Traffic characteristic prediction method, electronic device and storage medium

Abstract

The disclosed embodiment discloses a method for predicting traffic characteristics, an electronic device, and a storage medium, the method comprising: obtaining prediction data of a to-be-predicted intersection; the prediction data comprising feature data of the to-be-predicted intersection, an upstream section of the to-be-predicted intersection, and a downstream section of the to-be-predicted intersection, as well as known traffic characteristics of entering the downstream section from the upstream section of the to-be-predicted intersection after passing through the to-be-predicted intersection before prediction; inputting the prediction data into a prediction model to obtain predicted traffic characteristics of entering the downstream section of the to-be-predicted intersection after passing through the to-be-predicted intersection from the upstream section of the to-be-predicted intersection. This technical solution can avoid confusion between the traffic characteristics of the intersection and the traffic characteristics of the upstream and downstream sections by taking the to-be-predicted intersection and its upstream section as a whole to count the traffic characteristics.

Description

Technical Field

[0001] The present disclosure relates to the field of navigation technology, and in particular to a traffic feature prediction method, an electronic device and a storage medium. Background Art

[0002] With the development of Internet technology, more and more applications need to provide services to users based on location service systems. One of the important services based on location service systems is dynamic traffic service. The information provided by dynamic traffic service includes whether the road section is congested, the average traffic characteristics of vehicles and / or the average traffic speed of vehicles. For example, when the user initiates a route planning request, an application with map navigation function usually plans a navigation route for the user based on dynamic traffic service and calculates the estimated travel time or estimated arrival time required for the user to reach the destination along the navigation route. However, due to the complex traffic conditions at the intersection, for example, the traffic conditions of different directions at the same intersection in the same period are different, which makes it difficult to predict the traffic characteristics of the intersection in a refined manner. Therefore, the existing technology predicts the traffic characteristics of the intersection as a whole, but this will cause the prediction result to not truly reflect the traffic capacity of the intersection, and then make the estimated travel time or estimated arrival determined based on the prediction result. Therefore, how to improve the degree of refinement of the prediction of traffic characteristics so that the prediction results are more in line with the actual situation is one of the technical problems that need to be solved at present. Summary of the invention

[0003] The embodiments of the present disclosure provide a traffic feature prediction method, an electronic device, and a storage medium.

[0004] In a first aspect, an embodiment of the present disclosure provides a method for predicting traffic characteristics, which includes:

[0005] Acquire prediction data of the intersection to be predicted; the prediction data includes feature data of the intersection to be predicted, an upstream section of the intersection to be predicted, and a downstream section of the intersection to be predicted, as well as known traffic features of passing through the intersection to be predicted from the upstream section of the intersection to be predicted and then entering the downstream section before prediction;

[0006] The prediction data is input into a prediction model to obtain predicted traffic characteristics of a downstream section of the intersection to be predicted after passing through the intersection to be predicted from the upstream section of the intersection to be predicted.

[0007] Furthermore, the known traffic characteristics include a known traffic characteristic sequence corresponding to a plurality of historical time segments arranged in time sequence; and the future traffic characteristics include a future traffic characteristic sequence corresponding to a plurality of future time segments arranged in time sequence.

[0008] Furthermore, the prediction data is input into a prediction model to obtain future traffic characteristics of a road section downstream of the intersection to be predicted after passing through the intersection to be predicted, including:

[0009] The time domain convolution layer in the prediction model is used to extract the sequence features of the known traffic feature sequence in the time domain dimension;

[0010] A hybrid model consisting of a linear sub-model and a deep neural network sub-model in the prediction model is used to predict the sequence features and feature data involving the intersection to be predicted, the upstream section and the downstream section of the intersection to be predicted, so as to obtain the predicted traffic feature sequence.

[0011] Furthermore, the feature data related to the intersection to be predicted, the upstream section of the intersection to be predicted and the downstream section of the intersection to be predicted include attribute features of the intersection to be predicted, attribute features of the upstream section and attribute features of the downstream section, as well as turning action features at the intersection to be predicted when entering the downstream section from the upstream section.

[0012] In a second aspect, an embodiment of the present disclosure provides a model training method, which includes:

[0013] Acquire sample data of a plurality of sample intersections and annotation information of the sample intersections; the sample data includes feature data related to the sample intersection, an upstream section and a downstream section of the sample intersection, and sample traffic features of entering the downstream section from the upstream section through the sample intersection before a historical time range; the annotation information includes actual traffic features of entering the downstream section from the upstream section through the sample intersection after the historical time range;

[0014] The sample data and the annotation information are used to train a prediction model so that the prediction model can be used to predict future traffic characteristics of the intersection.

[0015] Furthermore, the sample traffic features include a sequence of sample traffic features corresponding to a plurality of historical time segments arranged in time sequence; and the actual traffic features include a sequence of actual traffic features corresponding to a plurality of future time segments arranged in time sequence.

[0016] Further, the prediction model is trained using the sample data and the annotation information so that the prediction model is used to predict future traffic characteristics of the intersection, including:

[0017] Inputting the sample passage feature sequence into the time domain convolution layer in the prediction model to extract sequence features of the sample passage feature sequence in the time domain dimension;

[0018] Inputting the sequence features of the sample traffic feature sequence in the time domain dimension, and the feature data of the sample intersection, the upstream section and the downstream section of the sample intersection in the sample data into the hybrid model composed of the linear sub-model and the deep neural network sub-model in the prediction model to obtain an output result;

[0019] The model parameters of the prediction model are adjusted based on the gap between the output result and the actual traffic feature sequence.

[0020] Furthermore, the feature data involving the sample intersection, the upstream section and the downstream section of the sample intersection include attribute characteristics of the sample intersection, the upstream section and the downstream section, and turning action characteristics at the sample intersection when entering the downstream section from the upstream section.

[0021] In a third aspect, a location-based service provision method is provided in an embodiment of the present disclosure, which includes: providing location-based services to a service recipient using the intersection traffic characteristics predicted by the method described in the first aspect, or using the intersection traffic characteristics predicted by the prediction model obtained by the method described in the second aspect; the location-based services include: one or more of real-time traffic prediction, navigation route planning, or navigation route ETA prediction.

[0022] In a fourth aspect, an embodiment of the present invention provides a traffic feature prediction device, comprising:

[0023] A first acquisition module is configured to acquire prediction data of a to-be-predicted intersection; the prediction data includes feature data of the to-be-predicted intersection, an upstream section of the to-be-predicted intersection, and a downstream section of the to-be-predicted intersection, as well as known traffic features before prediction of a vehicle passing through the upstream section of the to-be-predicted intersection and then entering the downstream section;

[0024] The input module is configured to input the prediction data into the prediction model to obtain the predicted traffic characteristics of the downstream section of the intersection to be predicted after passing through the intersection to be predicted from the upstream section of the intersection to be predicted.

[0025] In a fifth aspect, an embodiment of the present invention provides a model training device, including:

[0026] A second acquisition module is configured to acquire sample data of a plurality of sample intersections and annotation information of the sample intersections; the sample data includes feature data related to the sample intersection, an upstream section and a downstream section of the sample intersection, and sample traffic features of entering the downstream section from the upstream section through the sample intersection before a historical time range; the annotation information includes actual traffic features of entering the downstream section from the upstream section through the sample intersection after the historical time range;

[0027] The training module is configured to train a prediction model using the sample data and the annotation information so that the prediction model can be used to predict future traffic characteristics of the intersection.

[0028] In the sixth aspect, an embodiment of the present invention provides a location-based service providing device, which provides location-based services to the service object by using the intersection traffic characteristics predicted by the above-mentioned traffic characteristics prediction device, or the intersection traffic characteristics predicted by the prediction model obtained by the above-mentioned model training device; the location-based service includes: one or more of real-time traffic prediction, navigation route planning or navigation route ETA prediction.

[0029] The functions can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the functions.

[0030] In one possible design, the structure of the above-mentioned device includes a memory and a processor, the memory is used to store one or more computer instructions that support the above-mentioned device to execute the above-mentioned corresponding method, and the processor is configured to execute the computer instructions stored in the memory. The above-mentioned device may also include a communication interface for the above-mentioned device to communicate with other devices or communication networks.

[0031] In a seventh aspect, an embodiment of the present disclosure provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the method described in any one of the above aspects.

[0032] In an eighth aspect, an embodiment of the present disclosure provides a computer-readable storage medium for storing computer instructions used by any of the above-mentioned devices, and when the computer instructions are executed by a processor, they are used to implement the method described in any of the above-mentioned aspects.

[0033] In a ninth aspect, an embodiment of the present disclosure provides a computer program product, which includes computer instructions, and when the computer instructions are executed by a processor, they are used to implement the method described in any of the above aspects.

[0034] The technical solution provided by the embodiments of the present disclosure may have the following beneficial effects:

[0035] In the disclosed embodiment, when using a pre-trained prediction model to predict the future traffic characteristics of a to-be-predicted intersection, the prediction data acquired for the to-be-predicted intersection includes, in addition to the characteristic data of the to-be-predicted intersection, the characteristic data of the upstream section and the characteristic data of the downstream section, the known traffic characteristics of entering the downstream section after passing through the to-be-predicted intersection from the upstream section. The prediction model predicts the future traffic characteristics of entering the downstream section after passing through the to-be-predicted intersection from the upstream section based on the above prediction data. The disclosed embodiment can avoid confusion between the traffic characteristics of the intersection and the traffic characteristics of the upstream and downstream sections by statistically analyzing the traffic characteristics of the intersection as a whole and the upstream and downstream sections. In addition, the known traffic characteristics input in the disclosed embodiment are the traffic characteristics of entering the downstream section after passing through the predicted intersection from the upstream section, and the output future traffic characteristics are also the traffic characteristics of entering the downstream section after passing through the predicted intersection from the upstream section. That is, the disclosed embodiment subdivides the traffic characteristics corresponding to the same intersection according to different traffic directions, which can improve the degree of refinement of the prediction of the traffic characteristics, make the prediction results more in line with the timing, and avoid the statistical errors caused by using a unified historical statistical value for the same intersection in the existing technology.

[0036] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Other features, objectives and advantages 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:

[0038] Figure 1 A flow chart showing a method for predicting traffic characteristics according to an embodiment of the present disclosure;

[0039] Figure 2 A schematic diagram showing the determination of the intersection travel time to be predicted in the embodiment of the present disclosure is shown;

[0040] Figure 3 A schematic diagram of the structure of a prediction model according to one embodiment of the present disclosure is shown;

[0041] Figure 4 A flowchart of a model training method according to an embodiment of the present disclosure is shown;

[0042] Figure 5 A schematic diagram showing an application of a traffic feature prediction service in a path planning scenario according to an embodiment of the present disclosure;

[0043] Figure 6It is a structural diagram of an electronic device suitable for implementing the traffic characteristic prediction method, model training method and / or location-based service provision method according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0044] 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.

[0045] In the present disclosure, it should be understood that terms such as "include" or "have" are intended to indicate the presence of features, numbers, steps, behaviors, components, parts, or a combination thereof disclosed in the specification, and do not exclude the possibility that one or more other features, numbers, steps, behaviors, components, parts, or a combination thereof exist or are added.

[0046] 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 present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0047] The details of the embodiments of the present disclosure are described in detail below through specific examples.

[0048] Figure 1 FIG. 1 is a flow chart showing a method for predicting traffic characteristics according to an embodiment of the present disclosure. Figure 1 As shown, the traffic feature prediction method includes the following steps:

[0049] In step S101, prediction data of a to-be-predicted intersection is obtained; the prediction data includes characteristic data of the to-be-predicted intersection, an upstream section of the to-be-predicted intersection, and a downstream section of the to-be-predicted intersection, as well as known traffic characteristics of a vehicle passing through the upstream section of the to-be-predicted intersection and entering the downstream section before prediction;

[0050] In step S102, the prediction data is input into a prediction model to obtain predicted traffic characteristics of a road section that passes through the intersection to be predicted from an upstream section of the intersection to be predicted and then enters a downstream section of the intersection to be predicted.

[0051] In this embodiment, the traffic characteristics may be the average traffic speed or average traffic time of the vehicle passing through the intersection. In some embodiments, the traffic characteristics of the intersection to be predicted may be the entire traffic characteristics starting from when the vehicle enters the upstream section and passes through the intersection to be predicted until it enters the downstream section.

[0052] Figure 2 FIG. 2 shows a schematic diagram of determining the traffic time of the intersection to be predicted in the embodiment of the present disclosure. Figure 2As shown, link A is the turning intersection J AB The upstream link of link B is the turning junction J AB downstream section.

[0053] To avoid link A, link B and turning intersection J AB There is a dependency and coupling between the travel time. In the embodiment of the present disclosure, the link and the intersection (such as Figure 2 Links A and J shown in AB ) composed of a whole (link + steering action, denoted as V AB ) as a prediction unit at the road section level, and then predict the prediction unit V AB The travel time from Figure 2 It can be seen that V AB The travel time of the downstream link B can be obtained by subtracting the entry time of the upstream link A from the entry time of the downstream link B. Therefore, the statistics of the travel time depend on the time from link A to J. AB Enter the driving trajectory of link B, wherein the driving trajectory includes link A through J AB Enter multiple consecutive GPS track points of link B.

[0054] Therefore, the prediction model in the embodiment of the present disclosure is used to predict the travel time of each intersection, and the travel time is the time obtained by statistically analyzing the section entering the intersection, that is, the upstream section mentioned above and the intersection as a whole. It can be understood as the travel time from entering the upstream section and passing through the intersection to entering the downstream section.

[0055] In some embodiments, the prediction model may adopt a neural network model based on deep learning.

[0056] In the process of using the prediction model to predict the traffic characteristics, the prediction data to be input into the prediction model in the embodiment of the present disclosure is data related to the current intersection to be predicted, which may include but is not limited to the characteristic data of the intersection to be predicted, the characteristic data of the upstream section of the intersection to be predicted, and the characteristic data of the downstream section of the intersection to be predicted. In addition to considering the characteristic data of the intersection to be predicted, the characteristic data of the upstream section of the intersection to be predicted, and the characteristic data of the downstream section of the intersection to be predicted to predict the traffic characteristics of the intersection to be predicted, the prediction model in the embodiment of the present disclosure also considers the known traffic characteristics of the intersection to be predicted before the prediction. Therefore, in order to predict the future traffic characteristics of the intersection to be predicted, the prediction data obtained may include but is not limited to the characteristic data of the intersection to be predicted, the characteristic data of the upstream section of the intersection to be predicted, the characteristic data of the downstream section of the intersection to be predicted, and the known traffic characteristics of the intersection to be predicted before the prediction.

[0057] It should be noted that the known traffic characteristics may be the traffic characteristics of a certain period before the prediction, and may include the current real-time traffic characteristics and the previous historical traffic characteristics. For example, if the traffic characteristics of the intersection to be predicted in the next period (such as the next hour) need to be predicted, the known traffic characteristics may include the known traffic characteristics within the current one hour; for another example, if the traffic characteristics of the intersection to be predicted on the next day need to be predicted, the known traffic characteristics may include the known traffic characteristics of the current day, etc.

[0058] It should also be noted that the prediction model can be obtained by pre-training. It can be seen from the input of the prediction model above that during the training process of the prediction model, the sample data collected is data related to the sample intersection, which may include but is not limited to the feature data of the sample intersection, the feature data of the upstream section of the sample intersection, and the feature data of the downstream section of the sample intersection; in addition, the sample data may also include sample traffic features of entering the downstream section after passing through the sample intersection from the upstream section in the historical period, and the label information corresponding to the sample data may be the actual traffic features of entering the downstream section after passing through the sample intersection from the upstream section after the historical period.

[0059] In the process of training the prediction model, sample data is input into the prediction model, and then the difference between the output of the prediction model and the actual traffic characteristics in the label information is compared, and the model parameters of the prediction model are adjusted based on the difference, so that the output predicted by the prediction model with adjusted model parameters is closer to the actual traffic characteristics. After training with a large amount of sample data, a trained prediction model can be obtained.

[0060] In the disclosed embodiment, when using a pre-trained prediction model to predict the future traffic characteristics of a to-be-predicted intersection, the prediction data acquired for the to-be-predicted intersection includes, in addition to the characteristic data of the to-be-predicted intersection, the characteristic data of the upstream section and the characteristic data of the downstream section, the known traffic characteristics of entering the downstream section after passing through the to-be-predicted intersection from the upstream section. The prediction model predicts the future traffic characteristics of entering the downstream section after passing through the to-be-predicted intersection from the upstream section based on the above prediction data. The disclosed embodiment can avoid confusion between the traffic characteristics of the intersection and the traffic characteristics of the upstream and downstream sections by statistically analyzing the traffic characteristics of the intersection as a whole and the upstream and downstream sections. In addition, the known traffic characteristics input in the disclosed embodiment are the traffic characteristics of entering the downstream section after passing through the predicted intersection from the upstream section, and the output future traffic characteristics are also the traffic characteristics of entering the downstream section after passing through the predicted intersection from the upstream section. That is, the disclosed embodiment subdivides the traffic characteristics corresponding to the same intersection according to different traffic directions, which can improve the degree of refinement of the prediction of the traffic characteristics, make the prediction results more in line with the timing, and avoid the statistical errors caused by using a unified historical statistical value for the same intersection in the existing technology.

[0061] In an optional implementation of this embodiment, the feature data related to the intersection to be predicted, the upstream section of the intersection to be predicted, and the downstream section of the intersection to be predicted include attribute features of the intersection to be predicted, attribute features of the upstream section, and attribute features of the downstream section, as well as turning action features at the intersection to be predicted when entering the downstream section from the upstream section.

[0062] In this optional implementation, the prediction data input to the prediction model may include feature data of the intersection to be predicted, the upstream section and the downstream section, and the feature data may include the attribute features of the intersection to be predicted, the attribute features of the upstream section of the intersection to be predicted and the attribute features of the downstream section of the intersection to be predicted, etc. The feature data may also include the turning action features of entering the downstream section from the upstream section through the intersection to be predicted. The attribute features of the intersection to be predicted may include, but are not limited to, simple intersections, complex intersections, the presence or absence of traffic lights, etc., and the attribute features of the intersection to be predicted may be pre-set, without limitation. The attribute features of the upstream section and the downstream section may include section attributes, and the section attributes may include, but are not limited to, the number of lanes, the length of the section, the grade of the section (such as national highways, expressways, county roads, sections in the city, main roads, auxiliary roads, etc.), the grade of the city, etc. The section attributes may also be pre-set, without limitation.

[0063] In the disclosed embodiment, when predicting the traffic characteristics of the intersection to be predicted, the prediction model is based on the turning action characteristics of the intersection to be predicted, in addition to the attribute characteristics of the intersection to be predicted, the attribute characteristics of the upstream section of the intersection to be predicted, and the attribute characteristics of the downstream section of the intersection to be predicted. The turning action characteristics refer to the turning action characteristics of entering the downstream section from the upstream section through the predicted intersection. In other words, when predicting the traffic characteristics of the intersection to be predicted, the prediction model will distinguish the factors affecting the traffic characteristics of different turns in different lanes or the same lane of the same intersection, and further refine the traffic characteristics of the intersection according to different turns, so that the traffic characteristics predicted by the prediction model are more accurate.

[0064] In an optional implementation of this embodiment, the known traffic features include a known traffic feature sequence corresponding to multiple historical time segments arranged in time sequence; the future traffic features include a future traffic feature sequence corresponding to multiple future time segments arranged in time sequence.

[0065] In this optional implementation, the traffic characteristics usually adopt the average traffic characteristics, that is, the traffic characteristics of multiple vehicles passing through the intersection within a period of time are collected, and then the average value is calculated to obtain the corresponding traffic characteristics. In order to make the predicted traffic characteristics more accurate, the embodiment of the present disclosure divides the time into multiple time segments, and the average traffic characteristics are obtained for each time segment, such as the average travel time or the average travel speed. In some embodiments, the time range of a time segment can be in minutes. For example, the known traffic characteristics can be the known traffic characteristics of the intersection to be predicted one hour before, then the known traffic characteristics can include the average traffic characteristics corresponding to all historical time segments in the previous hour. If the length of the time segment is 5 minutes, the known traffic characteristics are a sequence of 12 average traffic characteristics. Similarly, the predicted future traffic characteristics are also a sequence of multiple average traffic characteristics corresponding to multiple future time segments.

[0066] In an optional implementation of this embodiment, step S102, that is, the step of inputting the prediction data into the prediction model to obtain the future traffic characteristics of the downstream section of the intersection to be predicted after passing through the intersection to be predicted from the upstream section of the intersection to be predicted, further includes the following steps:

[0067] The time domain convolution layer in the prediction model is used to extract the sequence features of the known traffic feature sequence in the time domain dimension;

[0068] A hybrid model consisting of a linear sub-model and a deep neural network sub-model in the prediction model is used to predict the sequence features and feature data involving the intersection to be predicted, the upstream section and the downstream section of the intersection to be predicted, so as to obtain the predicted traffic feature sequence.

[0069] In this optional implementation, the prediction model may include multiple time-domain convolution layers and a hybrid model consisting of a linear sub-model and a deep neural network sub-model. The time-domain convolution layer is used to perform time-domain convolution on a known traffic feature sequence, thereby extracting sequence features based on the time dimension corresponding to the known traffic feature sequence. In some embodiments, the time-domain convolution layer may use a gated time-domain convolution, and the prediction model may set up two gated time-domain convolution layers. The known traffic feature sequence is input to the first time-domain convolution layer, and the output of the first time-domain convolution layer is used as the input of the second time-domain convolution layer. The output of the second time-domain convolution layer is the sequence feature corresponding to the known traffic feature sequence in the time domain dimension.

[0070] The sequence characteristics, the attribute characteristics of the intersection to be predicted, the attribute characteristics of the upstream section, the attribute characteristics of the downstream section, and the turning action characteristics performed at the intersection to be predicted when entering the downstream section from the upstream section are all input into the hybrid model composed of the linear sub-model and the deep neural network sub-model, and finally the future traffic feature sequence is output.

[0071] In some embodiments, the hybrid model composed of the linear sub-model and the neural network sub-model can adopt the existing wide&deep model. The specific details of the wide&deep model can refer to the existing technology and will not be repeated here.

[0072] Figure 3 FIG. 2 shows a schematic diagram of a prediction model according to an embodiment of the present disclosure. Figure 3 As shown in the figure, the prediction model consists of two sequentially connected temporal gated convolution layers. Given a known sequence of passing features (v1, v2, ..., v n )express Figure 2 V AB The speed sequence corresponding to n time segments within the current time range (for example, within the current hour) is obtained. After the known traffic feature sequence passes through two time domain convolutional network layers in sequence, the feature expression of the speed sequence is obtained. The feature expression of the speed sequence is consistent with the attributes of linkA, linkB, and J AB The properties of the intersection and J AB The turning action at the intersection is input to the Wide&Deep model, and the Wide&Deep model outputs the predicted value. The predicted value can be V AB The speed sequence of .

[0073] Figure 4 FIG. 1 is a flow chart of a model training method according to an embodiment of the present disclosure. Figure 4 As shown, the model training method includes the following steps:

[0074] In step S401, sample data of a plurality of sample intersections and annotation information of the sample intersections are obtained; the sample data includes feature data related to the sample intersection, the upstream section and the downstream section of the sample intersection, and sample traffic features of entering the downstream section from the upstream section through the sample intersection before a historical time range; the annotation information includes actual traffic features of entering the downstream section from the upstream section through the sample intersection after the historical time range;

[0075] In step S402, a prediction model is trained using the sample data and the annotation information, so that the prediction model is used to predict future traffic characteristics of the intersection.

[0076] In this embodiment, the sample traffic characteristics may be the average traffic speed or average traffic time of the vehicle passing through the intersection. In some embodiments, the sample traffic characteristics of the sample intersection may be the entire sample traffic characteristics starting from when the vehicle enters the upstream section and passes through the sample intersection and enters the downstream section.

[0077] In some embodiments, the prediction model may adopt a neural network model based on deep learning.

[0078] In the process of training the prediction model, sample data of various sample intersections can be collected based on navigation data. The navigation data may include but is not limited to the trajectory of the vehicle passing through the sample intersection. Based on the trajectory, the upstream section and downstream section of the vehicle passing through the sample intersection can be determined. The sample traffic characteristics such as the travel time and travel speed of the vehicle from entering the upstream section and passing through the sample intersection to the downstream section can also be calculated. Therefore, in order to train the prediction model, the collected sample data is data related to the sample intersection, which may include but is not limited to the characteristic data of the sample intersection, the characteristic data of the upstream section of the sample intersection, and the characteristic data of the downstream section of the sample intersection. In addition to considering the characteristic data of the sample intersection, the characteristic data of the upstream section of the sample intersection, and the characteristic data of the downstream section of the sample intersection to predict the sample traffic characteristics of the sample intersection, the prediction model in the embodiment of the present disclosure also considers the known sample traffic characteristics of the sample intersection before the prediction. Therefore, in order to enable the prediction model to accurately predict the future sample traffic characteristics of the intersection, the sample data for training it may include but is not limited to the characteristic data of the sample intersection, the characteristic data of the upstream section of the sample intersection, the characteristic data of the downstream section of the sample intersection, and the known sample traffic characteristics from entering the upstream section and passing through the sample intersection to the downstream section within the historical time range. The annotation information corresponding to the sample data may include the known sample traffic characteristics from entering the upstream section and passing through the sample intersection to the downstream section after the historical time range.

[0079] It should be noted that when collecting sample data, a sample intersection and a historical time range may be selected first. In some embodiments, multiple sample intersections may be selected, and multiple historical time ranges may also be selected. For each sample intersection, trajectory data within one or more selected historical time ranges and a period of time after the historical time range may be obtained from the navigation system, and then the trajectory data may be processed to obtain sample data corresponding to each sample intersection. It should be noted that the historical time range may be a time range in units of days, hours, etc., and the period of time after the historical time range may also be a time range in units of days, hours, minutes, etc.

[0080] In the process of training the prediction model, sample data is input into the prediction model, and then the difference between the output of the prediction model and the actual traffic characteristics in the label information is compared, and the model parameters of the prediction model are adjusted based on the difference, so that the output predicted by the prediction model with adjusted model parameters is closer to the actual traffic characteristics. After training with a large amount of sample data, a trained prediction model can be obtained.

[0081] In the embodiment of the present disclosure, during the process of training the prediction model, the sample data collected are data related to the sample intersection. In addition to the feature data of the sample intersection, the upstream section and the downstream section, the sample traffic characteristics of entering the downstream section from the upstream section after passing through the sample intersection within a historical time range are also included. The sample data is used as input data of the prediction model, and the label information used as supervision information is the actual traffic characteristics of entering the downstream section from the upstream section after passing through the sample intersection after the historical time range. The output of the prediction model is a predicted value for representing the traffic characteristics of entering the downstream section from the upstream section after passing through the sample intersection after the historical time range. The disclosed embodiment can avoid confusion between intersection traffic characteristics and traffic characteristics of upstream and downstream sections by statistically analyzing sample traffic characteristics of the sample intersection and its upstream section as a whole. In addition, since the sample traffic characteristics included in the sample data in the disclosed embodiment are sample traffic characteristics of entering the downstream section after passing through the sample intersection from the upstream section, and the label information of the prediction model is the actual traffic characteristics of entering the downstream section after passing through the sample intersection from the upstream section, that is, in the process of training the prediction model, the disclosed embodiment separately counts the sample traffic characteristics of vehicles heading in different directions corresponding to the same intersection. Therefore, the statistical errors caused by using a unified historical statistical value for the same intersection in the prior art can be avoided, and the prediction accuracy of the prediction model for intersection traffic characteristics can be improved.

[0082] In an optional implementation of the present embodiment, the feature data involving the sample intersection, the upstream section and the downstream section of the sample intersection include attribute features of the sample intersection, the upstream section and the downstream section, and turning action features at the sample intersection when entering the downstream section from the upstream section.

[0083] In this optional implementation, the sample data input to the prediction model may include feature data of the sample intersection, the upstream section and the downstream section, and the feature data may include attribute features of the sample intersection, attribute features of the upstream section of the sample intersection and attribute features of the downstream section of the sample intersection, etc. The feature data may also include the turning action features of entering the downstream section from the upstream section through the sample intersection. The attribute features of the sample intersection may include, but are not limited to, simple intersections, complex intersections, the presence or absence of traffic lights, etc., and the attribute features of the sample intersection may be pre-set, without limitation. The attribute features of the upstream section and the downstream section may include section attributes, and the section attributes may include, but are not limited to, the number of lanes, the length of the section, the grade of the section (such as national highways, expressways, county roads, urban sections, main roads, auxiliary roads, etc.), the grade of the city, etc. The section attributes may also be pre-set, without limitation.

[0084] In the disclosed embodiment, in the process of training the prediction model, in addition to the attribute features of the sample intersection, the attribute features of the upstream section of the sample intersection, and the attribute features of the downstream section of the sample intersection, the turning action features of the sample intersection are also based on the turning action features, which refer to the turning action features of entering the downstream section from the upstream section through the predicted intersection. In other words, in the process of training the prediction model, the prediction model will distinguish and learn the sample traffic features of different turns in different lanes or the same lane at the same intersection, so that the sample traffic features predicted by the prediction model are more accurate.

[0085] In an optional implementation of this embodiment, the sample traffic features include a sequence of sample traffic features corresponding to multiple historical time segments arranged in time sequence; the actual traffic features include a sequence of actual traffic features corresponding to multiple future time segments arranged in time sequence.

[0086] In this optional implementation, the sample traffic characteristics usually adopt the average traffic characteristics, that is, by collecting the traffic characteristics of multiple vehicles passing through the intersection within a period of time, and then calculating the average value to obtain the corresponding sample traffic characteristics. In order to enable the sample traffic characteristics to more accurately describe the traffic characteristics of the sample intersection, the embodiment of the present disclosure divides the time into multiple time segments, and statistically obtains the average traffic characteristics for each time segment, and the obtained sample traffic characteristics are a sequence composed of multiple average traffic characteristics corresponding to multiple time segments within the historical time range. In some embodiments, the time range of a time segment can be at the minute level. For example, the sample traffic characteristics can be the sample traffic characteristics of the sample intersection within a historical time range, such as a certain day and a certain hour. Then the sample traffic characteristics can include the average traffic characteristics corresponding to all historical time segments within the hour. If the time segment length is 5 minutes, then the sample traffic characteristics are a sequence of 12 average traffic characteristics corresponding to 12 historical time segments in the hour. Similarly, the actual traffic characteristics are a sequence of multiple actual traffic characteristics corresponding to all future time segments in a period of time after the historical time range. In the above example, the actual traffic characteristics can be the average traffic characteristic sequence corresponding to all future time segments within an hour after that hour.

[0087] In an optional implementation of this embodiment, step S402, i.e., the step of training a prediction model using the sample data and the annotation information so that the prediction model is used to predict future traffic characteristics of the intersection, further includes the following steps:

[0088] Inputting the sample passage feature sequence into the time domain convolution layer in the prediction model to extract sequence features of the sample passage feature sequence in the time domain dimension;

[0089] The sequence characteristics of the sample traffic feature sequence in the time domain dimension, and the feature data of the sample intersection, the upstream section and the downstream section of the sample intersection in the sample data are input into the hybrid model composed of the linear sub-model and the deep neural network sub-model in the prediction model, and the output result is the sequence characteristics of the sample traffic feature sequence in the time domain dimension;

[0090] The model parameters of the prediction model are adjusted based on the gap between the output result and the actual traffic feature sequence.

[0091] In this optional implementation, the prediction model may include multiple time-domain convolution layers and a hybrid model consisting of a linear sub-model and a deep neural network sub-model. The time-domain convolution layer is used to perform time-domain convolution on a known sample passage feature sequence, thereby extracting sequence features based on the time dimension corresponding to the known sample passage feature sequence. In some embodiments, the time-domain convolution layer may use a gated time-domain convolution, and the prediction model may set up two gated time-domain convolution layers. The sample passage feature sequence is input to the first time-domain convolution layer, and the output of the first time-domain convolution layer is used as the input of the second time-domain convolution layer. The output of the second time-domain convolution layer is the sequence feature corresponding to the sample passage feature sequence in the time domain dimension.

[0092] The sequence characteristics, attribute characteristics of the sample intersection, attribute characteristics of the upstream section, attribute characteristics of the downstream section, and the turning action characteristics performed at the sample intersection when entering the downstream section from the upstream section are all input into a hybrid model composed of a linear sub-model and a deep neural network sub-model, and finally the output result of the prediction model is obtained, which is the predicted value of the traffic feature sequence of the sample intersection after the historical time range.

[0093] By comparing the predicted value with the actual traffic feature sequence, for example, the predicted value and the actual traffic feature sequence can be substituted into a pre-constructed loss function to calculate the loss of this training, and then the loss is used to adjust the model parameters of the prediction model through the back propagation method. The model parameters may include model parameters of multiple time domain convolutional layers included in the prediction model, and model parameters of a hybrid model composed of a linear sub-model and a deep neural network model.

[0094] In some embodiments, the hybrid model composed of the linear sub-model and the neural network sub-model can adopt the existing wide&deep model. The specific details of the wide&deep model can refer to the existing technology and will not be repeated here.

[0095] According to one embodiment of the present disclosure, a location-based service providing method according to one embodiment of the present disclosure is used. The location-based service providing method uses the intersection traffic characteristics predicted by the above-mentioned traffic characteristics prediction method or the intersection traffic characteristics predicted by the prediction model obtained by the above-mentioned model training method to provide location-based services for the service object. The location-based service includes: one or more of real-time traffic prediction, navigation route planning or navigation route ETA prediction.

[0096] In this embodiment, the location-based service provision method can be executed on a terminal, which is a mobile phone, iPad, computer, smart watch, vehicle, etc. In the disclosed embodiment, the traffic characteristics of the intersection in the future period of time can be predicted using the above-mentioned method for any road section or a pre-selected road section such as a popular road section, and then in the location-based service process, the traffic characteristics can be used to provide more accurate location services for the service object, such as real-time traffic prediction services, navigation route planning services and / or navigation route ETA (estimated arrival time) prediction services, etc.

[0097] The service object may be a mobile phone, iPad, computer, smart watch, vehicle, robot, etc. When providing the service object with real-time traffic prediction service, navigation route planning service and / or navigation route ETA (estimated time of arrival) prediction service, the traffic characteristics of the road section can be obtained based on the above method, and then the travel time and speed of the service object on the road section can be predicted based on the traffic characteristics in real-time traffic prediction and navigation route planning, and the time to arrive at the navigation destination can be predicted based on the travel time and speed in ETA prediction. For details, please refer to the above description of the road section traffic characteristics prediction method, which will not be repeated here.

[0098] Figure 5 FIG. 1 is a schematic diagram showing an application of a traffic feature prediction service in a path planning scenario according to an embodiment of the present disclosure. Figure 5 As shown, in an offline scenario, the prediction server obtains the navigation data of vehicles in city A in a few months from the navigation server, and selects the vehicle trajectories of representative intersections from the navigation data. For each representative intersection, the average speed or average travel time of vehicles in all time segments of each day or week can be counted in units of days or weeks, and the average speed or average travel time corresponding to each time segment is expressed in the form of a sequence; in addition, it is also necessary to determine the attributes of the upstream and downstream sections corresponding to each intersection, the attributes of the intersection, and the turning action from the upstream section through the intersection to the downstream section. The prediction server uses the above data obtained to train the prediction model. The structure of the prediction model can be found in Figure 3 The structure shown.

[0099] In the online prediction stage, for each intersection to be predicted, the prediction server obtains the attribute features of the intersection, the attribute features of the upstream and downstream sections of the intersection, and the turning action features of the upstream section passing through the intersection to enter the downstream section from the navigation server, and obtains the vehicle trajectory passing through the intersection within a period of time before the prediction time point, and obtains the average travel time or average travel speed from the upstream section through the intersection to the downstream section in each time segment within the period of time based on the vehicle trajectory, forming a traffic feature sequence corresponding to the period of time. The traffic feature sequence is input into the time domain convolution layer of the prediction model to obtain the sequence features of the traffic feature sequence in the time domain dimension, and then the sequence features, the attribute features of the intersection, the attribute features of the upstream and downstream sections, and the turning action features of the upstream section passing through the intersection to enter the downstream section are input into the hybrid model composed of the linear sub-model and the deep neural network sub-model in the prediction model to obtain the traffic feature sequence of the intersection in the next time period.

[0100] The prediction server sends the predicted traffic feature sequence of each intersection to the navigation server. The navigation server can use the above traffic feature sequence to predict the future traffic features of the upstream and downstream sections related to the intersection, and then perform path planning based on the traffic features when receiving the vehicle's navigation request.

[0101] 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.

[0102] According to a traffic feature prediction device of 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. The traffic feature prediction device includes:

[0103] A first acquisition module is configured to acquire prediction data of a to-be-predicted intersection; the prediction data includes feature data of the to-be-predicted intersection, an upstream section of the to-be-predicted intersection, and a downstream section of the to-be-predicted intersection, as well as known traffic features before prediction of a vehicle passing through the upstream section of the to-be-predicted intersection and then entering the downstream section;

[0104] The input module is configured to input the prediction data into the prediction model to obtain the predicted traffic characteristics of the downstream section of the intersection to be predicted after passing through the intersection to be predicted from the upstream section of the intersection to be predicted.

[0105] In an optional implementation of the present embodiment, the feature data involving the intersection to be predicted, the upstream section of the intersection to be predicted and the downstream section of the intersection to be predicted include attribute features of the intersection to be predicted, attribute features of the upstream section and attribute features of the downstream section, as well as turning action features at the intersection to be predicted when entering the downstream section from the upstream section.

[0106] In an optional implementation of this embodiment, the known traffic features include a known traffic feature sequence corresponding to multiple historical time segments arranged in time sequence; the future traffic features include a future traffic feature sequence corresponding to multiple future time segments arranged in time sequence.

[0107] In an optional implementation of this embodiment, the input module includes:

[0108] An extraction submodule, configured to extract sequence features of the known traffic feature sequence in the time domain dimension using a time domain convolution layer in the prediction model;

[0109] The processing submodule is configured to use a hybrid model composed of a linear submodel and a deep neural network submodel in the prediction model to predict the sequence features and feature data involving the intersection to be predicted, the upstream section and the downstream section of the intersection to be predicted, so as to obtain the predicted traffic feature sequence.

[0110] The traffic characteristic prediction device in this embodiment corresponds to the traffic characteristic prediction method described above. For specific details, please refer to the description of the traffic characteristic prediction method above, which will not be repeated here.

[0111] According to a model training device in one 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. The model training device includes:

[0112] A second acquisition module is configured to acquire sample data of a plurality of sample intersections and annotation information of the sample intersections; the sample data includes feature data related to the sample intersection, an upstream section and a downstream section of the sample intersection, and sample traffic features of entering the downstream section from the upstream section through the sample intersection before a historical time range; the annotation information includes actual traffic features of entering the downstream section from the upstream section through the sample intersection after the historical time range;

[0113] The training module is configured to train a prediction model using the sample data and the annotation information so that the prediction model can be used to predict future traffic characteristics of the intersection.

[0114] In an optional implementation of the present embodiment, the feature data involving the sample intersection, the upstream section and the downstream section of the sample intersection include attribute features of the sample intersection, the upstream section and the downstream section, and turning action features at the sample intersection when entering the downstream section from the upstream section.

[0115] In an optional implementation of this embodiment, the sample traffic features include a sequence of sample traffic features corresponding to multiple historical time segments arranged in time sequence; the actual traffic features include a sequence of actual traffic features corresponding to multiple future time segments arranged in time sequence.

[0116] In an optional implementation of this embodiment, the training module includes:

[0117] A first input submodule is configured to input the sample passage feature sequence into the time domain convolution layer in the prediction model to extract sequence features of the sample passage feature sequence in the time domain dimension;

[0118] The second input submodule is configured to input the sequence characteristics of the sample traffic feature sequence in the time domain dimension, and the feature data of the sample intersection, the upstream section and the downstream section of the sample intersection in the sample data into the hybrid model composed of the linear submodel and the deep neural network submodel in the prediction model, and obtain the output result of the sequence characteristics of the sample traffic feature sequence in the time domain dimension;

[0119] The adjustment submodule is configured to adjust the model parameters of the prediction model based on the gap between the output result and the actual traffic feature sequence.

[0120] The model training device in this embodiment corresponds to the model training method described above. For specific details, please refer to the description of the model training method above, which will not be repeated here.

[0121] According to a location-based service providing device in one 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. The location-based service providing device uses the intersection traffic characteristics predicted by the above-mentioned traffic characteristics prediction device, or the intersection traffic characteristics predicted by the prediction model obtained by the above-mentioned model training device to provide location-based services to the service object; the location-based services include: one or more of real-time traffic prediction, navigation route planning, or navigation route ETA prediction.

[0122] The location-based service providing device in this embodiment corresponds to the location-based service providing method described above. For specific details, please refer to the description of the location-based service providing method above, which will not be repeated here.

[0123] Figure 6 It is a structural diagram of an electronic device suitable for implementing the traffic characteristic prediction method, model training method and / or location-based service provision method according to an embodiment of the present disclosure.

[0124] like Figure 6As shown, the electronic device 600 includes a processing unit 601, which can be implemented as a processing unit such as a CPU, a GPU, an FPGA, and an NPU. The processing unit 601 can perform various processes in the embodiments of any of the above methods of the present disclosure according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage part 608 into a random access memory (RAM) 603. In RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing unit 601, ROM 602, and RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0125] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed into the storage section 608 as needed.

[0126] In particular, according to an embodiment of the present disclosure, any method in the above referenced embodiment of the present disclosure 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 machine-readable medium, and the computer program includes a program code for executing any method in the embodiment of the present disclosure. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 609, and / or installed from a removable medium 611.

[0127] 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.

[0128] 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.

[0129] As another aspect, the present disclosure further provides a computer-readable storage medium, which may be a computer-readable storage medium included in the device described in the above 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 method described in the present disclosure.

[0130] 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 present disclosure is not limited to the technical solution 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 the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other.

Claims

1. A traffic feature prediction method, wherein: include: Acquire prediction data of the intersection to be predicted; the prediction data includes feature data of the intersection to be predicted, an upstream section of the intersection to be predicted, and a downstream section of the intersection to be predicted, as well as known traffic features of passing through the intersection to be predicted from the upstream section of the intersection to be predicted and then entering the downstream section before prediction; The prediction data is input into a prediction model to obtain future traffic characteristics of a road section downstream of the intersection to be predicted after passing through the intersection to be predicted from the upstream section of the intersection to be predicted.

2. The method according to claim 1, wherein: The known traffic features include a known traffic feature sequence corresponding to a plurality of historical time segments arranged in time sequence; the future traffic features include a future traffic feature sequence corresponding to a plurality of future time segments arranged in time sequence.

3. The method according to claim 2, wherein: The prediction data is input into a prediction model to obtain future traffic characteristics of a road section downstream of the intersection to be predicted after passing through the intersection to be predicted, including: The time domain convolution layer in the prediction model is used to extract the sequence features of the known traffic feature sequence in the time domain dimension; By using a hybrid model consisting of a linear sub-model and a deep neural network sub-model in the prediction model, the sequence features and feature data involving the intersection to be predicted, the upstream section and the downstream section of the intersection to be predicted are predicted to obtain the future traffic feature sequence.

4. The method according to any one of claims 1 to 3, wherein: The feature data involving the intersection to be predicted, the upstream section of the intersection to be predicted and the downstream section of the intersection to be predicted include the attribute features of the intersection to be predicted, the attribute features of the upstream section and the attribute features of the downstream section, as well as the turning action features at the intersection to be predicted when entering the downstream section from the upstream section.

5. A model training method, wherein: include: Acquire sample data of a plurality of sample intersections and labeling information of the sample intersections; The sample data includes characteristic data related to the sample intersection, the upstream section and the downstream section of the sample intersection, and sample traffic characteristics of entering the downstream section from the upstream section through the sample intersection before the historical time range; The annotation information includes actual traffic characteristics from the upstream section through the sample intersection into the downstream section after the historical time range; The sample data and the annotation information are used to train a prediction model so that the prediction model can be used to predict future traffic characteristics of the intersection.

6. The method according to claim 5, wherein: The sample traffic features include a sequence of sample traffic features corresponding to a plurality of historical time segments arranged in time sequence; the actual traffic features include a sequence of actual traffic features corresponding to a plurality of future time segments arranged in time sequence.

7. The method according to claim 6, wherein: The prediction model is trained using the sample data and the annotation information so that the prediction model is used to predict future traffic characteristics of the intersection, including: Inputting the sample passage feature sequence into the time domain convolution layer in the prediction model to extract sequence features of the sample passage feature sequence in the time domain dimension; Inputting the sequence features of the sample traffic feature sequence in the time domain dimension, and the feature data of the sample intersection, the upstream section and the downstream section of the sample intersection in the sample data into the hybrid model composed of the linear sub-model and the deep neural network sub-model in the prediction model to obtain an output result; The model parameters of the prediction model are adjusted based on the gap between the output result and the actual traffic feature sequence.

8. The method according to any one of claims 5 to 7, wherein: The feature data involving the sample intersection, the upstream section and the downstream section of the sample intersection include attribute features of the sample intersection, the upstream section and the downstream section, and the turning action features at the sample intersection when entering the downstream section from the upstream section.

9. A method for providing location-based services, wherein: The location-based service providing method uses the intersection traffic characteristics predicted by the method described in any one of claims 1 to 4, or the intersection traffic characteristics predicted by the prediction model obtained by the method described in any one of claims 5 to 8 to provide location-based services to the service object; the location-based services include: one or more of real-time traffic prediction, navigation route planning or navigation route ETA prediction.

10. An electronic device, wherein: The method comprises a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the method according to any one of claims 1 to 9.

Images