above, with the development and progress of science and technology and data technology, the demand for accurate data prediction is becoming more and more widespread. For example, in the field of navigation, the route deviation rate refers to the probability that a user will deviate from a certain route. The lower the route deviation rate, the more the user prefers this route or the more the user confirms that there is no problem with this route. Subsequently, the planning and presentation of navigation routes can be implemented based on routes with lower route deviation rates, thereby improving the user's navigation experience and the service quality of the navigation platform. Therefore, the accuracy of route deviation rate prediction plays a very important role in the planning of navigation routes.
[0048] In view of the above problems, in this embodiment, a yaw rate prediction method is proposed. The method is based on the additional feature data obtained from the second yaw rate prediction sub-model in the pre-trained yaw rate prediction model, combined with the basic feature data of the target route, to jointly participate in the prediction of the preset indicators of the target route. This technical solution can effectively improve the accuracy of the prediction of indicators such as the route yaw rate, thereby improving the user experience and the service quality of the data platform.
[0049] In one embodiment of the present disclosure, the yaw rate prediction method may be applicable to a computer, a computing device, an electronic device, a server, a service cluster, etc. that can perform yaw rate prediction.
[0050] In one embodiment of the present disclosure, the target route refers to a route consisting of multiple road sections for which yaw rate prediction is required.
[0051] In one embodiment of the present disclosure, the yaw rate prediction model refers to a model that is pre-trained based on training data and can realize the function of predicting the yaw rate of a certain route, and the extracted characteristic data of the target route is input into the yaw rate prediction model to obtain a yaw rate prediction result for the target route. The yaw rate prediction model includes a first yaw rate prediction sub-model and a second yaw rate prediction sub-model, the first yaw rate prediction sub-model is used to obtain the yaw rate prediction result of the target route according to the characteristic data of the target route, and the second yaw rate prediction sub-model is used to obtain part of the characteristic data of the target route input into the first yaw rate prediction sub-model.
[0052] In one embodiment of the present disclosure, the characteristic data of the target route refers to data that can reflect the characteristics of the target route. In one embodiment of the present disclosure, the characteristic data of the target route includes first characteristic data and second characteristic data, wherein the first characteristic data is data that can reflect one or several aspects of the characteristics of the target route, the second characteristic data is data used to obtain the third characteristic data that can reflect another or several aspects of the characteristics of the target route, and the third characteristic data is obtained based on the second characteristic data, and can be considered as data that can reflect some implicit characteristics of the target route, that is, the first characteristic data, the second characteristic data and the third characteristic data are all characteristic data of the target route, but they respectively reflect different aspects of the characteristics of the target route.
[0053] For example, if the target route is a route composed of multiple road sections, the first characteristic data of the target route may be the basic characteristic data of the target route. For example, the first characteristic data of the target route may include one or more of the following characteristic data: route-level characteristic data such as the estimated arrival time of the target route, the length of the target route, and the congestion characteristics of the target route, i.e., route characteristics, and user-level characteristic data such as user portrait information and user preference information, i.e., navigation object characteristics. The second characteristic data of the target route may be segment-level characteristic data, i.e., segment characteristic data of the segments included in the target route. For example, assuming that the target route includes N segments, the segment sequence may be expressed as L=(l1,l2,...,l N ), each road segment contains M attribute data, and the attribute data vector can be expressed as f i =(x1,x2,...,x M ), wherein i represents the ith road section, and the M attribute data include static attribute data, dynamic attribute data and position attribute data, wherein the static attribute data may include data such as section length, section width, number of section lanes, and section road grade, the dynamic attribute data may include data such as congestion conditions on the section, traffic incident conditions on the section, and estimated arrival time of the section, and the position attribute data may include data such as the cumulative length and cumulative estimated arrival time from the first section to the current section in the target route, thereby, the second feature data of the target route can be represented as an N×M section-level feature matrix F, wherein each row of the feature matrix F represents a section in the target route, i.e., N represents the number of sections, and each column represents an attribute of the section, i.e., M represents the number of attributes of a certain section.
[0054] The first characteristic data and its type, the second characteristic data and its type of the target route can be determined according to the needs of actual application, and the present disclosure does not specifically limit the specific content and type of the characteristic data.
[0055] In the above embodiment, the target route for which the yaw rate is to be predicted and the pre-trained yaw rate prediction model can be first determined; then the first feature data and the second feature data of the target route are extracted; the second feature data is input into the second yaw rate prediction sub-model in the yaw rate prediction model to obtain the third feature data; finally, the first feature data and the third feature data of the target route are input into the first yaw rate prediction sub-model together to jointly obtain the yaw rate prediction result for the target route.
[0056] In one embodiment of the present disclosure, the method may further include the following steps:
[0057] The yaw rate prediction model is trained.
[0058] In one embodiment of the present disclosure, the step of training the yaw rate prediction model may include the following steps:
[0059] Determine a yaw rate prediction model to be trained that includes a first initial yaw rate prediction sub-model and a second initial yaw rate prediction sub-model;
[0060] Acquire a training data set, wherein the training data set includes first feature data, second feature data of the training data, and label data corresponding to the training data;
[0061] The yaw rate prediction model to be trained is progressively trained based on the training data set to obtain a trained yaw rate prediction model.
[0062] In this embodiment, when training the yaw rate prediction model, firstly, a yaw rate prediction model to be trained is determined, which includes a first initial yaw rate prediction sub-model and a second initial yaw rate prediction sub-model, wherein the yaw rate prediction model prototype can be selected according to the needs of actual applications, for example, the first initial yaw rate prediction sub-model can be selected as a deep neural network (DNN), a feature cross network (DCN) or other deep learning models, and the second initial yaw rate prediction sub-model can be selected as a multi-layer convolutional neural network (CNN), ResNet50 or other convolutional models, and the model parameters of the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model need to be determined with the help of subsequent training.
[0063] Then, a training data set including first feature data of the training data, second feature data and label data corresponding to the training data is obtained, wherein the label data corresponding to the training data refers to known yaw rate indicator data corresponding to the training data. For example, if the training data is route data, the label data corresponding to the training data refers to the actual yaw rate corresponding to the route data.
[0064] Then, based on the training data set, the yaw rate prediction model to be trained, which is composed of the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model, is progressively trained to obtain a trained yaw rate prediction model, that is, to obtain the trained first yaw rate prediction sub-model and the second yaw rate prediction sub-model.
[0065] In one embodiment of the present disclosure, the step of performing progressive training on the yaw rate prediction model to be trained based on the training data set may include the following steps:
[0066] The second feature data of the training data is used as the input of the second initial yaw rate prediction sub-model, the output of the second initial yaw rate prediction sub-model and the first feature data of the training data are used together as the input of the first initial yaw rate prediction sub-model, the label data corresponding to the training data is used as the output of the first initial yaw rate prediction sub-model, the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model are trained, and a yaw rate prediction model consisting of the first yaw rate prediction sub-model and the second yaw rate prediction sub-model is obtained.
[0067] When the yaw rate prediction model to be trained is progressively trained, the yaw rate prediction model including the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model is considered as a whole for training, and the progressive training means that the output of the second initial yaw rate prediction sub-model is used as one of the training inputs of the first initial yaw rate prediction sub-model. Specifically, when the yaw rate prediction model to be trained is progressively trained, the second feature data of the training data is used as the training input of the second initial yaw rate prediction sub-model, the output of the second initial yaw rate prediction sub-model is used as the third feature data of the training data, and the first feature data of the training data is used as the common input of the first initial yaw rate prediction sub-model, and the label data corresponding to the training data is used as the output of the first initial yaw rate prediction sub-model to train the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model. When the training results converge, the yaw rate prediction model can be obtained, and the first yaw rate prediction sub-model and the second yaw rate prediction sub-model that have been trained are obtained.
[0068] In one embodiment of the present disclosure, the step of training the yaw rate prediction model may further include the following steps:
[0069] The first characteristic data and the second characteristic data of the target route and the corresponding yaw rate prediction result data are added to the training data set as new training data to train the yaw rate prediction model.
[0070] In order to improve the completeness of the training data set of the yaw rate prediction model and ensure the comprehensiveness of the yaw rate prediction learning and training results, in this embodiment, a feedback mechanism is used to perform yaw rate prediction, that is, after the yaw rate prediction result is obtained based on the target route using the yaw rate prediction model, the target route and the corresponding yaw rate prediction result data are added as new training data to the training data set of the yaw rate prediction model, and then the yaw rate prediction model is subjected to the above-mentioned progressive training to enrich the training data, improve the accuracy of the yaw rate prediction, and obtain a more complete yaw rate prediction model to participate in the output of the next yaw rate prediction result.
[0071] Figure 2 FIG. 1 shows an overall flow chart of a yaw rate prediction method according to an embodiment of the present disclosure, as shown in FIG. Figure 2 As shown, first determine the yaw rate prediction model to be trained, which includes the first initial yaw rate prediction submodel and the second initial yaw rate prediction submodel. Then obtain the basic features of the training route, such as route features such as the estimated arrival time, length, and congestion characteristics, and navigation object features such as user portrait information and user preference information, as the first feature data; obtain the segment feature matrix F of the training route segment level with a size of N×M as the second feature data; and obtain the label data of the training route. During training, the second feature data is used as the input of the second initial yaw rate prediction submodel, the output of the second initial yaw rate prediction submodel, that is, the implicit feature of the training route, that is, the third feature data, and the first feature data are used as the input of the first initial yaw rate prediction submodel, and the label data of the training route is used as the output of the first initial yaw rate prediction submodel. The first initial yaw rate prediction submodel and the second initial yaw rate prediction submodel are trained to obtain the first yaw rate prediction submodel and the second yaw rate prediction submodel, and then the yaw rate prediction model is obtained. When in use, the target route for which the yaw rate prediction is required can be first determined, and then the first characteristic data and the second characteristic data of the target route can be obtained, and the second characteristic data can be input into the second yaw rate prediction sub-model to obtain the third characteristic data, and then the first characteristic data and the third characteristic data can be input into the first yaw rate prediction sub-model together to obtain the yaw rate prediction result for the target route.
[0072] The following are embodiments of the apparatus of the present disclosure, which can be used to execute embodiments of the method of the present disclosure.
[0073] Figure 3 FIG. 1 is a block diagram of a yaw rate prediction device according to an embodiment of the present disclosure. The device can be implemented as part or all of an electronic device through software, hardware, or a combination of both. Figure 3 As shown, the yaw rate prediction device includes:
[0074] A determination module 301 is configured to determine a target route and a pre-trained yaw rate prediction model, wherein the yaw rate prediction model includes a first yaw rate prediction sub-model and a second yaw rate prediction sub-model;
[0075] The acquisition module 302 is configured to acquire first feature data and second feature data of the target route, and input the second feature data into a second yaw rate prediction sub-model to obtain third feature data, wherein the first feature data includes route features and / or navigation object features, the second feature data is segment feature data of a segment included in the route, represented as an N×M feature matrix, wherein N represents the number of segments, and M represents the number of attributes of a certain segment, and the third feature data is an implicit feature obtained based on the second feature data;
[0076] The prediction module 303 is configured to input the first characteristic data and the third characteristic data of the target route into the first yaw rate prediction sub-model to obtain a yaw rate prediction result for the target route.
[0077] As mentioned above, with the development and progress of science and technology and data technology, the demand for accurate data prediction is becoming more and more widespread. For example, in the field of navigation, the route deviation rate refers to the probability that a user will deviate from a certain route. The lower the route deviation rate, the more the user prefers this route or the more the user confirms that there is no problem with this route. Subsequently, the planning and presentation of navigation routes can be implemented based on routes with lower route deviation rates, thereby improving the user's navigation experience and the service quality of the navigation platform. Therefore, the accuracy of route deviation rate prediction plays a very important role in the planning of navigation routes.
[0078] In view of the above problems, in this embodiment, a yaw rate prediction device is proposed, which is based on the additional feature data obtained from the second yaw rate prediction sub-model in the pre-trained yaw rate prediction model, combined with the basic feature data of the target route, to jointly participate in the prediction of the preset indicators of the target route. This technical solution can effectively improve the accuracy of the prediction of indicators such as the route yaw rate, thereby improving the user experience and the service quality of the data platform.
[0079] In one embodiment of the present disclosure, the yaw rate prediction device may be implemented as a computer, a computing device, an electronic device, a server, a service cluster, etc. that can perform yaw rate prediction.
[0080] In one embodiment of the present disclosure, the target route refers to a route consisting of multiple road sections for which yaw rate prediction is required.
[0081] In one embodiment of the present disclosure, the yaw rate prediction model refers to a model that is pre-trained based on training data and can realize the function of predicting the yaw rate of a certain route, and the extracted characteristic data of the target route is input into the yaw rate prediction model to obtain a yaw rate prediction result for the target route. The yaw rate prediction model includes a first yaw rate prediction sub-model and a second yaw rate prediction sub-model, the first yaw rate prediction sub-model is used to obtain the yaw rate prediction result of the target route according to the characteristic data of the target route, and the second yaw rate prediction sub-model is used to obtain part of the characteristic data of the target route input into the first yaw rate prediction sub-model.
[0082] In one embodiment of the present disclosure, the characteristic data of the target route refers to data that can reflect the characteristics of the target route. In one embodiment of the present disclosure, the characteristic data of the target route includes first characteristic data and second characteristic data, wherein the first characteristic data is data that can reflect one or several aspects of the characteristics of the target route, the second characteristic data is data used to obtain the third characteristic data that can reflect another or several aspects of the characteristics of the target route, and the third characteristic data is obtained based on the second characteristic data, and can be considered as data that can reflect some implicit characteristics of the target route, that is, the first characteristic data, the second characteristic data and the third characteristic data are all characteristic data of the target route, but they respectively reflect different aspects of the characteristics of the target route.
[0083] For example, if the target route is a route composed of multiple road sections, the first characteristic data of the target route may be the basic characteristic data of the target route. For example, the first characteristic data of the target route may include one or more of the following characteristic data: route-level characteristic data such as the estimated arrival time of the target route, the length of the target route, and the congestion characteristics of the target route, i.e., route characteristics, and user-level characteristic data such as user portrait information and user preference information, i.e., navigation object characteristics. The second characteristic data of the target route may be segment-level characteristic data, i.e., segment characteristic data of the segments included in the target route. For example, assuming that the target route includes N segments, the segment sequence may be expressed as L=(l1,l2,...,l N ), each road segment contains M attribute data, and the attribute data vector can be expressed as f i =(x1,x2,...,x M), wherein i represents the ith road section, and the M attribute data include static attribute data, dynamic attribute data and position attribute data, wherein the static attribute data may include data such as section length, section width, number of section lanes, and section road grade, the dynamic attribute data may include data such as congestion conditions on the section, traffic incident conditions on the section, and estimated arrival time of the section, and the position attribute data may include data such as the cumulative length and cumulative estimated arrival time from the first section to the current section in the target route, thereby, the second feature data of the target route can be represented as an N×M section-level feature matrix F, wherein each row of the feature matrix F represents a section in the target route, i.e., N represents the number of sections, and each column represents an attribute of the section, i.e., M represents the number of attributes of a certain section.
[0084] The first characteristic data and its type, the second characteristic data and its type of the target route can be determined according to the needs of actual application, and the present disclosure does not specifically limit the specific content and type of the characteristic data.
[0085] In the above embodiment, the target route for which the yaw rate is to be predicted and the pre-trained yaw rate prediction model can be first determined; then the first feature data and the second feature data of the target route are extracted; the second feature data is input into the second yaw rate prediction sub-model in the yaw rate prediction model to obtain the third feature data; finally, the first feature data and the third feature data of the target route are input into the first yaw rate prediction sub-model together to jointly obtain the yaw rate prediction result for the target route.
[0086] In one embodiment of the present disclosure, the device may further include a training module, and the training module may be configured as follows:
[0087] The yaw rate prediction model is trained.
[0088] In one embodiment of the present disclosure, the training module may be further configured as follows:
[0089] Determine a yaw rate prediction model to be trained that includes a first initial yaw rate prediction sub-model and a second initial yaw rate prediction sub-model;
[0090] Acquire a training data set, wherein the training data set includes first feature data, second feature data of the training data, and label data corresponding to the training data;
[0091] The yaw rate prediction model to be trained is progressively trained based on the training data set to obtain a trained yaw rate prediction model.
[0092] In this embodiment, when training the yaw rate prediction model, firstly, a yaw rate prediction model to be trained is determined, which includes a first initial yaw rate prediction sub-model and a second initial yaw rate prediction sub-model, wherein the yaw rate prediction model prototype can be selected according to the needs of actual applications, for example, the first initial yaw rate prediction sub-model can be selected as a deep neural network (DNN), a feature cross network (DCN) or other deep learning models, and the second initial yaw rate prediction sub-model can be selected as a multi-layer convolutional neural network (CNN), ResNet50 or other convolutional models, and the model parameters of the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model need to be determined with the help of subsequent training.
[0093] Then, a training data set including first feature data of the training data, second feature data and label data corresponding to the training data is obtained, wherein the label data corresponding to the training data refers to known yaw rate indicator data corresponding to the training data. For example, if the training data is route data, the label data corresponding to the training data refers to the actual yaw rate corresponding to the route data.
[0094] Then, based on the training data set, the yaw rate prediction model to be trained, which is composed of the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model, is progressively trained to obtain a trained yaw rate prediction model, that is, to obtain the trained first yaw rate prediction sub-model and the second yaw rate prediction sub-model.
[0095] In one embodiment of the present disclosure, the part of performing progressive training on the yaw rate prediction model to be trained based on the training data set may be configured as follows:
[0096] The second feature data of the training data is used as the input of the second initial yaw rate prediction sub-model, the output of the second initial yaw rate prediction sub-model and the first feature data of the training data are used together as the input of the first initial yaw rate prediction sub-model, the label data corresponding to the training data is used as the output of the first initial yaw rate prediction sub-model, the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model are trained, and a yaw rate prediction model consisting of the first yaw rate prediction sub-model and the second yaw rate prediction sub-model is obtained.
[0097] When the yaw rate prediction model to be trained is progressively trained, the yaw rate prediction model including the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model is considered as a whole for training, and the progressive training means that the output of the second initial yaw rate prediction sub-model is used as one of the training inputs of the first initial yaw rate prediction sub-model. Specifically, when the yaw rate prediction model to be trained is progressively trained, the second feature data of the training data is used as the training input of the second initial yaw rate prediction sub-model, the output of the second initial yaw rate prediction sub-model is used as the third feature data of the training data, and the first feature data of the training data is used as the common input of the first initial yaw rate prediction sub-model, and the label data corresponding to the training data is used as the output of the first initial yaw rate prediction sub-model to train the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model. When the training results converge, the yaw rate prediction model can be obtained, and the first yaw rate prediction sub-model and the second yaw rate prediction sub-model that have been trained are obtained.
[0098] In one embodiment of the present disclosure, the training module may also be configured as follows:
[0099] The first characteristic data and the second characteristic data of the target route and the corresponding yaw rate prediction result data are added to the training data set as new training data to train the yaw rate prediction model.
[0100] In order to improve the completeness of the training data set of the yaw rate prediction model and ensure the comprehensiveness of the yaw rate prediction learning and training results, in this embodiment, a feedback mechanism is used to perform yaw rate prediction, that is, after the yaw rate prediction result is obtained based on the target route using the yaw rate prediction model, the target route and the corresponding yaw rate prediction result data are added as new training data to the training data set of the yaw rate prediction model, and then the yaw rate prediction model is subjected to the above-mentioned progressive training to enrich the training data, improve the accuracy of the yaw rate prediction, and obtain a more complete yaw rate prediction model to participate in the output of the next yaw rate prediction result.
[0101] The present disclosure also discloses an electronic device, Figure 4 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown. Figure 4 As shown, the electronic device 400 includes a memory 401 and a processor 402; wherein,
[0102] The memory 401 is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor 402 to implement the above method steps.
[0103] Figure 5 It is a structural schematic diagram of a computer system suitable for implementing a yaw rate prediction method according to an embodiment of the present disclosure.
[0104] like Figure 5 As shown, the computer system 500 includes a processing unit 501, which can perform various processes in the above-mentioned embodiments according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage part 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the computer system 500 are also stored. The processing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.
[0105] The following components are connected to the I / O interface 505: an input part 506 including a keyboard, a mouse, etc.; an output part 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage part 508 including a hard disk, etc.; and a communication part 509 including a network interface card such as a LAN card, a modem, etc. The communication part 509 performs communication processing via a network such as the Internet. The drive 510 is also connected to the I / O interface 505 as needed. Removable media 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., are installed on the drive 510 as needed, so that the computer program read therefrom is installed into the storage part 508 as needed. Among them, the processing unit 501 can be implemented as a processing unit such as a CPU, a GPU, a TPU, an FPGA, an NPU, etc.
[0106] In particular, according to an embodiment of the present disclosure, the method described above may be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program tangibly contained on a readable medium thereof, the computer program including a program code for executing the yaw rate prediction method. In such an embodiment, the computer program may be downloaded and installed from a network through the communication part 509, and / or installed from a removable medium 511.
[0107] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the road map or block diagram can represent a part of a module, program segment or code, and the part of the module, program segment or code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a different order from the order marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0108] The units or modules involved in the embodiments described in the present disclosure may be implemented by software or hardware. The units or modules described may also be set in a processor, and the names of these units or modules do not constitute limitations on the units or modules themselves in some cases.
[0109] As another aspect, the embodiments of the present disclosure further provide a computer-readable storage medium, which may be a computer-readable storage medium included in the device described in the above-mentioned embodiment; or a computer-readable storage medium that exists independently and is not assembled into the device. The computer-readable storage medium stores one or more programs, and the programs are used by one or more processors to execute the methods described in the embodiments of the present disclosure.
[0110] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure.
Claims
1. A yaw rate prediction method, comprising: Determine a target route and a pre-trained yaw rate prediction model, wherein the yaw rate prediction model includes a first yaw rate prediction sub-model and a second yaw rate prediction sub-model; Acquire first characteristic data and second characteristic data of the target route; Inputting the second feature data into a second yaw rate prediction sub-model to obtain third feature data, wherein the first feature data includes route features and / or navigation object features, the second feature data is segment feature data of a segment included in the route, represented as an N×M feature matrix, wherein N represents the number of segments, and M represents the number of attributes of a certain segment, and the third feature data is data that can reflect some features of the target route obtained based on the second feature data; The first characteristic data and the third characteristic data of the target route are input into the first yaw rate prediction sub-model to obtain a yaw rate prediction result for the target route.
2. The method according to claim 1, further comprising: Determine a yaw rate prediction model to be trained that includes a first initial yaw rate prediction sub-model and a second initial yaw rate prediction sub-model; Acquire a training data set, wherein the training data set includes first feature data, second feature data of the training data, and label data corresponding to the training data; The yaw rate prediction model to be trained is progressively trained based on the training data set to obtain a trained yaw rate prediction model.
3. The method according to claim 2, wherein the step of performing progressive training on the yaw rate prediction model to be trained based on the training data set comprises: The second feature data of the training data is used as the input of the second initial yaw rate prediction sub-model, the output of the second initial yaw rate prediction sub-model and the first feature data of the training data are used together as the input of the first initial yaw rate prediction sub-model, the label data corresponding to the training data is used as the output of the first initial yaw rate prediction sub-model, the first initial yaw rate prediction sub-model and the second initial yaw rate prediction sub-model are trained, and a yaw rate prediction model consisting of the first yaw rate prediction sub-model and the second yaw rate prediction sub-model is obtained.
4. The method according to claim 2 or 3, further comprising: The first characteristic data and the second characteristic data of the target route and the corresponding yaw rate prediction result data are added to the training data set as new training data to train the yaw rate prediction model.
5. The method according to any one of claims 1 to 4, wherein: The first yaw rate prediction sub-model is a deep learning model, and the second yaw rate prediction sub-model is a convolutional model.
6. A yaw rate prediction device, comprising: A determination module is configured to determine a target route and a pre-trained yaw rate prediction model, wherein the yaw rate prediction model includes a first yaw rate prediction sub-model and a second yaw rate prediction sub-model; an acquisition module, configured to acquire first feature data and second feature data of the target route, and input the second feature data into a second yaw rate prediction sub-model to obtain third feature data, wherein the first feature data includes route features and / or navigation object features, the second feature data is segment feature data of a segment included in the route, represented as an N×M feature matrix, wherein N represents the number of segments, and M represents the number of attributes of a certain segment, and the third feature data is data that can reflect some features of the target route obtained based on the second feature data; The prediction module is configured to input the first characteristic data and the third characteristic data of the target route into the first yaw rate prediction sub-model to obtain a yaw rate prediction result for the target route.
7. The apparatus according to claim 6, further comprising a training module, wherein the training module is configured to: Determine a yaw rate prediction model to be trained that includes a first initial yaw rate prediction sub-model and a second initial yaw rate prediction sub-model; Get the training data set, where The training data set includes first feature data, second feature data and label data corresponding to the training data; The yaw rate prediction model to be trained is progressively trained based on the training data set to obtain a trained yaw rate prediction model.
8. An electronic device comprising a memory and at least one processor; wherein: The memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the at least one processor to implement the method steps described in any one of claims 1-5.
9. A computer-readable storage medium having computer instructions stored thereon, wherein the computer instructions, when executed by a processor, implement the method steps described in any one of claims 1 to 5.
10. A computer program product comprising a computer program / instructions, wherein: When the computer program / instructions are executed by a processor, the method steps described in any one of claims 1 to 5 are implemented.
